University of Southern California Dissertations and Theses

Three essays in international macroeconomics and finance

(USC Thesis Other)

Three essays in international macroeconomics and finance

PDF

Download a page range

Download transcript

Copy asset link

Request this asset

Transcript (if available)

Content Three Essays in International Macroeconomics and Finance by Bada Han A Dissertation Presented to the FACULTY OF THE GRADUATE SCHOOL UNIVERSITY OF SOUTHERN CALIFORNIA In Partial Fulfillment of the Requirements for the Degree DOCTOR OF PHILOSOPHY (Economics) May 2021 Copyright 2021 Bada Han I dedicate this thesis to my wife, Jueun Cha, for being with and supporting me during my long journey. ii Acknowledgements My research would not have been possible without the help of a lot of people. First, I owe a lot to my wife, Ju-Eun Cha. My wife Ju-Eun has been with me during my long journey to pursue a Ph.D. degree in economics. In particular, I would not have been able to bear the stress of finishing my thesis amid the global pandemic, if she has not been staying with me since the beginning of the pandemic. It has not been so easy to live in the US during the pandemic, isolating ourselves from the virus, and she has not gone back to Korea just to support my studies. I am also grateful for my parent, Sang-Yool Han, and Shin-Gu Kim. They have supported my desire to pursue graduate studies in the US and believed in me even when the going got tough. I am thankful for the opportunity to be guided by Professors Joshua Aizenman, Caroline Betts, Romain Ranciere, and Wenhao Li. In particular, it was so great for me to be exposed to different approaches in different fields in economics. Fortunately, I was guided by the benevolent advisors whose fields and preferred methodologies are different among themselves. The discussion with them has led me to examine different approaches so that I can form a comprehensive view on my research agendas and exhibit a set of different skills in the dissertation. More precisely, Joshua introduced me to the debate of the global financial cycle and encouraged me to go through the re- lated literature. Furthermore, he urged me to study the histories and institutional details in different emerging market countries, to make sure my theoretical and empirical studies are firmly grounded on reality. Romain helped me with forming my research question more precisely and sharpening my idea when I was in the fourth year of my doctoral study. The experience as a research assis- tant for him also provided me with an opportunity to observe the development of a high-quality forefront research project. Betts also helped me with sharpening my idea and presentations. I still have a lot to improve, but her training altered me to a better presenter. The advice from Wenhao iii has enabled me to look at my research from a more finance-oriented perspective. He advised me to look at more disaggregated data to have more rigorous empirical evidence for the theoretical findings in my job market paper. Also, I was able to frame my job market in the spirit of joint hypothesis testing, thanks to his suggestions. I am also benefited from numerous interactions with my classmates and friendly environments at USC. I especially thank Rashad Ahmed, Yu Cao, Dario Laudati, Dongwook Kim, Yimeng Xie, Weining Xin, and all other students for listening to me and discussing my researches. The professional supports from the department staff have enabled me to focus on my researches. Lastly, I am also indebted to my colleagues at Bank of Korea. The dissertation is in part moti- vated by my observations and experiences at the bank. I am lucky to join the research department in 2009 January when the Korean economy was severely hit by massive capital outflows during the Global Financial Crisis. All the working experiences and discussions with my colleagues helped me with developing my research questions and finding the answers to the questions. I especially pay my respects to Ho-Il Oh for his incredible insights. iv Table of Contents Dedication ii Acknowledgements iii List of Tables viii List of Figures ix Abstract xi Chapter 1: Original Sin Dissipation and Currency Exposures in Emerging Markets 1 1.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1 1.2 Data . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8 1.2.1 Sources and Strategy . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8 1.2.2 Aggregate Level Data . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10 1.2.3 Sectoral Level Data . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17 1.3 Currency Exposures in EMEs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19 1.3.1 Aggregate Level Currency Exposures . . . . . . . . . . . . . . . . . . . . 19 1.3.2 Sectoral Level Currency Exposures . . . . . . . . . . . . . . . . . . . . . 22 1.4 Original Sin Dissipation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 30 1.5 Concluding Remarks . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 37 References 39 Chapter 2: Transmission of Global Financial Shocks: Which Capital Flows Matter? 43 2.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 43 2.2 Empirical Analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 52 2.2.1 Empirical Strategies . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 52 2.2.2 Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 55 2.3 Model . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 61 2.3.1 Simple Model . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 62 2.3.1.1 Environments . . . . . . . . . . . . . . . . . . . . . . . . . . . 62 2.3.1.2 Market Equilibrium . . . . . . . . . . . . . . . . . . . . . . . . 69 2.3.1.3 Inspecting the Mechanism . . . . . . . . . . . . . . . . . . . . . 70 2.3.2 Microlevel Evidence of the Capital Market Channel . . . . . . . . . . . . 83 2.4 Quantitative Exercise . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 87 v 2.4.1 Environments . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 87 2.4.2 Calibration . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 96 2.4.3 Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 102 2.4.3.1 Transmission of Global Financial Shocks in Korea . . . . . . . . 102 2.4.3.2 Quantitative evaluation of the importance of the capital market channel . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 108 2.5 Concluding Remarks . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 110 References 112 Chapter 3: International Reserve Accumulation: Balancing Private Inflows with Public Outflows (with Dongwook Kim) 118 3.1 Empirical Regularities . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 126 3.1.1 General Facts . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 126 3.1.2 Reserve Accumulation and ”Extra” Capital Inflows . . . . . . . . . . . . . 131 3.2 Model . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 136 3.2.1 Model Setup . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 136 3.2.1.1 Discussion of assumptions . . . . . . . . . . . . . . . . . . . . 143 3.2.2 Solving the Model . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 146 3.2.2.1 Decentralized Equilibrium . . . . . . . . . . . . . . . . . . . . . 146 3.2.2.2 Equilibrium with Social Planner . . . . . . . . . . . . . . . . . 155 3.2.2.3 Reserve accumulation and currency manipulation . . . . . . . . 166 3.3 Extensions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 172 3.3.1 The Model with Heterogeneous Agents . . . . . . . . . . . . . . . . . . . 173 3.3.2 The Infinite Horizon Model . . . . . . . . . . . . . . . . . . . . . . . . . 176 3.3.3 Endogenous Direct Investments and Capital Price . . . . . . . . . . . . . . 182 3.4 Concluding Remarks . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 185 References 189 Appendices 194 Chapter 4: Appendices 194 A Appendix to Chapter 1 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 194 A.1 Local currency debt and equity liability data . . . . . . . . . . . . . . . . . 194 A.1.1 Sector Level Foreign Currency Assets and Liabilities . . . . . . 199 A.1.2 Additional Tables . . . . . . . . . . . . . . . . . . . . . . . . . 200 B Appendix to Chapter 2 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 203 B.1 Other Data . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 203 B.2 Portfolio of Global Investors . . . . . . . . . . . . . . . . . . . . . . . . . 204 B.3 Pricing to Markets and Exchange Rate Channel . . . . . . . . . . . . . . . 206 B.4 Who Can Borrow More in Equity and LC Bond? . . . . . . . . . . . . . . 212 B.5 Omitted Algebras and Proofs . . . . . . . . . . . . . . . . . . . . . . . . . 215 B.6 Estimation of the Parameters . . . . . . . . . . . . . . . . . . . . . . . . . 221 B.7 Contagion to Credit Market . . . . . . . . . . . . . . . . . . . . . . . . . 225 B.8 Regressions with Sector Level Currency Mismatches . . . . . . . . . . . . 229 vi B.9 Additional Empirical Results . . . . . . . . . . . . . . . . . . . . . . . . . 232 C Appendix to Chapter 3 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 233 C.1 List of Countries and Omitted Figures and Tables . . . . . . . . . . . . . . 233 C.1.1 List of Countries used in Regressions . . . . . . . . . . . . . . . 233 C.1.2 Omitted Figures . . . . . . . . . . . . . . . . . . . . . . . . . . 233 C.1.3 Omitted Tables . . . . . . . . . . . . . . . . . . . . . . . . . . . 234 C.2 Omitted Algebras and Proofs . . . . . . . . . . . . . . . . . . . . . . . . . 236 C.3 Other Mircofoundations of Frictions on Capital Outflows . . . . . . . . . . 245 C.4 Reserve Accumulation of Saving Glut EMEs . . . . . . . . . . . . . . . . 247 C.5 Capital Outflow Restrictions in EMEs . . . . . . . . . . . . . . . . . . . . 249 vii List of Tables 1.1 Foreign Currency Positions of EMEs . . . . . . . . . . . . . . . . . . . . . . . . . 23 2.1 Exchange Rate Regressions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 55 2.2 Stock Indices Regressions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 57 2.3 Capital Market Channel Regressions . . . . . . . . . . . . . . . . . . . . . . . . . 84 2.4 Assigned Parameters . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 99 2.5 Estimated Parameters . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 101 2.6 Variance Decomposition . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 109 3.1 (a) Extra capital inflows . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 134 3.2 The U.S Treasury’s Foreign Exchange Report . . . . . . . . . . . . . . . . . . . . 170 A.1 Sources for the Local Currency Equities and Debts . . . . . . . . . . . . . . . . . 201 A.2 External Liabilities of EMEs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 202 A.3 Correlations with Local Currency External Liabilities . . . . . . . . . . . . . . . . 202 B.1 GMM estimation results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 223 B.2 Sensitivity of the Interest Rates to VIX . . . . . . . . . . . . . . . . . . . . . . . . 224 B.3 Contagion to Credit Market . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 228 B.4 Exchange Rate Regressions Sector level . . . . . . . . . . . . . . . . . . . . . . . 230 B.5 Stock Indices Regressions Sector level . . . . . . . . . . . . . . . . . . . . . . . 231 C.1 (a) The whole sample period (1998-2017) . . . . . . . . . . . . . . . . . . . . . . 236 viii List of Figures 1.1 Local External Debts in Korea . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11 1.2 International Investment Position and Local Currency Debt . . . . . . . . . . . . . 12 1.3 Foreign Currency Borrowing and Lending . . . . . . . . . . . . . . . . . . . . . . 17 1.4 Changes in Net Foreign Currency Positions . . . . . . . . . . . . . . . . . . . . . 20 1.5 Sector Level Currency Mismatches . . . . . . . . . . . . . . . . . . . . . . . . . . 25 1.6 Net Foreign Currency Position to Capital Ratios . . . . . . . . . . . . . . . . . . . 28 1.7 Local Currency External Liabilities . . . . . . . . . . . . . . . . . . . . . . . . . . 31 1.8 Capital Markets and Local Currency Borrowing . . . . . . . . . . . . . . . . . . . 35 2.1 VIX and Financial Markets in Selected EMEs . . . . . . . . . . . . . . . . . . . . 45 2.2 The Loop Mechanism in the Model . . . . . . . . . . . . . . . . . . . . . . . . . . 47 2.3 Stock Index Betas and the increase in LC external liabilities . . . . . . . . . . . . . 60 2.4 Capital Market Channel . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 73 2.5 Market Crashes from Risk-Off Shocks . . . . . . . . . . . . . . . . . . . . . . . . 82 2.6 Simulated Capital Prices and Exchange Rates with VIX . . . . . . . . . . . . . . . 103 2.7 Impulse Response Functions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 105 2.8 Impulse Response Functions, LC vs. FC . . . . . . . . . . . . . . . . . . . . . . . 107 3.1 Reserve Accumulation of EMEs . . . . . . . . . . . . . . . . . . . . . . . . . . . 127 3.2 NFA ex-IR and International Reserves . . . . . . . . . . . . . . . . . . . . . . . . 128 3.3 External Liability and Asset of EMEs . . . . . . . . . . . . . . . . . . . . . . . . 129 3.4 Private External Assets and Reserves . . . . . . . . . . . . . . . . . . . . . . . . . 130 3.5 Current Account and Reserve Accumulation . . . . . . . . . . . . . . . . . . . . . 131 ix 3.6 Reserves and FDI & Equity Liability . . . . . . . . . . . . . . . . . . . . . . . . . 132 3.7 Reserve Outflows and Extra Capital Inflows on Selected EMEs . . . . . . . . . . . 133 3.8 Comparative statics of the households decision . . . . . . . . . . . . . . . . . . . 152 3.9 Decentralized Equilibrium . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 156 3.10 Reserve accumulation without capital control . . . . . . . . . . . . . . . . . . . . 160 3.11 Passive Reserve Accumulation . . . . . . . . . . . . . . . . . . . . . . . . . . . . 169 3.12 Reserve Accumulation and Currency Manipulation . . . . . . . . . . . . . . . . . 171 3.13 Flow of Funds . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 174 3.14 Reserve Depletion during Sudden Stops . . . . . . . . . . . . . . . . . . . . . . . 180 B.1 Stock Market Capitalization, Trade Openness, and LC external liabilities . . . . . . 213 B.2 Contagion to Credit Market . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 226 C.1 Reserve Outflows and Extra Capital Inflows . . . . . . . . . . . . . . . . . . . . . 234 C.2 External Asset Structure and GDP per capita . . . . . . . . . . . . . . . . . . . . . 235 C.3 Capital Outflow Restrictions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 250 x Abstract In this dissertation, I study the changes in external liabilities and assets of emerging market economies since the early-2000s and the following implications of the changes for the financial stability and the optimal policies of the economies. In chapter 1, I construct a dataset, which measures the ex- ternal liability composition of emerging market economies in different instruments and currencies. The new dataset shows emerging market economies have much lower currency exposures than in the past. Also, the observed pattern in the dataset suggests that the ever-increasing local currency external borrowings of the emerging market economies since the early 2000s, original sin dissi- pation, is related to the capital market development in emerging market economies. Chapter 2 is a study of channels through which risk-appetite shocks to global investors, i.e., global financial shocks, are transmitted to emerging market economies. First, I empirically show that much of the transmission of global financial shocks to emerging market economies is reflected in equity and lo- cal currency bond portfolio investment capital flows. I then develop a small open economy model which, augmented with leverage constrained banks and foreign investors who purchase equities and bonds, can replicate these empirical findings qualitatively. Quantitative analysis of the model suggests that global financial shocks can account for 50% of the equity price volatility and 30% of the investment volatility in Korea, in which most of the external liabilities of the country are Korean won-denominated equities and debts. In short, all the analysis in chapter 2 implies that to a substan- tial extent, risk-appetite shocks to global investors are transmitted to emerging market economies via fickle portfolio capital flows to equity and local currency bond markets in the economies. In chapter 3, Dongwook Kim and I provide a novel theory of international reserve accumulation of emerging market economies. We view reserve accumulation as capital outflows by the public sec- tor which supplements insufficient capital outflows by the private sector. In our model, when an xi emerging market economy receives large capital inflows in the form of direct or equity portfolio investment, the emerging market economy must invest abroad to maintain macroeconomic balance and prepare for a possible future sudden stop. If the private sector in the emerging market economy cannot invest externally sufficiently or invests inefficiently due to low financial expertise or poor institutional quality, supplemental international investments must be accomplished by the public sector as international reserve outflows. xii Chapter 1 Original Sin Dissipation and Currency Exposures in Emerging Markets 1.1 Introduction Until the mid-2010s, it had been widely believed that Emerging Market Economies (EMEs) can- not borrow abroad in their currency, local currency, as foreign investors are not willing to hold local currency-denominated assets. 1 The seminal paper by Eichengreen and Hausmann (1999) formalized the idea, the famous ”Original Sin Hypothesis,” saying most countries are not able to borrow abroad in their domestic currency. Accordingly, the reliance of EMEs on foreign currency borrowings has been pointed out as a source of the fragility of those economies, and potential risk from foreign currency debts has been extensively studied in the literature. However, a few recent articles such as Arslanalp and Tsuda (2014) and Du and Schreger (2016 a,b) documented that the ability of EMEs to borrow in their local currency has significantly improved: the original sin has been ”dissipated.” This paper contributes to the literature by illustrating the original sin dissipation in EMEs since the mid-2000s. In this paper, by deploying different sources, I construct a dataset, which shows the currency composition of external assets and liabilities in EMEs at both the aggre- gate and the sectoral levels. While I discuss much detail in the data, I focus on the two aspects: 1) 1 Throughout this paper, I use nonresidents and foreign investors interchangeably. Of course, the two definitions differ in a certain context; for example, off-shore foreign currency debts. However, the dataset in this paper follows the traditions concept in the International Investment Positions. I corrected the amounts of foreign currency debts based on the nationality criteria, if necessary. 1 the amounts of local currency external liabilities in EMEs and accordingly 2) currency exposures in EMEs. The information on the currency composition of external assets and liabilities is not readily available yet, and there is no single complete and credible dataset. However, central banks or sta- tistical authorities in many EMEs begin including the currency composition information in their own International Investment Position (IIP) or external debt data. Those data are in public in dif- ferent forms or close to the outside of the institutions. I hand-collected the different data from sixteen EMEs, 2 while understanding institutional details in different EMEs. The level of informa- tion varies among the EMEs, and I make a reasonable guess in cases that available information in the data is not sufficient enough. In addition, while going through the data of each country, I learned the local currency debts can be approximated in a certain way, if necessary. In most of the EMEs, the local currency debts excluding local currency FDI debts, are the sum of domestic debt securities purchased by foreign investors and ”related” local currency deposits held by the foreign investors. In fact, most of the domestic debt securities held by foreign investors are local currency- denominated sovereign bonds, and most short-term deposits held by foreign investors are local currency-denominated deposits in many countries. 3 Then using the sovereign debt dataset by Ar- slanalp and Tsuda (2014) and different national sources, I construct the local currency debts of four more EMEs. 4 By combining this information with IIP data from IMF, I can identify the currency composition of different categories in IIP data at both the aggregate and the sectoral levels. The currency composition of external liabilities and assets is still not sufficient to capture cur- rency exposures at the sectoral level. As it has been discussed in the dollarization literature, entities in emerging markets lend and borrow in foreign currency among themselves. Using the dollariza- tion data constructed by Dalgic, I can approximate both foreign currency lending and borrowing among different sectors in EMEs, and therefore I can reasonably estimate net foreign currency 2 Barzil, Chile, Czech Republic, Hungary, India, Indonesia, Korea, Malaysia, Mexico, Philippine, Poland, Russia, South Africa, Thailand and Turkey 3 More precisely, the local currency deposits of foreign investors are related to the portfolio investments in local currency denominated security markets. I will discuss in detail later. 4 Argentina, Columbia, Peru, and Romania 2 debts, i.e., currency exposures of the different sectors — central bank, government, household, financial corporate sector, and nonfinancial corporate sector — in each of the EMEs. The first major finding from the data is the decline of currency exposures in EMEs. The decline of the currency exposures is clear at the aggregate level in the sample EMEs: almost all the EMEs in the sample have improved their aggregate foreign currency positions, measured by the ratio of net foreign currency debt to GDP. 5 In 2001, the average net foreign currency debt to GDP ratio across the twenty countries was around 7%, but the same ratio is around -22% in 2019: the negative net foreign currency debts and hence the positive net foreign currency assets. In short, the EMEs were net short in foreign currency in 2001, but now the countries are net long in foreign currency. The improvement becomes much more clear and stronger once I include portfolio equity assets in the foreign currency assets as the portfolio equity assets of the EMEs are mostly denominated in foreign currency. The improvement is attributable to both the decline of foreign currency debts and the accumulation of foreign currency assets, but more to the foreign currency asset accumulation. The increase in the foreign currency assets was mainly driven by international reserve accumulation by central banks in the 2000s, but, in the 2010s, it is attributable more to foreign currency asset accumulation by private sectors in the EMEs. The improvement in the foreign currency position is also observable at the sectoral level. For the sample period of 2010-2018, I identify the net foreign currency debts of the five different sectors I listed above; households, government, central bank, financial corporate sector, and nonfi- nancial corporate sector. The first observation from the data is the public sectors in the EMEs, as the agglomeration of government and central bank, has large net long positions in foreign currency. Central banks have accumulated tremendous amounts of international reserves, while many gov- ernments have rapidly decreased foreign currency bond issuance as discussed in the local currency sovereign debt literature. The most noteworthy result is the ”balanced” foreign currency positions 5 Throughout this paper, foreign currency denotes a hard currency, which is used in international transactions, such as US dollar, Euro, Japanese Yen or Swiss Franc. It is well known that almost all foreign currency debts in EMEs are denominated in US dollar, except for Eastern European countries where much of the foreign currency debts are denominated in Euro. Most of the portfolio investments or loans in foreign currency in EMES are also mostly hard currency denominated as documented in B´ en´ etrix et al. (2020). 3 in the financial sectors in most of the EMEs. The financial intermediaries have significant amounts of foreign currency debts, but those financial intermediaries also hold corresponding amounts of foreign currency assets, foreign currency loans to domestic companies, or foreign currency bonds issued in international markets. In contrast, nonfinancial corporate sectors in the EMEs have siz- able net foreign currency debts. As of 2018, the average of net foreign currency debts of the nonfinancial corporate sector to GDP across the EMEs is around 15% and the ratio is above 30% of GDP in some EMEs such as Bulgaria or Turkey. This is in line with recent papers such as Chui et al. (2016). However, unlike the preceding papers, I abstain from concluding that nonfinancial corporations in EMEs have serious currency exposures. First, nonfinancial corporate sectors in EMEs have low leverage ratios as documented in Beltran et al. (2017). In many EMEs, the aver- age nonfinancial corporate debt to equity ratio is less than 1 and thus the net foreign currency debt to the capital ratios as of 2018 suggest modest currency exposures of the nonfinancial corporate sectors in many of the EMEs. The average net foreign currency debt to the capital ratio is around 0.25, which is far lower than the levels at the dawn of the East Asian Crisis in 1997. Second, many nonfinancial corporations in the EMEs are obviously exporters and local currency deprecia- tion boosts the profitability of the exporters because the revenues from exports, computed in local currency, will increase if the exporting goods are foreign currency-denominated, as suggested in the Dominant Currency Pricing (DCP) hypothesis. Another important goal in this paper is to trace out the original sin dissipation of the EMEs so that we can have a hint on the underlying reasons for the important changes. I document and discuss detailed patterns of the original sin dissipation in different EMEs: when an EME began borrowing abroad in their local currency equity and debt and how the local currency borrowings have evolved in the EMEs. The most important finding regarding the original sin dissipation is the relationship between the capital market developments in the EMEs and the original sin dissipation. I find that the external borrowings of local currency equities and bonds are strongly correlated with domestic capital market developments. The correlation between the foreign nonresident investors’ local currency equity holdings to GDP ratio and the stock market capitalization to GDP ratio is 4 0.82, and the correlation between the nonresident portfolio investors’ local currency bond holdings to GDP ratio and the public bond market capitalization to GDP ratio is 0.61. Such high correlations are clear in both cross-country comparison and time series data of the individual country. Further- more, none of the other variables of economic fundamentals, such as measures of institutional quality, GDP per capita, or financial openness index, show such high correlations with external borrowings of local currency equities and bonds. A complete explanation of the correlations is beyond the scope of this paper, but one com- pelling explanation is as follows. The original sin has been dissipated over the last two decades, but the hypothesis is still valid in that almost all EMEs cannot issue local currency loans or bonds in international markets even these days. The most of external borrowings of local currency equity and debt are the foreign portfolio flows into the local currency-denominated stock and bond mar- kets in EMEs. Then a straightforward explanation is that investors from advanced economies were willing to hold equities and bonds in EMEs since the early-2000s, but there were just not enough stocks and bonds to be sold to the investors. As the capital markets in EMEs, the stock and bond markets, have grown, more equities and bonds have become available for the investors to purchase. Related Literature First and foremost, this paper belongs to a strand of literature that studied currency exposures in EMEs. After the sudden stops in the 1990s, Mexico peso crisis in 1994 and the East Asian crisis in 1997, several papers such as Eichengreen et al (2003), Goldstein and Turner (2004), and Eichengreen and Hausmann (2005) tried to identify the currency mismatches in EMEs and discussed the potential risk from foreign currency external debts in the EMEs. The seminal works by Lane and Milesi-Ferretti (2001, 2007a) constructed the national balance sheet data and traced out the evolution of the structures of external assets and liabilities of different countries in the world. Based on the progress in Lane and Milesi-Ferretti (2001, 2007a), Lane and Shambaugh (2010) constructed a database of international currency exposures for a large panel of countries, and documented that EMEs had reduced their currency exposures from the 1990s to the mid-2000s. B´ en´ etrix et al. (2015) and B´ en´ etrix et al. (2020) extended and enhanced the dataset in Lane and Shambaugh (2010). Those papers also documented the decline of currency exposures in EMEs. 5 In a similar fashion, B´ en´ etrix et al. (2015) and Hale and Juvenal (2020) estimated the valuation effects of exchange rate movements, local currency depreciation against US dollar or Euro, in EMEs, during the Global Financial Crisis and the early stage of COVID-19 crisis respectively. Similar to the preceding papers, they concluded that the valuation effects were not devastating to EMEs, at least on the aggregate level. The dataset constructed in this paper complements the dataset developed in the papers mentioned. While the recent dataset in B´ en´ etrix et al. (2020) covers a longer period for a larger set of countries, this paper provides more detailed and possibly more accurate information for a shorter period of a smaller number of countries. Unlike the B´ en´ etrix et al. (2020) that used synthetic data, the dataset in this paper is mostly based on the actual data from central banks, treasuries, or statistical authorities in the EMEs. More importantly, the dataset in this paper provides an estimate of currency exposures at the sectoral level, in particular the exposures of the financial and nonfinancial corporate sectors, which are absent in the B´ en´ etrix et al. (2020). On the other hand, this paper echoes the findings from the preceding papers. EMEs have reduced their currency exposures and the level of risk from foreign currency debts is incomparably lower than what it was in the 1990s or the early 2000s. Another strand of literature studied potential risk from foreign currency debts, missed in the aggregate level data. A few papers after the Global Financial Crisis focused on the possibility that the conventional residence classification misses important aspects of currency exposures in EMEs. Several papers by the authors at BIS (McCauely et al., 2015; Chui et al., 2016) pointed out that corporations in EMEs began borrowing through their funding vehicles established in for- eign countries, which they called “off-shore” borrowing, and this is missed in the conventional residency-based statistics. Some articles in the literature studied potential risk from domestic bor- rowing and lending in foreign currency, the prevalent phenomenon in some EMEs that the literature named as “dollarization.” The influential early paper Levy-Yeyati (2006) analyzed destabilizing effects of dollarization in EMEs and a more recent paper by Bocola and Lorenzoni (2020) built a stylized small open economy model where dollarization of domestic deposits, due to insurance motivation, amplifies the negative impacts of local currency depreciation on the economy. Several 6 recent papers, Dalgic (2020), Christiano et al. (2020), and Montamat (2020), suggest an opposite view. Those papers view the dollarization as a result of risk-hedging of entities in EMEs, and provided suggestive evidence along with a theoretical model consistent with the evidence. This paper accommodates all the empirical findings into the dataset in this paper. I construct different measures of currency exposures with dollarization of domestic borrowing and the off-shore bor- rowing. I found no evidence that the consideration of those foreign currency borrowing can alter the trend in EMEs for the last two decades: the currency exposures have decreased although we consider the dollarization or off-shore borrowings. The off-shore foreign currency debts raise the currency exposures of certain EMEs, for example, Brazil or Russia, but the decline of the currency exposures are even stronger in those countries. This paper is also related to a growing literature about local currency-denominated sovereign debts in EMEs. Unlike the Original Sin hypothesis, many EMEs have been able to issue local currency sovereign bonds and sell the bonds to foreign investors since the early-2000s, and the literature has explored what are the empirical characteristics of local currency sovereign bonds and what derives these features. Usually, the papers in the literature assumed the limited commitment of the EME governments and studied how the incentive problem derives empirical facts observed in the data. Noteworthy papers are Du and Schreger (2016a,b), Du et al. (2016), Engel and Park (2018), Ottonello and Perez (2018), and Alfaro and Kanczuk (2018). In these papers, governments in EMEs are incentivized to inflate away the local currency debts to reduce ex-ante payments to foreign investors, but costs of inflating away debts such as the existence of currency mismatches in domestic corporate sectors or exogenous costs of inflation work as a discipline device. This paper contributes to the literature by providing up-to-date information on the local currency debts along with related institutional details. Further, unlike preceding papers that looked at only external debts of EMEs, this paper tries to understand the states in “whole balance sheets of EMEs” including equity external liabilities and external assets. Besides the overall picture of IIPs of the EMEs, another important and interesting empirical finding uncovered in this paper is that the size of the domestic capital market seems to determine how much an EME can borrow abroad in equities 7 or LC debts: EMEs with larger stock markets can sell more equities to foreign investors, and similarly larger bond markets make it possible for an EME to sell more local currency bonds to foreign investors. Layout The rest of the paper is organized as follows. Section 2 illustrates how I constructed the dataset using different sources. Section 3 describes the evolution of external assets and liabilities of EMEs since the early 2000s. In particular, I introduce different measures of currency exposures and show how the exposures have been reduced. Section 4 describes how EMEs have increasingly borrowed abroad in equity and local currency debt, hence the original sin dissipation. I found several patterns observed in the data. Section 5 concludes. 1.2 Data In this section, I illustrate how I constructed the dataset. First, I list the sources I used for different countries and explain how I combined the different data. Then I describe the construction of the aggregate level data and the sectoral level data. 1.2.1 Sources and Strategy To construct the dataset, I choose twenty major EMEs including almost all sizable developing economies. 6 My strategy is to correct the International Investment Position (IIP) data from Inter- national Monetary Fund (IMF), using other sources that I hand-collected. The IIP dataset from IMF provides much information about external assets and liabilities at the country and the sectoral levels. What is missing in the rich dataset is the currency composition information. To fill in the gap, I used the data from the central banks or government agencies in the EMEs and the sovereign debt data from Arslanalp and Tsuda (2014). Below, I explain the steps in detail. 6 Argentina, Brazil, Bulgaria, Chile, Colombia, Czech Republic, Hungary, India, Indonesia, Korea, Malaysia, Mex- ico, Peru, Philippines, Poland, Romania, Russia, South Africa, Thailand and Turkey 8 The IIP dataset from IMF shows the composition of external assets and liabilities of each of the countries in the dataset. It classifies the assets and liabilities into FDI, portfolio equity and debts, derivatives, and other instruments, which are mostly composed of loans and deposits. Furthermore, the IIP data show the assets and liabilities of different instruments for each of the five different sectors; central bank, general government, deposit-taking financial corporations, other financial corporations and others, which are mostly nonfinancial corporations. What is missing in the IIP dataset is the currency denomination, and thus I make the following simplification. Throughout this paper, I classify all assets and liabilities into two different currencies, local (domestic) currency and foreign currency. In principle, external assets and liabilities can be de- nominated in any foreign currency, but, in reality, US dollar-denominated instruments take account of more than 70% of the external assets and liabilities and other international key currencies such Euro, Japanese yen or Swiss franc take account of most of the remaining. Furthermore, a few recent articles document the fall of Euro as a key currency and the share of US dollar in cross- border financial transactions has risen since the Global Financial Crisis. 7 In practice, in countries I can identify more detailed currency denomination information, e.g., Korea and few East European countries, US dollar takes account of most of the foreign currency-denominated external assets and liabilities. The exception is the East European countries, but the share of US dollar is still higher than 50% even in those countries. In addition, local currencies depreciate (appreciate) against all the key currencies US dollar, Euro, or Japanese yen — during almost all negative (positive) do- mestic or global events. That is, more detailed identification of the foreign currency denomination shall be of course useful, but not critical for the assessment of valuation effects due to exchange rate movements. In addition, I assume all the external assets, regardless of whether it is equity, bond, or loan, are denominated in foreign currency. Because the local currencies of the EMEs in the sample are not international currencies, it is just hard for residents in the EMEs to find their own local currency assets in international markets. In fact, domestic local currency-denominated assets account for 7 See B´ en´ etrix et al. (2020) and Maggiori et al. (2020) 9 less than 2% of the total external assets in Korea, where I can identify Korean won-denominated external assets. In the Korean data, throughout the sample period 2001-2019, more than 80% of the total external assets of portfolio equity are denominated in US dollar, Euro, Japanese yen, and Swiss franc. The share of the four key currencies is higher than 95% for the debt claims of Korea, excluding Bank of Korea international reserve, on nonresidents. Then I only need to identify local currency equities and debts and compute the amounts of foreign currency equities (in the liability side) and debts by extracting the local equities and debts from the IIP data. For the local currency equities and debts, I use the data from central banks, treasuries, and statistical authorities in the EMEs, and the sovereign debt data from Arslanalp and Tsuda (2014). Quarterly External Debt Statistics (QEDS) from World Bank provides the currency composition of external debts of many EMEs. However, the dataset provides currency compositions for only the aggregate external debts, and I need to disentangle FDI debts from the aggregate; FDI debts are mostly intercompany lending (or transactions) and therefore it is often meaningless to separately identify local currency debts and foreign currency debts. Furthermore, the imprecision or inconsistency across the countries in the dataset seems to be nonnegligible. Therefore, I try to use data from the central banks or government agencies in the EMEs as long as it is available. 1.2.2 Aggregate Level Data The dataset in this paper is constructed based on the IIP data. Thus, it is ideal to have a country- level data in which the classification is consistent with the IIP data. Central banks of Korea and Russia (Bank of Korea and Bank of Russia) provide their data almost consistent with IIP. Those data are in public. The central bank of Poland (National Bank of Poland) provided me with their data which is not in public, but almost consistent with the IIP data. The currency composition section in the IIP data naturally provides sufficiently rich information for the debt claims and liabilities, but the complementary section is filled in only for Hungary, except for Russia and Poland. 10 Figure 1.1: Local External Debts in Korea Other central banks or statistical authorities in EMEs provide relevant information in different formats. Central banks of some EMEs such as Indonesia or South Africa publish periodicals where the amounts of total local currency external debts, along with some classification of the local currency debts, are documented. Central banks in other EMEs such as Chile or Turkey provide the data of domestically issued debt securities held by nonresidents. For the remaining EMEs in the sample, available information is different among the EMEs. I summarize the information available in the appendix. To exploit the information to construct the dataset within the classification of the IIP, I used the following stylized facts about the local currency-denominated external liabilities. 1. Most of the equity liabilities in the category of portfolio investment in the IIP are domesti- cally issued equities, which are almost all local currency-denominated. 2. Most of the local currency debts, except for FDI debts, are domestically issued local currency- denominated debt securities (bonds) in the category of portfolio investment in the IIP. 3. Most of the local currency-denominated bonds held by nonresidents are the central govern- ment bonds, and central bank bonds in Asian EMEs, Korea, Malaysia, and Thailand. 11 Figure 1.2: International Investment Position and Local Currency Debt < Foreign Direct Investments > < Portfolio Investments > < Financial Derivatives > < Other Investments (Loans & Deposits> Equity Foreign Direct Investment Debt Foreign Direct Investment Equity Portfolio Investment Debt (Bond) Portfolio Investment Debt (Bond) Portfolio Investment LC FDI Debt LC Bond (Govt or CB) LC Deposit LC Debt 4. Most of the local currency external debts, except for local currency bonds and FDI debts, are the local currency deposits possessed by nonresidents. Hence, the local currency debts are classified into FDI debt, portfolio investment debt securities or other debt instruments in other investments in the classification of the IIP. The debt securities in the category of portfolio investment are the bonds issued by central governments or central banks. The local currency debts in the category of other investments in the IIP are mostly local currency deposits. This is evident in figure 1 where I plot the data of Korea. To understand the facts, note that most of the local currency external debts in EMEs, except for FDI debts whose context is much different from other loca l currency debts, are local currency bonds issued by governments or central banks. That is, foreign investors, who are likely to be residents in advanced economies, are willing to purchase local currency equities and bonds. The local currency bond purchase of the foreign investors in the bond markets in EMEs are centered on the sovereign bonds or central bank bonds whose risk and characteristics are similar to sovereign 12 bonds. On the other hand, corporations and governments in EMEs are still not able to borrow abroad in local currency debt; they cannot raise local currency loans or issue local currency bonds in international (foreign) markets. In this sense, the precise description of the original sin dissipa- tion is the participation of international investors in local currency capital markets in EMEs. The local currency equities and bonds are not exclusively issued to foreign investors, but the foreign investors participate in the capital markets in EMEs to purchase local currency equities and bonds. Local currency deposits held by foreign investors are related to the local currency equity and bond portfolio investments. Foreign investors in the capital markets in EMEs may want to dispose of their local currency assets and quit the market. Notice that the foreign investors in EMEs are exposed to exchange rate risk as their assets are denominated in local currency, not their home currency, key currencies like US dollar. Thus, the foreign investors who just sold their local cur- rency assets may not want to convert their local currency cashes to foreign currency right away, depending on the conditions in foreign exchange markets in the EME. Then the foreign investors may deposit their local currency cashes in EMEs until they will convert it to foreign currency, if it is possible or not too much cumbersome. 8 In some EMEs, for example, India, where a substantial number of citizens work abroad, local currency deposits are held by the citizens as the citizens are classified as nonresidents. Figure 2 shows how the local currency debts can be allocated according to the classification in the IIP. Using the identity, we can separate FDI local currency debts from the total local currency debts. Some central banks announce the total amounts of local currency debts along with domestic debt securities and deposits held by nonresidents. Hence, I can reasonably approximate the total local currency external debts, except for FDI debts, by adding the local currency bonds and de- posits. As a result, I can classify the local currency debts according to the categories in the IIP, if the total amounts of local currency debts are known. If the total local currency debt information is 8 The local currency deposits are observed and grow with the portfolio investments in most of the EMEs. However, relative volume of the loca l currency deposits in the Latin American EMEs, considering the foreign local currency equity and bond portfolio investments, is low compared with the Asian EMEs or East European EMEs. This may be because the statistics in the Latin American countries omit the local currency deposits of foreign investors. Or some institutional features in the area make it difficult for foreign investors to hold the deposits. 13 not available, then I can, with reasonable precision, estimate the amounts of local currency debts excluding FDI debts. Moreover, this simple identification strategy is particularly useful for some EMEs where central banks or other authorities do not publish relevant statistics. Sovereign debt dataset in Arslanalp and Tsuda (2014) 9 show the amounts of local currency bonds held by foreign investors for a number of EMEs. Moreover, QEDS shows the amounts of nonresident deposits. As a result, I can approximate the amounts of local currency external debts, even for the EMEs with no national sources. It is important to separate out FDI debts from the local currency debts as the FDI debts are mere inter-company ”transactions.” Regardless of the local currency denomination, there are no valua- tion effects from the debt, due to exchange rate movements. 10 The increase in FDI debts after the Global Financial Crisis may reflect the behaviors of global corporations to avoid taxes or manage risk following the direct investments in EMEs. But, it is unclear how it is related to the original sin dissipation as the debts are not actual debts from advanced economies to EMEs. Of course, I do not argue that FDIs are the ”green lights.” As discussed in Ostry et al. (2010), FDI direct in- vestments can contribute to large outflows and destabilize EMEs. 11 But, it is reasonable to assume that the FDI flows are more stable than the capital flows of other forms and furthermore, there are no valuation effects in the FDI debts, inter-company lending. Henceforth, the local currency debts indicate the local currency debt excluding FDI local currency debts. Discussion of nationality versus residence Some readers may question the validity of the ap- proach in this paper, using the statistics based on the classification of residence. Several recent 9 The dataset is regularly updated by the authors. The version I used is updated on October 30th, 2020 10 That could be different for reverse direct investments. For example, if a globally working Russian company issue foreign currency bonds using their funding vehicle unit in Europe and channels back the money to the mother company, then it will be included in the category debt FDI reverse investments. If Russian ruble depreciates against the foreign currency, the cost of serving the foreign currency debts will actually increase. This issue arises because the economic characteristic of the FDI debts are much closer to foreign currency debts in the category of portfolio debt securities in the classification of the IIP. However, such missed foreign currency debts might be captured in nationality based debt statistics by BIS. Below, I discuss it and show how to capture the missed foreign currency debts. 11 A different interpretation of the events reported in the paper is that the FDI debt outflows are actually the outflows driven by a form of carry trades. As I will discuss later in this paper, branches of global banks are often actively engaged in carry trades and they borrow in foreign currency from their headquarters to engage in carry trades in emerging markets. Then, the FDI debt flows are closer to local currency bond portfolio investments in terms of economic characteristic. 14 articles written by authors at BIS pointed out that the traditional classification of external assets and liabilities based on the residence of the entity may miss some important dimensions of cur- rency exposures in EMEs. To briefly describe their arguments, they showed that, after the Global Financial Crisis, some multi-national companies in EMEs have raised foreign currency debts us- ing their branches located in foreign countries, usually financial centres such as Luxembourg or Hongkong. The branches in the financial centres are often found to raise the foreign currency bor- rowings, and in this sense, Chui et al (2016) called the branches “funding vehicles.” In practice, the foreign currency debts belong to the mother companies in the EMEs. However, those debts are missed in the statistics that classify external debts based on the residence of debt issuers. Sev- eral papers (McCauley et al., 2015; Chui et al., 2016; Kuruc et al. 2017) documented the risk of underestimating foreign currency debts in EMEs in residency-based statistics. The authors of the papers insist that the foreign currency debts based on ”nationality” better capture true currency exposures in EMEs. The claims of the authors have been well accepted in the literature and BIS developed the dataset providing the foreign currency debts in both the residency-based statistics and the nationality-based statistics. It it true that the nationality-based statistics show missed currency exposure in the residency- based statistics. In a similar idea, the measures of currency exposures in this paper discard the FDI debt claims and liabilities. The problem is that the dataset in this paper is constructed based on the IIP, which is residency-based. Since the dataset from BIS does not include asset sides, there is no way for me to convert the residency-based data into nationality-based data. Instead, I conduct a simple robustness check. As I stated, BIS debt securities dataset provides the amounts of the debts in the two different classifications; residency-based and nationality-based. By adding the difference between the two different statistics to my dataset based on the IIP, I can reflect possible increases or decreases of foreign currency debts. As a robustness check, I correct my residency- based data as follows, to modify it into nationality-based data. FC Debt N i;t = FC Debt R i;t + BIS Debt Securities N i;t BIS Debt Securities R i;t 15 where ”FC” is an abbreviation of foreign currency, and N and R denote the nationality-based and the residency-based respectively. BIS debt securities data show the amounts of debt securities issued in international markets for different countries and sectors, and the different entities are classified according to nationality or residency. 12 Therefore, FC Debt R i;t denotes the EME foreign currency debt data in this paper, based on the IIP data and national sources. By adding the differ- ence between the two classifications in the BIS data, I can check how much foreign currency debt can increase once I revise mine to the nationality-based one. Again, I emphasize that this is not a precise way to resolve the discrepancy between the residency and nationality, but rather it is a simple robustness check. The contribution from BIS of course helps us better capture the nationality-based foreign cur- rency debts in EMEs, but the data from BIS may overestimate the foreign currency debts. Once we reclassify the external debts of EMEs according to nationality rather than residency, we not only have to add up the off-shore borrowings to the external debts, but we also have to deduct the substantial borrowings by branches of foreign corporations in EMEs. What significantly matters is borrowings by branches of global banks in EMEs. The debts of the branches, mostly borrowed from international interbank markets, account for substantial parts of foreign currency debts in EMEs, if it is identifiable. The foreign currency debts of global bank branches account for 60% of total foreign currency debts in Korea in 2017, and 27% in Malaysia for the same year. It is almost impossible to trace out how the branches use the borrowings. However, usual behaviors in the branches in Korea are such that those branches convert their foreign currencies to local cur- rencies in foreign exchange markets and then purchase some fixed-income securities: in fact, they are heavily engaged in carry trades. In other words, once we reclassify the borrowings by those branches according to nationality, some of the foreign currency borrowings are more similar to typical LC debts: portfolio investments in local currency-denominated bonds. 13 12 One issue in this approach is that the two different datasets are not directly comparable to each other. IIP data and other national sources I used are all measured by market values. In contrast, BIS debt securities data are based on the book value, although the data reflect the debt redemption. This difference should create some errors, which might be non-negligible, but not sizable to alter the reported trends in this paper. 13 For the analysis of the foreign bank branches in Korea, see Shin and Shin (2012). 16 Figure 1.3: Foreign Currency Borrowing and Lending Nonfinancial Corporations Financial Corporations Government Rest of the World (Nonresidents) Household Central Bank 1.2.3 Sectoral Level Data Although the aggregate foreign currency positions are in good shape, if some important sectors in the economy, such as banking sectors, have substantial currency exposures, the aggregate positions do not provide precise information about potential risks from currency mismatch. Therefore, it should be useful to examine currency exposures of the different sectors in the EMEs. To measure currency exposures of the different sectors, I need data on foreign currency lend- ing and borrowing among different sectors in each EME since the data above only includes the transactions between domestic parties and foreign parties (nonresidents). While the precise infor- mation between foreign currency transaction among different entities in EMEs is unavailable, the literature on ”dollarization” 14 has traced how much of domestic deposits and loans in EMEs are denominated in foreign currencies. Combining the collected data in the literature with mine, I can draw a ”blueprint” of currency mismatches of different sectors in EMEs. 14 See Levy-Yeyati (2006), Dalgic (2020), Bocola and Lorenzoni (2020), and Christiano et al. (2020) 17 Let’s think of an economy composed of the household sector, financial corporate sector, nonfi- nancial corporate sector, and government. It is easy to breakdown the aggregate foreign currency assets and liabilities into the four different sectors. Then, the missed foreign currency assets and liabilities (debts) of the four sectors in the aggregate level data are the foreign currency deposits of households and nonfinancial corporations in the domestic financial corporate sector and the lending from the financial corporations to the households and nonfinancial corporations. This is depicted in figure 3 above. The financial assets and liabilities with nonresidents (solid line in the figure) are captured in the IIP data and national sources, but domestic foreign currency borrowings and lending among the different sectors (green line) are not captured. The dataset in Christiano et al. (2020) filled in the missing parts. Dr. Dalgic, one of the authors in the paper, thankfully provided his data, and the data includes foreign currency-denominated domestic deposits and domestic loans in EMEs. Furthermore, the data separately shows foreign currency loans from domestic financial corporations to households and nonfinancial corporations for some EMEs. What is missing in the data is who the depositors are; the depositors can be either households or nonfinancial corporations, but I cannot see the share of each group in the foreign currency deposits. However, in the dollarization literature, it has been assumed that most foreign currency deposits belong to households since households are incentivized to hold foreign currency deposits to insure themselves from sudden local currency depreciation. 15 Hence, I simply assume that all the foreign currency deposits belong to households. Certainly, it must underestimate the foreign currency assets of nonfinancial corporations in EMEs, net foreign currency debts of the nonfinancial corporations. However, the goal here is to have a rough blueprint, not to have precise information about the sectoral currency exposures in EMEs. With the dollarization data, under the simplification, I can construct a dataset in which all the foreign currency debt flows in figure 3 are identified. 15 See Christiano et al. (2020) 18 1.3 Currency Exposures in EMEs This section describes an important observed pattern in the constructed dataset; the decline of currency exposures in EMEs. First, I introduce the findings at the aggregate level. The aggregate level data clearly shows that EMEs have reduced their currency exposures, which corresponds to a few preceding papers since Lane and Shambaugh (2010). Then, I introduce measured sectoral level currency exposures. It turns out that only nonfinancial corporate sectors in the EMEs, among the five different sectors, have significant amounts of net foreign currency debts although the results of course vary along with countries. This also corresponds to a several articles in the 2010s, which casted concerns about the rising foreign currency debts of corporations in EMEs. Although precise assessments of the risk of the nonfinancial corporate foreign currency debts are beyond the scope of this paper, I view the nonfinancal corporate foreign currency debts from a different angle: states of the EMEs where only nonfinancial corporate sectors have short positions in foreign currency indicate that foreign currency positions in EMEs have improved even at the sectoral level. 1.3.1 Aggregate Level Currency Exposures Figure 4 shows the evolution of the average net foreign currency assets in the twenty EMEs, along with foreign currency debts and assets. The left panel in the figure shows the evolution of net foreign currency assets of different definitions. I used the three different definitions of foreign currency assets, so the different definitions of net foreign currency assets. The line at the bottom shows the net foreign currency debt claims, excluding international reserves by central banks. Then I added international reserves by central banks, and equity portfolio investments, which I assumed it is foreign currency-denominated. These are the lines in the middle and at the top. Regardless of how I define the foreign currency asset, it is clear that currency exposures in EMEs have declined. 16 16 As many EMEs have become net long in foreign currency, in some sense, the EMEs face the foreign currency depreciation risk. In this paper, currency exposures are the exposures to local currency depreciation as it is often assumed in discussions of EMEs. 19 Figure 1.4: Changes in Net Foreign Currency Positions Note: 1) Simple average of the 20 EMEs For illustrative purpose, I first describe the pattern before the Global Financial Crisis and then explain the observed patterns after the crisis. After illustrating the average of the twenty EMEs, I will briefly describe the different evolution across the EMEs. From 2001 to 2007, the beginning of the Global Financial Crisis, significant improvement in the net foreign currency positions are clear and the improvements are mostly due to the decreases in foreign currency debts and the international reserve accumulation by central banks. The decreases in the foreign currency debts are attributed to improvements in current accounts of the countries and also to large FDI and equity portfolio capital inflows during the periods, as documented in Lane and Milesi-Ferretti (2007a); the inflows of FDI and equity portfolio investments substituted for foreign currency debt inflows. The international reserve accumulation during the periods accounts for most of the increases in foreign currency assets. While the literature has not formed a consensus about whether (or how much) the international reserve accumulation leads to a higher net foreign asset position, 17 a few recent influential studies (e.g., Gabaix and Maggiori, 2015; Fanelli and Straub, 17 Several early studies of international reserve accumulation in the early 2000s (e.g., Rodrik, 2006; and Alfaro and Kanczuk, 2009) doubt casts on role of reserve accumulation against sudden stops. Other studies following the papers (Aizenman and Lee, 2007; and Jeanne and Ranci` ere, 2011) suggested theoretical models where international reserve accumulation does a role of preventing sudden stops. After the Global Financial Crisis, there have been different empirical studies of the effects of international reserve accumulation on the financial and economic stability of EMEs during the crisis. While different studies reported different results, a few influential studies such as Bussiere et al. (2015) showed that the countries with more reserves before the crisis weather the crisis better than other countries. 20 2020) suggest that reserve accumulation, i.e., foreign exchange market intervention, improves the net foreign asset position. The net foreign currency assets of the EMEs have kept increasing after the Global Financial Crisis, but the pace has been slower than before the crisis. A noteworthy difference from the pre- Global Financial Crisis is that the improvement is mostly attributed to the foreign currency asset accumulation by private sectors. The levels of foreign currency debts and international reserves of the EMEs have been stable since 2011, but the foreign currency debt claims and portfolio equity assets, mostly held by private sectors in EMEs, have notably risen. 18 This is important in the sense that 1) the increase of the foreign currency asset accumulation of private sectors in EMEs have not received much attention and therefore it has not been extensively studied, and 2) the capacity of private entities to protect themselves against the exchange rate risk should be more desirable than reliance on the international reserves of central banks. It is well-known that central banks are often hesitant to deplete their international reserves during sudden stops, possibly due to concerns about the moral hazard or heightened uncertainty. 19 Having private corporations and individuals hedge their own exchange rate risk is an important improvement regardless of the improvement in the aggregate positions. Figure 4 and the discussions above are all about the average of the twenty different EMEs. Of course, there is a great heterogeneity along with countries and regions. Table 1 shows the different evolution in the EMEs. Table 1 confirms that the patterns observed in the average are not driven by a small number of countries. Most of the countries were ”net short” in foreign currency in 2001, but now are ”net long” in foreign currency. The improvement in the net foreign currency positions is sizable, except for Romania and Turkey. The net foreign currency debt of Romania was less than Papers in different strands of the literature have built quantitative models to understand the role of international reserves against sudden stops or sovereign default crises (e.g., Bianchi et al., 2018; Shousha, 2019; Arce et al. 2020; and Samano, 2020). The literature oveall agrees that the reserve accumulation is an effective tool against sudden stops or sovereign default crisis, but different papers are still seeking exact mechanisms o how the reserve accumulation has an effect or has been chosed over other policy tools, e.g., capital controls. 18 The foreign currency debt claims and portfolio equity assets include the assets of sovereign wealth funds and public pension funds, who shall be included in the public sectors rather than private sectors. However, exclusion of the state-owned financial institutions hardly convert the trends. 19 See Aizenman and Sun (2012) and Shousha (2019) 21 5% of GDP even in 2001. Turkey seems to be the only EME in the sample, which corresponds to the old stereotype of EMEs, large amounts of net foreign currency debts over 30% of GDP. Note that, in the table, I did not include portfolio equity assets, which have significantly increased since the early 2010s and are mostly foreign currency-denominated liquid assets. 20 Once I include the equities in the foreign currency assets, the improvements in the foreign currency positions are even more dramatic. Only Turkey remains as a net borrower in foreign currency and twelve EMEs are net long in foreign currency, even without central bank international reserves if I count the foreign currency equity assets. The portfolio equity assets in 2019 are particularly sizable in Chile, Korea, and South Africa. In table 1, I also included the modified foreign currency debts, the nationality-based foreign currency debt, computed by the way that I described above. The difference between the residency- based and the nationality-based was negligible in 2001, but the difference between the two dif- ferent classifications is not negligible any more in some EMEs in 2019. Foreign currency debts in Brazil, Malaysia, Russia, and South Africa in 2019 considerably increase when computed in the nationality-based classification. The increments are 5 - 10% of GDP. However, even after adding the foreign currency debts, we can still see the improvement in the four EMEs. In Brazil, Malaysia, and Russia, the increased net foreign currency assets in the residence-based classifica- tion are around 20% of GDP, which are far larger than the increments. In South Africa, the foreign currency asset accumulation of the country over the last decade has been done mostly through the portfolio equities, and once I calculate the foreign currency assets of South Africa, including the portfolio equities, the improvement in the foreign currency position is clear. 1.3.2 Sectoral Level Currency Exposures The constructed foreign currency exposures of the five sectors in EMEs clearly show who is short or long in foreign currency in EMEs. As shown in figure 5, households and central banks in EMEs have positive net foreign currency assets (net long in foreign currency). Financial corporate 20 In contrast, most of the portfolio equity liabilities are mostly local currency-denominated. 22 Table 1.1: Foreign Currency Positions of EMEs FC Debt Claims 2) FC Debt Liab. Net FC Position 3) 2001 2019 2001 2019 2001 2019 Argentina 36.8 ( 5.0) 70.5 (10.1) 35.7 [34.6] 43.4 [43.0] 1.1 ( -3.9) 27.1 ( 17.1) Brazil 9.8 ( 6.4) 23.6 (19.4) 37.6 [38.3] 17.5 [25.9] -27.7 (-34.1) 6.0 (-13.4) Bulgaria 63.8 (25.2) 77.7 (41.1) 71.5 [71.5] 40.5 [40.5] -7.7 (-32.9) 37.2 ( -3.9) Chile 38.4 (20.3) 38.1 (14.4) 46.2 [42.9] 44.4 [42.7] -7.8 (-28.1) -6.2 (-20.6) Columbia 27.6 (10.2) 37.5 (16.3) 39.9 [40.1] 37.1 [38.0] -12.3 (-22.5) 0.4 (-15.9) Czech Rep. 57.7 (21.3) 85.5 (59.8) 17.4 [16.3] 30.8 [30.4] 40.4 ( 19.0) 54.7 ( -5.1) Hungary 37.9 (20.1) 45.9 (19.8) 38.1 [37.9] 36.0 [36.3] -0.2 (-20.3) 10.0 ( -9.8) India 14.2 (11.2) 18.0 (16.1) 17.8 [18.4] 12.0 [13.7] -3.6 (-14.8) 6.0 (-10.1) Indonesia 24.8 (16.1) 24.8 (11.5) 77.5 [81.4] 24.4 [25.5] -52.7 (-68.8) 0.4 (-11.1) Korea 28.3 (18.8) 53.5 (24.8) 19.3 [19.7] 18.3 [18.7] 9.0 ( -9.7) 35.2 ( 10.4) Malaysia 49.7 (29.5) 56.9 (28.4) 47.5 [48.6] 35.3 [40.2] 2.2 (-27.3) 21.6 ( -6.8) Mexico 14.1 ( 5.9) 26.8 (14.5) 20.8 [21.6] 29.3 [29.1] -6.7 (-12.6) -2.5 (-17.0) Peru 23.1 (17.3) 35.8 (29.6) 53.5 [53.3] 27.4 [24.3] -30.4 (-47.7) 8.4 (-21.2) Philippines 35.5 (19.9) 36.0 (23.3) 67.6 [69.3] 21.9 [22.7] -32.0 (-51.9) 14.1 ( -9.3) Poland 25.3 (13.9) 32.4 (21.7) 22.5 [25.4] 28.0 [28.3 2.7 (-11.2) 4.5 (-17.2) Romania 26.4 (12.0) )28.5 (16.8) 29.8 [29.1] 29.3 [30.4] -3.3 (-15.4) -0.8 (-17.6) Russia 67.0 (11.1) 58.2 (32.6) 38.9 [39.4] 13.9 [18.7 ] 28.1 ( 16.9) 44.3 ( 11.8) South Africa 15.0 ( 6.1) 29.7 (15.7) 18.9 [21.8] 28.2 [39.2] -3.9 (-10.1) 1.6 (-14.1) Thailand 42.8 (27.5) 65.7 (41.3) 49.5 [50.3] 20.8 [22.5] -6.7 (-34.2) 44.9 ( 3.6) Turkey 23.2 ( 9.9) 26.3 (13.9) 51.7 [52.6] 50.4 [49.6] -28.4 (-38.3) -24.1 (-38.0) Note: 1) All the numbers are are % of GDP. 2) Numbers in parentheses are central bank official reserves to GDP ratios. 3) Numbers in parenthesis are private sectors net position (excluding official reserves). 23 sectors have balanced foreign currency assets and debts (square position in foreign currency), but nonfinancial corporate sectors have sizable net foreign currency debts (net short in foreign currency). Governments in some of the sample EMEs have nonnegligible foreign currency debts, but it is still much lower than the historical averages. The most important feature uncovered from the sector level data is that the currency mismatches in EMEs, measured in the levels of net foreign currency debts, are centered on nonfinancial corporations. As expected, households and central banks are net long in foreign currency as the households and the central banks hold foreign currency deposits and international reserves respectively. Gov- ernments in the EMEs have net foreign currency debts varying from nearly zero to around 15% of GDP, 21 except for Argentina whose foreign currency government external debt to GDP ra- tio jumped in 2018 due to large Argentine peso depreciation following the sovereign debt crisis. While the foreign currency debts will be burdensome for the governments when the local curren- cies depreciate, but the level of the foreign currency debts are much lower than in the 1990s or the early 2000s when the governments could not sell local currency government bonds to foreign investors. 22 Accurate assessment of the underlying risk of foreign currency debts of the EMEs is beyond the scope of this paper, but it is evident that the governments now face lower exchange rate risk than in the 1990s or the early 2000s. Below, I focus on the discussion of financial and nonfinancial corporate sectors. Financial Corporate Sectors The net foreign currency debts of the financial corporate sectors in the figure 1, measured as the ratios of GDP, look relatively small. The financial corporate sectors borrow in foreign currency from both foreign parties and domestic households, but they also lend in foreign currency and invest in foreign currency assets. As a result, the foreign currency assets of the financial corporate sectors, even without equity assets, are sizable enough to protect them from 21 Positive net foreign currency positions of general governments in Chile and Korea are mainly due to government- owned financial institutions, in particular sovereign wealth funds. 22 For the currency composition of government debts in EMEs and relevant information such as the debt holders, see Arslanalp and Tsuda (2014). Other relevant papers are Du and Schreger (2016) and Ottonello and Perez (2019) 24 Figure 1.5: Sector Level Currency Mismatches Households Non-Financial Corporate Sectors Financial Corporate Sectors Government Note: 1) All data is as of the 2018. 2) Net foreign currency debts: foreign currency debt instrument assets minus foreign currency debts, and hence negative numbers are positive net debts. 3) In the figure of Non-Financial Corporate Sectors, the orange bars indicate domestic foreign currency loans and the blue bars indicate net foreign currency debts from financial contracts with only non-residents. 4) The data of foreign currency deposits and loans are missing in my data, and therefore I let the foreign currency deposits and loans in Brazil be zero. 25 the exchange rate risk. In other words, the risk from foreign currency debts is hedged by their own foreign currency debt claims. However, the GDP ratio might not be an ideal measure of the currency exposure of the financial corporate sectors. While I can take different measures, I take the net foreign currency debt to the capital (equity) ratios. How much the exchange rate risk matters for a financial institution would be determined by the factors, 1) share of the net foreign currency debt in the total asset and 2) the leverage (total asset to the capital ratio). The net foreign currency debt to the capital ratios summarize the two factors. Also, I divide the financial corporate sectors into two different groups; deposit-taking financial corporations (banks) and other financial corporations including insurance companies, mutual funds and pension funds. Figure 6 shows the net foreign currency debt to capital ratios for the two groups of financial corporations in EMEs in 2018. 23 The compute the net foreign currency debt to the capital ratios show a little different picture from figure 5. Although the net foreign currency debt to the GDP ratios are low, the net foreign currency debt to the capital ratios significantly vary along with coun- tries because the sizes of the banking sectors and other financial sectors are very much different among the EMEs. The net short foreign currency position to the capital ratios are above 30% for some EMEs; Argentina, Brazil, Chile, India, Peru, Philippines, and Romania, and above 50% for a smaller set of EMEs; Argentina, Brazil, Peru and Philippines. In Peru, the short position to the capital ratio is higher than 100%. The sectoral level backs to only 2010 and thus it is hard to find its historical benchmark. However, this ratios are still lower than foreign currency debt to the capital (net worth) ratios used as target moments in many recent quantitative studies. The net foreign 23 I used financial development dataset by World Bank. The dataset provides the bank capital to GDP ratios for the EMEs. One issue is that there might be discrepancy between the coverage of the deposit taking financial corporations in the IIP and the coverage of banks in the world bank data. However, errors due to the different definitions of the banks must be small in practice, considering the dominance deposit taking banks in financial system in most of countries in the world. The World Bank dataset also shows the total assets of insurance companies, mutual funds and pension funds. Two issues I have are I do not have any information of leverage ratios of the financial sectors and have no clue of how the different sectors are interconnected. For the convenience, I just summed the assets of the three financial sectors, and then I assumed that the leverage ratio is 5 as such non-banking financial corporations take lower leverage ratios as they take higher risk in their investments. 26 currency positions to the capital ratios are often over 100% in many quantitative studies, but only Peru in the sample EMEs has such high currency exposures in the banking sectors. Also, the ratios are note enough to generate the falls in bank capitals as large as typically observed in a banking crisis. Baron and Xiong (2017) studied the history of banking crises and defined a banking crisis as an event where the bank capitals fall by more than 30%. During a currency crisis, e.g., East Asian crisis, local currency depreciated by 70-80%. If the depreciation is 75%, we need net foreign currency debts to be more than 40% of the capital. The net foreign currency debt to the capital ratios are over 40% only in Argentina, Brazil, Peru nad Philippines. Another noteworthy observation is that other financial corporations, financial corporations other than deposit-taking banks, tend to have long positions (or larger long positions) compared with the deposit-taking financial corporations. If I add the portfolio equities to the foreign currency assets, the foreign currency net long positions of the financial corporations will be even larger. In Chile, Colombia, Peru, Philippines, and Russia, deposit-taking financial corporations are net short in foreign currency, but other financial corporations are net long in foreign currency. In those EMEs, local currency deprecation will cause opposite valuation effects to the two different groups of financial sectors. In EMEs like Chile, the net long positions of nonbanking financial corpora- tions must help the Chilean economy when the banks are forced to deleverage due to loss from her own net short positions in foreign currency. Depending on the institutional details, insurance companies or mutual funds with stronger balance sheets may replace the sluggish intermediation by the banks or support the banks by providing more funds to the interbank markets. Another realistic scenario is that the banks may own the insurance companies or mutual funds. In such a case, the banks directly benefit from her subsidiaries so that the banks can cover the losses from the foreign currency debts. Nonfinancial Corporate Sectors Similarly with the financial sectors, I computed the net for- eign currency position to the capital ratios for the nonfinancial corporate sectors in the EMEs. I estimated the capitals of the nonfinancial corporate sectors, using the corporate debt database in 27 Figure 1.6: Net Foreign Currency Position to Capital Ratios Deposit Taking Financial Sector Nonfinancial Corporate Sector Other Financial Sector -150 -100 -50 0 50 100 150 Argentina Brazil Bulgaria Chile Colombia Czech Republic Hungary India Indonesia Korea Malaysia Mexico Peru Philippines Poland Romania Russia South Africa Thailand Turkey (% of Capital) -150 -100 -50 0 50 100 150 Argentina Brazil Bulgaria Chile Colombia Czech Republic Hungary India Indonesia Korea Malaysia Mexico Peru Philippines Poland Romania Russia South Africa Thailand Turkey (% of Capital) -80 -60 -40 -20 0 20 Argentina Brazil Bulgaria Chile Colombia Czech Republic Hungary India Indonesia Korea Malaysia Mexico Peru Philippines Poland Romania Russia South Africa Thailand Turkey (% of Capital) 28 Mbaye et al. (2018) and the average corporate leverages in Beltran et al. (2017). 24 Figure 6 shows the computed net foreign currency position to the capital ratios. Despite the sizable foreign cur- rency debts of the nonfinancial corporations, the net foreign currency short positions are below 30% of the capital in most of the countries. This is due to the leverage ratios of nonfinancial corpo- rations that tend to be low in most of the countries in the world. Then, as discussed in Christiano et al. (2020), the nonfinancial corporations better absorb negative valuation effects from the foreign currency debts and it is illustrated in the relatively low ratios in figure 6. More importantly, much of the revenues of the nonfinancial corporations in the EMEs are in fact foreign currency-denominated. Many of the corporations are exporters and the current consensus about the export pricing in the literature is the ”Dominant Currency Pricing (DCP)” hypothesis. According to the DCP hypothesis, most of the tradable goods are de facto denominated in the key currency, US dollar. Then, certain shares of the revenues of exporters in EMEs are foreign currency-denominated, and the revenues from exports computed in local currency increase in the local currency depreciation against the foreign currency. In contrast, much of the costs, for example, wages, must be local currency-denominated. 25 Given the foreign currency denomination in exports, as one can easily see, the local currency depreciation actually raises the mark-up on the exporting goods, which eventually boosts the profitability of the exporters. In other words, domestic local currency depreciation benefits many of the nonfinancial corporations in EMEs, net exporters. In a companion paper Han (2021), I build a small open economy model to formalize the idea. Auclert et al. (2021) separately developed a similar idea in their paper. To sum up, despite the net foreign currency debts of nonfinancial corporations in the EMEs, how risky the foreign currency debts are is unclear. The nonfinancial corporations have low lever- ages than financial corporations and they can generate foreign currency revenues by themselves. 24 The corporate debts in Mbaye et al. (2018) significantly vary along with countries and the level of corporate debts are quite low; even lower than 20% of GDP in Argentina and Romania. This reflects the large informal sectors in EMEs; corporations borrow informally or rely on net worth 25 Those exporter must import intermediate goods to produce the exporting goods. However, considering consid- erable shares of imports are consumption goods, exports from the corporations must be more than the intermediate goods import. 29 More importantly, under the DCP hypothesis, the local currency depreciation itself is likely to raise the profitability of exporters in EMEs. Because of the facts and findings from the recent consensus in the trade literature, foreign currency debts of nonfinancial corporate sectors in EMEs are probably not great threats to the financial stability of those countries. Furthermore, it is clear that the nonfinancial corporations are the sectors least fragile to local currency depreciation and in some sense, the foreign currency net short positions of the nonfinancial corporations seem to be efficient. Therefore, considering that governments and financial corporations also took large net short positions in foreign currency in the past, current states in EMEs where only nonfinancal corporate sectors are net short in foreign currency show the improvement of EMEs in terms of the sectoral level foreign currency positions. 1.4 Original Sin Dissipation In this section, I illustrate how the original sin has been dissipated since the early 2000s in different EMEs. I focus on when and which EMEs began borrowing abroad in local currency equities and debts, and how the local currency borrowings have increased. In particular, I try to find economic fundamentals which are highly correlated with the local currency borrowings. For example, EMEs with low and stable inflation could have borrowed abroad more than other EMEs? or higher fi- nancial openness makes some EMEs able to borrow abroad more in their own local currency? It turns out that the most tightly correlated variable with the local currency borrowings is the capital market depth, stock market and bond market capitalization to GDP ratios. Key trends in the original sin dissipation EMEs began receiving portfolio equity investments into their stock markets from the 1990s, as documented in Lane and Milesi-Ferretti (2001, 2007a). However, in the very early 2000s, the amounts of equity portfolio investments were still less than 5% of GDP in most of the EMEs. The local currency external debts in the very early 2000s were even more scarce than the local currency equities. 30 Figure 1.7: Local Currency External Liabilities 0 2 4 6 8 10 12 14 2001 2004 2007 2010 2013 2016 2019 LC Equity LC Debt % of GDP Note: 1) Simple average of the twenty EMEs 2) Both the local currency equity liabilities and local currency debt liabilities exclude FDI equities and debts. 31 The local currency equities and debts held by nonresidents (mostly foreign investors) have steadily increased from 2000 to 2007, just before the Global Financial Crisis, but the inflows into local currency equity and bond markets were particularly large in 2006 and 2007. Amounts of the equities held by foreign portfolio investors rapidly increased during the two years. The increases were in part due to the global stock market booms at that time, but also due to large purchases of the stocks by foreign investors. The large inflows into the bond markets in some EMEs at that time are also observable. Actually, 2006 and 2007, the two years was the time when local currency bonds in EMEs began being sold to investors from advanced economies. It is not clearly observable in the average as many EMEs were still not able to attract foreign investors into their local currency bond markets, but the bond inflows were large in Brazil, Korea and Malaysia. During the Global Financial Crisis in 2008, there were obviously large outflows from EMEs, which significantly decreased the local currency equity liabilities and local currency debts. How- ever, from 2010 until the early-2010s, the external liabilities of local currency equity and debt in the EMEs had rapidly increased. Both the local equity and local currency debt had increased, but the increases in the local currency debts were larger in terms of growth rate. In particular, international investors began purchasing local currency bonds in the EMEs that they had not been significantly holding in the 2000s; India, Indonesia, Peru, Romania, and Thailand. The large inflows into stock and bond markets in the EMEs were probably fueled by the near- zero interest rates and quantitative easing in major economies such as US or Eurozone. During the late-2010s when Fed was gradually normalizing its policy, the overall growth of the external liabilities of local currency equities and debts in the EMEs became sluggish. However, there were no considerable outflows from the stock and bond markets in the EMEs: although Fed raised its policy rate and rolled back the quantitative easing, there was no notable regress in the original sin dissipation. Depending on the economic and political conditions (e.g., Brazil) or the amounts of inflows in the late and mid-2010s (e.g., Hungary and South Africa), there were moderate amounts of outflows in the bond markets, but there were more capital inflows into local currency bond markets in other EMEs during the same periods; e.g., Chile, India, and Indonesia. 32 Finally, I summarize the key trends in the original sin dissipation as follows 1. In 2005-07, there was a beginning of the original sin dissipation: large inflows into local currency stock markets in many of the EMEs and meaningful inflows into local currency bond markets in a few EMEs 2. In 2010-12, there was a surge of capital inflows into the local currency stock and bond markets, although amounts of the inflows and relative weights between the local currency equity inflows and local currency bond inflows vary among the EMEs. 3. In 2016-19, amid the monetary policy normalization by Fed, there was a slowdown of the inflows, but there were no large outflows from the stock and bond markets, equivalently no significant regress in the original sin dissipation Original sin dissipation and capital market developments An important question related to the findings illustrated so far in this paper is ”What is behind the original sin dissipation?” Several preceding papers have touched on this question. Chang and Wei (2020) argued that countries with better institutional quality can attract more equity investments into the countries. Ottonello and Perez (2019) suggested that better management of domestic inflation, such as inflation targeting, might be behind the surge of local currency sovereign debts. Du et al. (2016) provided empirical and theoretical results of similar insights. I suggest a different view on the original sin dissipation. However, I do not formalize my idea. I simply examine different economic fundamental variables and pick up the variable that is most correlated with the local currency borrowings since the early-2000s in different EMEs. Then, I discuss my empirical findings and suggest an interpretation to the results, but the interpretation should not be understood as a formal claim. I computed correlations between different economic fundamental variables and the local cur- rency borrowings. I examined different variables, level of inflation, inflation volatility, GDP per capita, institutional quality index, financial openness, and trade openness. Among the different variables, the most correlated variable with the local currency borrowing is the development of 33 capital markets, the stock and bond markets. 26 That is, an EME with larger stock markets relative to the size of the economy, EME of a higher stock market capitalization to GDP ratio, seems to borrow abroad more in local currency equity by attracting more foreign investors into the market. Similarly, an EME of higher bond market capitalization to GDP ratio seems to borrow abroad more in local currency debt by attracting more foreign investors into the bond market. Figure 8 below shows the high correlations between the market sizes and the original sin dissipation, local currency borrowings. The correlation between the stock market capitalization to GDP and portfolio equity investment to GDP is 0.82 and none of the other variables is even close. Hence, the shares of non-resident portfolio investors in stock markets in the EMEs have been mostly in the range between 0.15-0.25 across the time and the countries. The bond market capitalization is a little hard to find a proper measure. As it is in many advanced economies, much of the corporate bond tradings are probably done in over-the-counter markets in the EMEs. Then, the sizes of the floor bond markets will be much proportional to the sizes of public debt markets. Hence, I used the outstanding public debt- securities to GDP ratios, 27 as a proxy for the bond market capitalization. The correlation for the bond market capitalization is 0.61, which is a little lower than in portfolio equity investments, but much higher than other variables. Again, these are mere correlations and I do not try to draw a firm causality from the observed correlations. I just suggest a plausible interpretation of the correlations. As I described, it is in the early 2000s that EMEs began attracting substantial portfolio investments into their domestic stock and bond markets where the securities are traded in local currency. Around the early 2000s, there were several important changes. Some EMEs that had suffered from sudden stops in the 1990s, for example, Korea and Indonesia, opened their capital markets from the early-2000s. A few Latin American countries ended hyperinflation in the mid or late-1990s and adopted inflation targeting, which has stabilized the inflation in those countries. Furthermore, progress in infor- mation technology in the early-2000s, wider use of the internet in developing countries, makes it 26 Here, the development mere refers the size of the market capitalizations 27 I used Arslanalp and Tsuda (2014) and added outstanding amounts of central bank debt securities, if necessary. 34 Figure 1.8: Capital Markets and Local Currency Borrowing 0 10 20 30 40 50 0 50 100 150 200 250 300 LC Portfolio Equity Liabilities / GDP (%) Stock Mkt Cap / GDP (%) Corr = 0.82, R² = 0.69 0 3 6 9 12 15 0 10 20 30 40 50 LC Bonds / GDP (%) Public Debt Mkt Cap / GDP (%) Corr = 0.61, R² = 0.34 Note: Each bin is 5 years avearge of each of the 20 EMEs. Hence, the window is either of 2001-05, 2006-10, 2011-15, or 2016-19 for equity and 2004-09, 20010-14, or 2015-19 for bond 35 easy for international investors to enter and exit the stock and bond markets in EMEs. Note that international investors’ portfolio investments in local currency equity and bond markets are the participation of the investors in the local markets. 28 To quickly change positions depending on the surrounding conditions, the investors need to enter and exit the markets whenever they want. To make all the process in the investments electronic should be crucial for the investors to be able to quickly dispose of their positions in the local markets. All the changes together incentivized international investors to invest in stock and bond markets in EMEs. However, in the early-2000s, there were not enough local currency stocks and bonds in EMEs for the investors to buy. When the stock and bond markets were really small and therefore the liquidity in the markets was really limited, the local markets could not accommodate many enough international investors. The share of international investors in domestic capital markets cannot increase indefinitely and might have some limits. As the stock and bond markets in EMEs have grown over time, more stocks and bonds in local markets have become available for the in- vestors and the ability of the markets to accommodate international investors has improved. Then, more international investors naturally join the local markets to enjoy higher expected returns in the markets. 29 Put differently, the changes around the early-2000s I listed above could be substantial enough to attract international investors into the local currency stock and bond markets. However, there were just not enough stocks and bonds in EMEs for the investors to purchase. As more and more stocks and bonds in the local markets in EMEs have become available, the international investors have been willing to hold more local currency stocks and bonds in EMEs. In a companion paper 28 Hence, the local currency equities and bonds in EMEs are mostly not issued in foreign markets and not exclusively to international investors. Rather, the securities are issued in domestic capital markets, and the international investors “come” to the markets to buy the local currency-denominated securities so that the international investors are another type of investors in the domestic capital markets along with domestic investors. Thus, the issuers of the equities and bonds do not have controls over the buyers: rather, the shares of international investors are determined by the investors, not by the domestic issuers. Such obvious institutional features are often overlooked, but it is important to build a correct model. 29 The positive linear relationship between the market growth and foreign investment in local currency bond invest- ment is not as clear as the portfolio investment in local currency equities. Possibly, this relfects the another apsect of the bond market development other than the depth of the markets. For example, the existence of the bonds of different maturities should matter for attracting more international investors into the domestic local currency bond markets. 36 Han (2021), I formalize the idea in a simple model, although it is not the main research agenda in the paper. 1.5 Concluding Remarks This paper provides another source for a researcher to see the evolution of currency exposures and the composition of external liabilities in EMEs. The information of the external liabilities in local currency is mainly available through the national sources that I hand-collected. I combined the different sources using my understanding of institutional details of the local currency equities and debts in different EMEs. Furthermore, I also added the dollarization data to my data so that I can reasonably approximate the currency exposures at the sectoral level. Throughout the data con- struction in this paper, I relied on the actual data from different central banks and other authorities in the EMEs. Although the information from the different sources is not so complete, the use of actual data from the national sources greatly improves the precision of the dataset. I focus on the two aspects of the information in the constructed dataset. First, I traced out how currency exposures in the EMEs have evolved since the early 2000s. Like a few preceding and contemporaneous papers (B´ en´ etrix et al., 2015; B´ en´ etrix et al., 2020), I concluded that currency exposures in the EMEs overall have greatly reduced for the last two decades. In the 2000s before the Global Financial Crisis, large FDI and portfolio inflow into the EMEs provided EMEs with sources to save in foreign currency. In the early and mid-2000s, there was a great reduction in the foreign currency debts of the countries and central banks in the EMEs had aggressively increased their international reserve holdings. After the Global Financial Crisis, there were other large port- folio inflows into the EMEs, which were probably fueled by the moneteary policy easing in large advanced economies, in particular the US. Unlike the pre-Global Financial Crisis, there was no reduction in foreign currency debts or international reserve accumulation of the central banks, but private sectors in the EMEs, financial and nonfinancial corporations, have accumulated foreign currency assets in both equities and debt clams. As a result, the foreign currency positions in the 37 EMEs have kept improving even in the 2010s, although the speed of the improvements has become slower than in the 2000s. I also confirm that the EMEs are currently in a good shape in terms of the ”distribution” of currency exposures across different sectors. Financial sectors in the EMEs are balanced between foreign currency debts and foreign currency assets. The financial corporations in the EMEs borrow in foreign currency from foreign parties or households, but they also invest abroad in foreign currency or make foreign currency loans to domestic companies. As a result, the financial corporate sectors have net long positions in foreign currency or the amounts of net foreign currency debts are manageable considering the net foreign currency debt to the capital ratios, except for a few EMEs where the banking sectors have relatively high net foreign currency debt to the capital ratios, e.g., Brazil and Peru. The nonfinancial corporate sectors have in general have net foreign currency debts by considerable amounts. But, the currency exposures in the nonfinancial corporate sectors are limited because the leverages in the sectors are usually low and many of the corporations are exporters; some forefront studies in the literature (Han, 2021; Auclert et al, 2021) pointed out that, under the DCP hypothesis, local currency depreciation itself leads to higher mark-up on the exports, thereby raising the profitability of the exporters. The higher profitability will offset the higher cost of serving foreign currency debts of the corporations. Lastly, I illustrated how the EMEs have increasingly borrowed abroad in local currency equity and debts, i.e., how original sin in the EMEs has been dissipated. I documented several patterns in the dissipation commonly observed across the countries. The most interesting and important em- pirical regularity regarding the original sin dissipation is the highly positive correlations between the market capitalization to GDP ratios and the external liabilities of local currency equities and bonds. Deep investigations into the correlations are beyond the scope of this paper, but I provide a plausible interpretation of the correlations. The important changes occurred in the EMEs around 2000, such as capital account liberalization or progress in information technology, probably incen- tivized international investors to put their money in the local currency stock and bond markets in the EMEs. However, the capital markets in the EMEs didn’t have enough depth at that time and 38 thus the inflows into the local markets were limited. However, as the capital markets have grown over time, deepening the markets, the international investors have become willing to invest more in the markets: larger capital markets have attracted more foreign capitals into the local currency stock and bond markets in the EMEs. 39 References [1] Aizenman, J. and Sun, Y ., 2012. “The financial crisis and sizable international reserves de- pletion: From ‘fear of floating’ to the ‘fear of losing international reserves’?,” International Review of Economics Finance, 24 (C): 250–269. [2] Arslanalp, S. and Tsuda, T. 2014. “Tracking Global Demand for Emerging Market Sovereign Debt,” IMF Working Paper, WP/14/39. [3] Akinci, O. and Queralto, A. 2019. “Balance Sheets, Exchange Rates, and International Mon- etary Spillovers.” FRB of New York Staff Report, No. 849. [4] Aoki, K., G. Beningno, and N. Kiyotaki. 2018. “Monetary and Financial Policies in Emerging Markets.” Manuscript, Princeton University. [5] Auclert, A., Rognlie, M., Souchier, M., and Straub, L. 2021. “Exchange Rates and Monetary Policy with Heterogeneous Agents: Resurrecting the Real Income Channel.” Manuscript, Stanford University. [6] Beltran, D., Garud, K., and Rosenblum, A. 2017. “Emerging Market Nonfinancial Corporate Debt: How Concerned Should We Be?,“ IFDP Notes. Washington: Board of Governors of the Federal Reserve System, June 2017, https://doi.org/10.17016/2573-2129.32 [7] B´ en´ etrix, A. S., Lane, P. R., and Shambaugh, J. C. 2015. “International currency exposures, valuation effects and the global financial crisis.” Journal of International Economics, 96 (S1): S98-S109. [8] B´ en´ etrix, A. S., D. Gautam, L. Juvenal, and M. Schmitz. 2020. “Cross-border currency ex- posures: new evidence based on an enhanced and updated dataset,” ECB Working Paper, No. 2417. [9] Bocola, Luigi, and Guido Lorenzoni. 2020. “Financial Crises, Dollarization, and Lending of Last Resort in Open Economies.” American Economic Review, 110 (8): 2524-57. [10] Bruno, V . and Shin, H. S. 2018. “Currency Depreciation and Emerging Market Corporate Distress.” BIS Working Papers, No. 753. [11] Burger, J. D. and Warnock, F. E. 2006. “Local Currency Bond Markets.” NBER Working Paper, No. 12552. 40 [12] Calvo. G. 1998. “Capital Flows and Capital-market Crises: the Simple Economics of Sudden Stops.” Journal of Applied Economics, 1 (1): 33–54. [13] Calvo, G., Izquierdo, A., and Talvi, E. 2006. “Phoenix Miracles in Emerging Markets: Recov- ering Without Credit From Systemic Financial Crises.” Inter-American Development Bank Research Department Working Paper, No. 570. [14] Chang, M. and Wei, S-J. 2020. “International Equity and Debt Flows to Emerging Market Economies: Composition, Crises, and Controls.” Manuscript, Fudan University [15] Christiano, L., Dalgic, H. C. and Nurbekyan, A. 2020. “Financial Dollarization in Emerging Markets: Efficient Risk Sharing or Prescription for Disaster?” Manuscript, Northwestern University. [16] Chui, M. KF, Kuruc, E., and Turner, P. 2016. “A New Dimension to Currency Mismatches in the Emerging Markets: Non-financial Companies.” BIS Working Papers, No. 550. [17] Dalgic, H. C., 2020. “Corporate Dollar Debt in Emerging Markets.” Manuscript, University of Mannheim. [18] Du, W. and Schreger, J. 2016. “Local Currency Sovereign Risk.” Journal of Finance, 71 (3): 1027–1070. [19] Du, W. and Schreger, J. 2016. “Sovereign Risk, Currency Risk, and Corporate Balance Sheets.” Manuscript, Columbia Business School. [20] Du, W., Pflueger, C. E., and Schreger, J. 2016. “Sovereign Debt Portfolios, Bond Risks, and the Credibility of Monetary Policy.” NBER Working Paper, No. 22592. [21] Eichengreen, B. and Hausmann, R. 1999. “Exchange Rates and Financial Fragility,” NBER Working Paper, No. 7418 [22] Eichengreen, B., Hausmann, R., and Panizza, U. 2002. “Original Sin: the Pain, the Mystery, and the Road to Redemption.” Inter-American Development Bank Conference Paper. [23] Engel, C. and Park, J. J. 2018. “Debauchery and Original Sin: The Currency Composition of Sovereign Debt.” NBER Working Paper, No. 24671. [24] Fanelli, S. 2019. “Monetary Policy Capital Controls, and International Portfolios.” Manuscript, Princeton University [25] Gopinath, G. and Stein, J. C. 2020. “Banking, Trade, and the Making of a Dominant Cur- rency,” Quarterly Journal of Economics, qjaa036. [26] Lane, Philip R., and Gian Maria Milesi-Ferretti. 2001. “THE EXTERNAL WEALTH OF NATIONS: Measures of Foreign Assets and Liabilities For Industrial and Developing Coun- tries.” Journal of International Economics, 55 (2): 263–294. 41 [27] Lane, Philip R., and Gian Maria Milesi-Ferretti. 2007. “The external wealth of nations mark II: Revised and extended estimates of foreign assets and liabilities, 1970-2004.” Journal of International Economics, 73 (2): 223–250. [28] Lane, Philip R., and Jay C. Shambaugh. 2010. “Financial Exchange Rates and International Currency Exposures.” American Economic Review, 100 (1): 518-40. [29] Levy-Yeyati, E. 2006. “Financial dollarization: evaluating the consequences.” Economic Pol- icy, 21 (45): 62-118. [30] Maggiori, M., Neiman, B., and Schreger, J. 2020. “International Currencies and Capital Al- location.” Journal of Political Economy, 128 (6): 2019–2066. [31] Mbaye, S., M. Moreno Badia, and K. Chae. 2018. “Global debt database: Methodology and sources.” IMF Working Paper, 18/111. [32] Montamat. G. 2020. “Stubborn Dollarization: Love for the Dollar and Fear of the Peso.” Manuscript, Harvard University. [33] Ottonello, P. and Perez, D., 2019. “The Currency Composition of Sovereign Debt.” American Economic Journal: Macroeconomics, 11 (3): 174-208. [34] Shin, H. S. and Shin, K. 2011. “Procyclicality and Monetary Aggregates.” NBER Working Paper No. 16836. 42 Chapter 2 Transmission of Global Financial Shocks: Which Capital Flows Matter? 2.1 Introduction Disruptive effects of cross-border capital flows in Emerging Market Economies (EMEs) have been thoroughly explored in recent international macroeconomic and financial research. The seminal papers by Rey (2013, 2016) coined the term ”Global Financial Cycle”, saying shocks to risk- appetite of global investors, measured by Chicago Board Options Exchange V olatility Index (Cboe VIX), 1 generate comovements of risky asset prices over the world. 2 In the context of EMEs, as it is shown in figure 1, we can see certain correlations between VIX and EME financial market variables; a higher VIX is associated with falls in EME stock indices and with EME local currency depreciation, and vice versa for lower VIX. Furthermore, when an early draft of this paper was written (2020 March), we clearly saw another big shock to risk-appetite of global investors, insti- gated by COVID-19 pandemic, resulting in historically large falls in stock indices and currency 1 Throughout this paper, VIX indicates Cboe VIX 2 Another important and even more famous claim in Rey (2013) is that monetary policies in peripheral economies can be autonomous only if the capital accounts are controlled, and moreover it is almost regardless of the exchange regime: hence the traditional trilemma has morphed into dilemma. Such a provocative claim ignites the debate of whether the peripheral economies are in the state of trilemma or dilemma. Throughout this paper, I do not try to answer whether EMEs are currently in the state of trilemma or dilemma. However, this paper contributes to the debate by elucidating more precise channel through which the risk-appetite shocks to the global investors propagate into financial markets and real economies in EMEs. For more details of the debate, see Obstfeld (2015), Edwards (2015), and Cerutti et al. (2017). Also see the excellent survey by Aizenman (2018). 43 values in EMEs. This paper aims at improving our understanding of the mechanism by which the risk-appetite shocks, ”Global Financial Cycles”, 3 impact financial markets and the real economy in small open economies and, in particular, Emerging Market Economies (EMEs). This paper focuses on the global financial shock implications for EMEs because, first, EMEs are usually considered more fragile to sudden reversals in capital flows than advanced economies (AEs) and therefore EME policy makers are more concerned about the capital flows. Second, this paper reviews our understanding of the key channel through which capital flows disrupt EMEs. Traditionally, it has been thought that EMEs must borrow in foreign currencies when they borrow abroad, and the resulting currency mismatches of external liabilities with domestic assets are the source of the fragility; the ”Original Sin” hypothesis as espoused by Eichengreen and Hausman (2002). However, there have been important changes in International Investment Positions (IIPs) of EMEs over the last twenty years, which brings into question about the continued validity of the Original Sin hypothesis. As it is documented in recent papers such as Du and Schreger (2016) and Perez and Ottonello (2019), currently substantial parts of external debts of EMEs are local currency (LC) denominated debts. Furthermore, in a companion paper Han (2021), I constructed a dataset, which shows EMEs have increasingly borrowed abroad in local currency equities and debt, and have reduced their currency mismatches. This suggests that we may no longer rely on solely the currency mismatch channel to explain the vulnerability of EMEs to global financial shocks. In this paper, I first empirically evaluate how different types of external liabilities are associated with different sensitivities to global financial shocks. Deploying the dataset in the companion pa- per Han (2021) and strategies similar to papers that studied financial market responses in EMEs to tapering tantrum in 2013 (Aizenman et al., 2014; Eichengreen and Gupta, 2014), I estimated how financial market variables in the EMEs — stock indices and exchange rates — respond to the risk appetite shocks, measured by changes in VIX, conditional on different types of external liabilities of each EME. Surprisingly, it turns out that more equity external liabilities and LC external debts 3 Throughout this paper, risks appetite shocks refer to the shocks to risk appetite of global financial intermediaries, which derive the global financial cycle. I interchangeably use risk-on/off shocks, risk appetite shocks, and global financial cycles. 44 Figure 2.1: VIX and Financial Markets in Selected EMEs Note: 1) Monthly data, Jan. 2008 to Dec. 2018. 2) I normalize the stock price indices and nominal exchange rates at a basis of the values at the beginning of 2007 equal to 100. 3) Exchange rates are price of US dollar in local currencies. Hence, higher exchange rates mean depreciation of the local currencies. are associated with higher sensitivities to global financial shocks at least in terms of financial mar- ket reactions in a short run at monthly frequency. By contrast, measures of currency mismatches, both in the aggregate and at the sectoral level, turn out to be much insignificant. The result is in line with a few earlier empirical studies (e.g., Dedola et al., 2017) that document no clear rela- tionship between country responses and likely relevant fundamentals such as US dollar exposure, but countries with larger capital markets, equity and bond markets, seem to be more fragile. My empirical findings together with these prior results suggest that there are alternative channels for global financial shocks to impact EME financial markets other than currency mismatches. Motivated by these facts, I develop a small open economy (SOE) model augmented by three distinctive features. First, to model equity markets in the SOE, I adopt assumptions in Gertler and Kiyotaki (2010) that firms issue claims on capital every period and the resulting equity-type securities must be purchased by leverage constrained domestic banks or global investors. Second, 45 following Miranda-Agrippino and Rey (2019), I assume that global investors are risk-neutral, but face Value at Risk (VaR) constraints so that they behave like risk-averse agents. Third, government in the SOE can issue sovereign bonds denominated in the local currency of the SOE (LC bonds), which can be purchased by either global banks or domestic banks. The domestic banks finance their investments through either deposit from households or foreign currency borrowing in international loan markets. To understand the mechanisms in the model, consider a “risk-off” scenario in which global investors face some negative shocks to their own capital. The damage to capital forces the global investors to dispose of their risky asset holdings in EMEs. Given initial conditions, the global investors sell off some of their EME equity and EME LC bonds. The other investors in EMEs, domestic banks, cannot absorb the sell-off because they are leverage constrained. The resulting insufficient asset demand reduces the price of capital, which in turn reduces the net worth of do- mestic banks and accordingly lowers the demand from the domestic banks. As a result, the sell-off by global investors generates a negative externality through the domestic capital price, dramatically weakening the total demand for the capital and resulting in a large fall in the asset price: Hence it is a form of fire sale mechanism ignited by sell-off by global investors. On the other hand, in foreign exchange markets, when the global investors sell off the assets in EMEs, they must also sell off their local currency proceeds and convert it to their own foreign currency. This depreciates the lo- cal currency. These impacts of the “risk-off shock” for stock prices and exchange rates, propagate to the real economy, resulting in lower investment in new capitals and higher net exports, which is typically observed in risk-off events. This describes the theoretical ”capital market channel”, for global risk appetite shocks to impact EME financial markets and the real economies, which is the main contribution of this paper. If the domestic banks have net foreign currency debts, the local currency depreciation reduces the net worth of the banks even further, and this precipitates a fall in the domestic capital price: I call this ”exchange rate channel.” Facing the shrinking net worth, banks deleverage reducing both the capital purchases and the foreign currency borrowings as they need to borrow less when they 46 Figure 2.2: The Loop Mechanism in the Model Note: Balance sheets in the middle are the balance sheets of the financial intermediaries in EMEs invest less. As a result, the local currency depreciation causes a fall in capital price and the lower capital price depreciate the local currency further through the deleveraging of the banks. Such as negative loop mechanism magnifies the impacts of a risk-off shock, resulting in larger falls in investment in capital and steeper increases in net exports. The effects are illustrated in figure 2. Based on the empirical and theoretical results, I build a medium scale new Keynesian DSGE model for more quantitative studies. The model is designed for a quantitative study of the impacts of the risk-appetite shocks on small open economy, and the purpose of the exercise is to evalu- ate the importance of the capital market channel quantitatively in a more general environment. I model the leverage constraint, following Gertler and Kiyotaki (2010) and added several necessary ingredients such as incomplete exchange rate pass-through, which might be important for quan- titative results. I calibrate the model to Korean economy where corporates and the government have no significant net foreign currency debts, and whose external liabilities are mostly Korean won denominated equities and bonds. I feed four different shocks into the calibrated model: TFP shock, trade shock — shock to foreign demand for Korean exports — , monetary policy shock, and global financial shock. The results of quantitative analysis illustrate that global financial shock is the most important and dominant force in financial markets in Korea. Approximately, it accounts 47 for 50% of the volatility of capital price, as a proxy for equity price in reality, 40% of real exchange rate volatility, and 30% of the government bond price volatility. The importance of global finan- cial shocks is relatively low for the real macroeconomic aggregates. The global financial shock accounts for approximately 30% of investment volatility, 20% of consumption, and 10% of GDP. These numbers are close to a recent estimate in Acalin and Rebucci (2020). The parts of GDP variations attributable to the risk appetite shocks are low, but it reflects that increases in net ex- ports largely offset the negative impacts on investments during a risk-off event and vice versa for a risk-on event. The model and quantitative studies above cannot accommodate rich institutional details in fi- nancial markets in reality. Since there is no direct evidence of the capital market channel in the literature, to the best of my knowledge, I test the validity of the channel, using more disaggregated data in Korea. Using the rich bank balance sheet data of Korean financial intermediaries, I empiri- cally test the core mechanism. The model predicts the financial intermediaries holding more risky assets and having higher leverages are affected by global financial cycles more than others. The results of the panel regressions show that the financial intermediaries behave as predicted by the model. Related Literature This paper is related to several strands of literature. First and foremost, this paper is a part of the literature that has studied mechanisms behind disruptive impacts of capital flows on EMEs. The literature has a really long history and backs to at least Calvo (1998). The preceding papers in the literature have focused on the potential risks from foreign currency external borrowings. Caballero and Krishnamurthy (2001) showed that the collateral constraints in EMEs generate the pecuniary externalities of foreign borrowings, which raise financial fragility in EMEs. After the Global Financial Crisis, there have been extensive studies on the pecuniary externalities from foreign currency external borrowings and related policies. Noteworthy papers in the literature are Bianchi (2011), Mendoza (2010), Beningo et al. (2016), and Jeanne and Korinek (2010b). The central idea in these papers is that decentralized agents do not internalize the impact of their actions on prices, real exchange rates in most papers, and capital controls are desirable policies to 48 handle the externalities. More recently, several papers emphasized the interaction between external shocks (e.g., sudden stops) and domestic banking sectors using a model based on Gertler and Kiyotaki (2010) or Gertler and Karadi (2013). Akinci and Queralto (2019), Aoki et al. (2018), and Jiang et al. (2019) belong to this fashion. This paper also constructs a small open economy model augmented with domestic banking sectors. The distinctive feature in this paper is that unlike the earlier papers, the mechanisms do not rely on currency mismatches. Based on the up- to-date empirical facts, I incorporate equity and local currency debt portfolio investments into the model and suggest a mechanism by which capital flows in the form of equity or LC debt portfolio investment generate large fluctuations in financial markets and the real economy. Handful recent papers have similar features with the model in this paper. Cavallino and Sandri (2019), Jeanne and Sandri (2019), and Caballero and Simpsek (2020) showed that sell-off of foreign investors in domestic financial markets can generate big falls in the market and a severe recession if domestic investors face a form of borrowing limit or collateral constraint. The model in this paper differs from Cavallino and Sandri (2019) and Jeanne and Sandri (2019) in that in my model, the asset price falls lower net worth of domestic investors so as to amplify the negative impacts. In this aspect, the model in this paper is close to Caballero and Simpsek (2020). However, despite the similarity, the model in this paper identifies different channels from different capital flows more precisely in richer environments. The model is also better grounded on empirical facts uncovered in this paper and more suitable for quantitative studies as the model has more realistic and richer features. 4 Another paper close to mine is Devereux and Yu (2019), which modeled different capital flows, equity and debt, and global investors who intermediate cross-border borrowing. Because of the similar features, this paper echoes one of key insights in Devereux and Yu (2019) that equity market participation of foreign investors transmits foreign shocks to the domestic markets, resulting in less severe but more frequent crises (or market turmoil). However, this paper emphasizes the role of asset price affected by fickle demands from foreign investors, which is absent in Devereux and Yu (2019). 4 In addition, the context of analysis in Caballero and Simpsek (2020) is on advanced economies, while this paper focuses on EMEs. 49 This paper is also related to the literature of global financial cycle and monetary policy spillover. It was Rey (2013) that coined the famous term “Global Financial Cycle” and suggested a provoca- tive claim that a small open economy loses its independent monetary policy as long as its capital account is open since the center economy monetary policy in fact determines the financial condi- tion of the small open economy through the changes in risk appetite of global investors: therefore, the traditional trilemma has morphed into the dilemma. There are two main related questions in the literature. The first question is whether SOEs are in the state of the trilemma or the dilemma. The literature has yet to reach a consensus. Aizenman et al. (2016) and Cerutti et al. (2017) provide evidence for the trilemma. Han and Wei (2016) argued SOEs lie somewhere between the trilemma and the dilemma. I do not directly address the question of the trilemma or the dilemma, but the findings in this paper imply that exchange rate regime does matter, but letting exchange rates float cannot be enough to insulate SOEs from global financial cycle: similarly with Han and Wei (2016), the state is between the trilemma and the dilemma. 5 Another important question in the literature of global financial cycle is what are the mecha- nisms behind global financial cycles? and relatedly which countries are more vulnerable to the risk appetite shocks, according to the mechanisms? Few papers such as Akinci and Queralto (2019), Aoki et al. (2018), and Cavallino and Sandri (2019) listed above, using structural mod- els, pioneered transmission mechanisms that risk appetite shocks transmit to EMEs. 6 Many more papers empirically examined different possible transmission channels using cross-country data or micro-level data in a specific EME. Papers worth mentioning here are Aizenman et al. (2016) and Eichengreen and Gupta (2014) that investigated the financial market reactions in EMEs during the tapering tantrum in 2013. Baskaya et al. (2017) document the transmission of the risk appetite shocks to local credit supplies, using Turkish bank-level data. 7 As explained above, this paper 5 This is also similar with Obstfeld (2016). 6 Another related strand of literature looks at how sovereign default risks evolve along with global financial cycle and how it affects EMEs. See Morelli et al. (2019) and Arellano et al. (2020) 7 See also Georgiadis and Zhu, (2019), Fendoglu et al. (2019), Avdjiev and Hale (2019), and Cesa-Bianchi et al. (2018). Other related influential works are Bruno and Shin (2015a, b). Bruno and Shin (2015a) empirically and theoretically showed that risk appetites of the banks are closely linked through cross asset holdings among the banks. Bruno and Shin (2015b) showed that the US monetary policy is an important factor in determining risk appetite of global investors. 50 contributes to the literature by providing a novel transmission mechanism of global risk-appetite shocks to SOEs. I also provide evidence of the transmission channel from the bank balance sheet data in Korea. In addition, this paper contributes to the literature by providing information of the currency mismatches in EMEs. Using a constructed data set showing the states of IIPs of 20 major EMEs in terms of currencies (local versus foreign) and instruments, I show that for many EMEs, it is hard to explain the transmission of risk appetite shocks as a result of the exchange rate channel since in many EMEs, foreign currency external net foreign currency debts are small or external foreign currency assets exceed the foreign currency debts. Finally, this paper shares some insights and features with papers that pioneered the implication of heterogeneous financial development among countries on risk sharing in the world and the global imbalance. The related influential papers are Gourinchas and Rey (2014), Mendoza et al. (2009), Caballero et al. (2008), and Maggiori (2017). The central idea in these papers is that in equilibrium, AEs with more developed financial markets will carry more risky assets than EMEs with less developed financial markets so that it generates higher income for AEs from the risky assets and the following current account deficits for AEs. For EMEs, vice versa. My contribution is I take the view to short run dynamics in EMEs and show how the changes in foreign investors’ demand for financial securities in EMEs induce fluctuations in the EMEs: The underlying reason behind this is that domestic banks in the EMEs have limited capacity to hold risk assets. Layout The rest of the paper is organized as follows. Section 2 conducts a simple empirical anal- ysis using the data set and uncover new facts. Section 3 introduces a small open economy model. I firstly will introduce a simple model by which I will derive some analytical results capturing the empirical findings. The model illustrates the new transmission channel that risk-appetite shocks to global investors impact EMEs through equity liabilities and LC debts. Section 4 introduces the results of more quantitative studies, using the medium-scale DSGE model based on the simple model. Section 5 concludes and discusses avenues for future researches. 51 2.2 Empirical Analysis In this section, I conduct a simple empirical analysis to see how the fragility of EMEs to global financial shocks are associated with different types of external liabilities — equities, LC debts and foreign currency debts. The data used in the regressions come from a companion paper Han (2021). In the companion paper, I showed that EMEs have increasingly borrowed abroad in LC equities and debts, and relatedly currency mismatches in the EMEs at the aggregate and the sectoral levels have greatly reduced. Surprisingly, the simple regression analysis using the novel data shows that in contrast to the usual belief, financial markets in EMEs that have more equity external liabilities and LC external debts seem to be more sensitive to the risk appetite shocks. 2.2.1 Empirical Strategies In this subsection, I conduct a simple empirical analysis using the data. Broadly speaking, the main purpose of the analysis is to find what kind of fundamentals are related to higher fragility to risk-appetite shocks so that from the information, I can guess specific channels of the transmission to EMEs. In practice, I examine which types of external liabilities — equities, LC debts, and FC debts— are associated with higher fragility to the global financial cycles. For this purpose, I need a measure of fragility and another measure of the risk appetite shocks. For the risk appetite shocks, I can conveniently use VIX as a measure of it. 8 Therefore, a rise in VIX indicates a higher risk appetite (risk-on), and naturally a fall in VIX indicates a lower risk appetite (risk-off). Henceforth, I use risk-on/off shocks for shocks to risk appetite of global investors, as the terminology is widely used in market participants and commentators. As measures of the fragility, I can use different variables; financial market prices such as stock indices or exchange rates, quantities in credit mar- kets such as credit growth or real economy variables like GDP growth. Although none of these are perfect, I decide to use monthly percentage changes in financial market prices — stock indices 8 One alternative approach is to use Factor model to estimate co-factor of risky prices in the world. Careful esti- mation can reveal more precise measure of global financial cycle, but may not provide a meaningfully different result. Other observable measures are US dollar index as argued in Shin (2016) and US monetary policy shocks. However, all these different measures are hardly much different from VIX. 52 and exchange rates — due to following considerations. The credit growth or real variables would adjust to global financial cycle with lags and the lags will be different among EMEs, which forces me not to use a simple and tractable approach. 9 Similarly, by taking responses of financial market prices in a relatively short run (in a month), I can suffer less from possible various endogeneities; for example, in a longer horizon policy authorities in EMEs that experienced bigger market falls take actions to boost the markets. 10 Hence, monthly data is a way to lessen the possible different endogeneities and avoid noises in daily or weekly data. Another problem I encounter is that the data on external liabilities has a low frequency. While the IIP from IMF and some of local currency debt and equity are quarterly or even monthly, local currency debt and local currency equity data in some EMEs are annual. Furthermore, foreign cur- rency deposits and loans are annual data in all the EMEs. Since the main interest in the regression is to identify different responses to a common risk on/off shock among the EMEs, I take the annual data of all different types of external liabilities; local currency bond, equity, net foreign currency debts, and so on, 11 except for reserves for which I have monthly data for all the sample EMEs. Therefore, in my monthly data, for example, net foreign currency debts of nonfinancial corporate sectors in 2012 are the same from January 2012 to December 2012. This is a little unsatisfactory. However, if I give different frequencies to different types of external liabilities in different EMEs, that might cause a bias toward certain EMEs or certain types of external liabilities. Therefore, taking the annual average is inevitable, although I admit the drawback in my regressions here. However, I would like to emphasize this is the best way to deploy all the available information, while not manipulating the data arbitrarily. Also, the external liabilities are stocks, not flows; hence it cannot change drastically in a month. Furthermore, taking the averages help me with handling possible endogeneities. 9 Moreover, estimation of the impacts on real economy variables or other quantities with some lags call for endo- geneity of policy response, which poses another challenge. 10 Another important benefit of monthly data is the number of observations. My data on local currency external debt has a relatively short time span (from 2011 to 2018), and hence using monthly data of stock indices and exchange rates has an advantage of more observations 11 Another possible solution to the problem is to take the average over the whole sample, soL j t =L j : This is possible in the regressions without sector level currency mismatches since the aggregate level data is rather stable. The results of the whole sample average is introduced in the appendix and the results are similar with the annual averages. 53 As a result, the regression is as in equation (2.1). The approach extends ideas in Aizenman et al. (2016) and Eichengreen and Gupta (2014) who studied the impact of the 2013 tapering tantrum shock on financial market variables in EMEs. I also borrow some features from Rey (2013). 12 Again, I would like to note that the main interest here is to see how the fragilities to the risk appetite shocks change along with the key variables, amounts of different types of external liabilities — equities, LC debts, and FC debts. y j t =a j +ry j t1 +d 0 vix t +d 1 ln(V IX t1 )+b 0 L j t vix t +G 0 0 c j t vix t +G 0 1 X j t +e j t (2.1) where y j t is either of the percentage changes in the nominal bilateral exchange rates of country j, 13 denoted byDEx j t , or the percentage changes in stock index in country j, denoted byDStock j t .L j t is the key variable in this regression equation. L j t is a vector of different types of external liabilities and assets to GDP ratios; LC equities to GDP ratio, and similarly for LC debts, FC debts, official reserves, and external assets by private sectors. For other term, vix t = log di f f erence o f V IX, X = the vector of controls. in f lation; IP; M2; i j ; i j i us ; Real e f f ective exchange rate(1) , and c j t = the vector of variables representing country characteristics such as trade openness, financial openness, government debt to GDP ratio and so on. I used Driscoll-Kraay standard errors to handle heteroskedasticity and cross-sectional dependence. However, the results I introduce below are robust to different methodologies to control heteroskedastic standard errors. 14 In the regression equation (2.1), VIX is almost exogenous to emerging market stock indices and exchange rates so as to relieve concerns about possible endogeneities. However, another key explanatory variableL j t is endogenously determined equilibrium outcomes. A reasonable concern is what ifL j t is related to fragility to global financial cycles. 15 Because of limited data, I cannot 12 The use of interaction terms between country characteristics and global financial cycle variables like VIX is popular in empirical studies of US monetary policy spillover. 13 Hence the higher exchange rate indicates a depreciation of the currency of country j) 14 Since our main interest is to compare different responsiveness of different countries to common shocks, it is crucial to use methodologies controlling heteroskedastic standard errors. 15 Since we take the annual data ofL j t , changes in external liabilities in the short run due to risk on/off shocks do not seriously matter. Also, in the appendix I introduce the results of using a laggedL j t following the idea of Bartik instrument. I obtain similar or even stronger results. 54 Table 2.1: Exchange Rate Regressions (1) (2) (3) (4) (5) (6) (7) (8) Dln(V IX) t 0.044*** 0.043*** 0.040*** 0.039*** 0.041*** 0.041*** 0.059*** 0.076* [0.012] [0.012] [0.014] [0.013] [0.012] [0.012] [0.022] [0.045] LCD GDP j t ” 0.105** [0.043] LCB GDP j t ” 0.176*** 0.175*** 0.184*** 0.254** 0.184** [0.066] [0.053] [0.069] [0.112] [0.080] LCE GDP j t ” 0.034 -0.022 -0.013 -0.056 [0.033] [0.071] [0.068] [0.052] FCD GDP j t ” 0.008 0.016 0.008 -0.012 -0.028 [0.043] [0.040] [0.044] [0.047] [0.052] FCA D GDP j t ” -0.092 † -0.095 † -0.074 † -0.088 [0.062] [0.064] [0.048] [0.084] FCA E GDP j t ” 0.010 0.019 0.006 0.026 [0.019] [0.038] [0.031] [0.028] Reserve GDP j t ” -0.039 -0.071*** -0.041 -0.058** -0.024 -0.017 -0.038 -0.053 † [0.029] [0.024] [0.029] [0.026] [0.044] [0.057] [0.044] [0.035] Country FE Yes Yes Yes Yes Yes Yes Yes Yes Time FE No No No No No No No Yes # of Obs. 1,660 1,660 1,660 1,660 1,660 1,660 1,660 1,660 R-squared 0.069 0.036 0.053 0.055 0.017 0.026 0.122 0.297 Note: 1) *** p¡0.01, ** p¡0.05, * p¡0.1, * p¡0.1, † p¡0.15. 2) LCD: Local Currency Debt, LCB: Local Currency Bond Portfolio, LCE: Local Currency Equity, FCD: Foreign Currency Debt, FCA D: Foreign Currency External Asset (Debt instrument), and FCA E: Foreign Currency Asset (Equity). 3) Driscoll-Kraay standard errors. 4) Regression (6) and (7) adds more controls (e.,g oil price, commodity price index and related groups) to regression (5). 5) Time fixed effects are not two-way fixed effects, but time dummies (random effects), because one of the key explanatory variables, the VIX log difference is the time series variable. rule out all the possible endogeneities, but at the end of this section, I show that a scenario that one easily comes up with does not correspond to the historical data. 2.2.2 Results I first introduce the results of the exchange rate regressions. For brevity, I introduce only the estimated coefficients of the key variables. The results for other control variables are relegated to the appendix. I denote local currency debts, local currency bond portfolio, local currency equity portfolio, foreign currency debt, foreign currency asset of debt instrument, and foreign currency asset of equity by LCD, LCB, LCE, FCD; FCA D; and FCA E respectively. 55 Surprisingly, it turns out that local currency denominated debts are highly associated with higher fragilities to the risk appetite shocks in terms of currency, and I get much stronger results once I replace LC debts with LC bonds; I extract LC deposits from LC debts. On the contrary, foreign currency debts are insignificant: no clear relationship between the amounts of foreign cur- rency debts and the measured sensitivities of a currency to the risk appetite shocks. Regarding the asset sides, external assets by private sectors are mostly insignificant, while official reserves are significant in some specifications; more official reserves are associated with lower fragility to risk appetite shocks. The results introduced above are only for nominal exchange rates and the sensitivities of the market variables to risk appetite shocks in short run. However, despite the limitation, considering that currency depreciation is often understood as a measure of magnitude of impact of an external shock on a small open economy, the results in the table must be unexpected and surprising. There are not many studies of the risk-sharing features of local currency denominated debts, but it is straightforward that local currency denominated bonds have some risk-sharing properties. If any negative shock to a small open economy results in depreciation of the local currency of the small open economy, then the depreciation will reduce the real debt burden, thereby limiting the local currency depreciation in turn. 16 Hence, the standard model predicts that we can see negative coefficients for local currency debt (or bonds), or smaller and less significant coefficients in terms of absolute value than foreign currency debts. A possible way to interpret the results looking seemingly counterintuitive is that LC bond portfolio investments, carry trades by another name, are more sensitive to risk on/off shocks than other types of capital flows such as foreign currency denominated debts: local currency denominated bonds are absolutely riskier for global investors. I will show more details using the model in this paper. Next, I introduce the results of monthly stock indices regressions. It turns out that all the measures of GDP ratios – the type of liabilities to GDP ratios – are insignificant, except for foreign currency liabilities. On the contrary to the GDP ratios, the local currency equity external liabilities 16 For the related mechanisms, see Fanelli (2018) and Korinek (2009). 56 Table 2.2: Stock Indices Regressions (1) (2) (3) (4) (5) (6) (7) (8) Dln(V IX) t -0.093*** -0.089*** -0.088*** -0.076*** -0.076*** -0.082*** -0.059*** -0.059** [0.016] [0.014] [0.015] [0.014] [0.014] [0.015] [0.016] [0.025] LCE GDP j t ” 0.010 [0.037] LCB GDP j t ” -0.003 0.091 0.084 0.145* 0.183* [0.092] [0.096] [0.100] [0.091] [0.089] LCE Mkt Cap j t ” -0.072* -0.093** -0.093** -0.103** -0.105** [0.037] [0.038] [0.039] [0.043] [0.051] FCL GDP j t ” 0.016 0.027 0.028 0.020 [0.041] [0.042] [0.040] [0.050] FCA D GDP j t ” -0.015 -0.011 -0.012 [0.082] [0.085] [0.093] FCA E GDP j t ” 0.003 -0.006 -0.012 [0.016] [0.015] [0.017] Reserve GDP j t1 ” 0.070** 0.073*** 0.071** 0.081*** 0.074** 0.076* 0.056 0.027 [0.030] [0.029] [0.030] [0.030] [0.030] [0.043] [0.044] [0.047] Country FE Yes Yes Yes Yes Yes Yes Yes Yes Time FE No No No No No No No Yes # of Obs. 1,660 1,660 1,660 1,660 1,660 1,660 1,660 1,660 R-squared 0.070 0.077 0.048 0.060 0.059 0.049 0.083 0.244 Note: 1) *** p¡0.01, ** p¡0.05, * p¡0.1, * p¡0.1, † p¡0.15, †† p¡0.20. 2) LCD: Local Currency Debt, LCB: Local Currency Bond Portfolio, LCE: Local Currency Equity, FCD: Foreign Currency Debt, FCA D: Foreign Currency External Asset (Debt instrument), and FCA E: Foreign Currency Asset (Equity). 3) Driscoll-Kraay standard errors. 3) Regression (6) adds more controls (oil and commodity price indices along with related groups, government debt) to regression (5). 4) Regression (6) adds trade openness and financial openness, whereas drops commodity prices. 5) Time fixed effects are not two-way fixed effects, but time dummies (random effects), because one of the key explanatory variables, the VIX log difference is the time series variable. to total stock market capitalization ratios, hence foreign investor shares in domestic equity markets, are negative and all significant at least 10% level. That is, the higher the foreign investor shares in the stock market are, the more fragile the stock market is to global financial shocks. Same as the exchange rate regressions, foreign currency debt and asset are much insignificant in all the specifications. Other noteworthy results are international reserves help the EME with reducing the impact of risk-on/off shocks on the stock markets, and local currency bond, which is highly correlated with more fragility of local currency in the exchange rate regressions, is positive and weakly significant. 57 This result looks puzzling as well, but interestingly the results are in line with several preceding papers. Eichengreen and Gupta (2014) documented that EMEs with higher stock market capital- ization to GDP ratio or more opening capital markets suffered more from the tapering tantrum shock and Aizenman et al. (2016) also report a similar result; more developed EMEs, which have probably larger capital markets, were dampened more during the market turbulence due to the ta- pering tantrum; more developed EMEs tend to have larger capital markets and high foreign investor shares in their stock markets. Dedola et al. (2017) showed that there exists a great heterogeneity in terms of the responses of economic variables in EMEs to the US monetary policy shocks, and there is no clear-cut relation between country responses and likely relevant country characteristics, such as income level and the USD exposures. In the regressions conducted above, foreign currency assets and liabilities are measured on the aggregate level. In a section in the appendix, I replace the aggregate level data with sector level currency mismatches. I add net foreign currency debts of the four different sectors — households, financial corporate sectors, nonfinancial corporate sectors, and government. Overall, the results are much the same as the regressions of the aggregate currency mismatch. Local currency bond portfolio is significant in all the regressions. All the net foreign currency assets in different sectors do not show strong enough significance although almost all signs are negative; more net foreign currency assets are associated with higher robustness of the local currency to risk-on/off shocks at least in terms of the sign. I relegate more details in the results and following interpretations to the appendix. I also conducted various robustness checks; adding more control variables, taking the whole sample period averages of the different types of external liabilities or one year lag of the external liabilities. All the different trials show similar results with the baseline model. Discussion of endogeneity One possible interpretation of the results is that some EMEs issue more local currency denominated securities to foreign investors because the EMEs are more fragile to global financial cycle. The idea follows from a typical risk sharing argument. Both equity and LC debt have properties that payments to foreign investors are counter-cyclical to global financial 58 cycle; payments decrease when there is a negative shock to the risk-appetite of global investors. If there is an EME whose business cycles follow the Global Financial Cycle, then the EME, given other conditions, is incentivized to issue more equities or LC debts to global investors if the global investors are risk-neutral. Although I cannot completely rule out the chance of such endogeneity since the data used is not rich enough, I show that at least the results above are unlikely to come from the endogeneity. The empirical results are not because some EMEs with higher exposures issue more equities and LC debts to foreign investors. First of all, the interpretation that fragile EMEs sell more equities and LC debts to global investors misses the risk appetite shock is a global systemic shock. As typically argued, VIX is a measure of a cofactor of risky assets in the world. Hence, risks measured by VIX are the risks to every investor and it is the same for the global investors, who manage different assets in different countries all over the world. Therefore, assets in EMEs whose business cycles are positively correlated with risk appetite shocks are less attractive to global investors in terms of risk sharing. Then it is straightforward that such EMEs need to provide higher premiums if they want to sell equities and LC debts to global investors. On the other hand, the issuers in EMEs are indifferent between sharing country specific risks and sharing systemic global risks as long as both risks are their own risks. In contrast, global investors would not care much about country-specific risks. As a result, in the argument of frictionless risk sharing, EMEs whose fundamentals are less or negatively correlated with global financial cycle are more incentivized to issue more equities or LC debts to global investors because they can share their risks at lower costs. Altogether, if the risk sharing argument in a frictionless economy works, the signs of the coefficients must be opposite from the two tables. In the appendix, I explain my counter-argument more formally using a simple small open economy model. 17 Second, historical evidence is unfavorable for the risk sharing argument that more fragile EMEs issue more equities and LC debts to global investors. If EMEs that were fragile to external financial 17 The overall argument here is related with Hassan et al. (2016) in that risk properties of a currency can attract more or less foreign capitals in the country. 59 Figure 2.3: Stock Index Betas and the increase in LC external liabilities Note: 1)DLC=LC external liabilities (equities and LC debts) to GDP ratios in 2018 minus the same ratios in 2001. 2) Left panel includes all the EMEs in the sample, except for Romania and right panel excludes an outlier, Russia. shocks have issued more local currency denominated external liabilities — equities and LC debts— to foreign investors, then we should see positive correlations between the fragility in the past (from 1995 to 2001) 18 and local currency denominated external liabilities in the present. To check this while avoiding possible complexities, I first estimate the “beta” of each currency and stock indices in 1990s as follows. y j t =a j +b j vix+e j t (2.2) where y j t is monthly percentage changes in either exchange rates or stock indices, and vix is the percentage changes in Cboe vix same as the regressions above. I run the regression for each country so that I have twenty betas of exchange rate for 20 EMEs and same for the stock indices. Then I plot the betas against the amounts of local currency liabilities, including both equities and LC debts, in the 20 EMEs. For exchange rates, for most of EMEs, the betas are not significant, reflecting on the facts that many of the EMEs were under fixed exchange rate regimes. Hence, I plot the stock index betas against the increase in local currency liabilities. As one can easily see, there is no clear relationship between the two variables. Although the exercise is a little crude, upon investigations that have been done so far, there is no clear relationship 18 This time period is to avoid the eras of hyperinflation in Latin American countries and the time that relative LC external liabilities among the EMES are similar with the present distribution. 60 between the fragility in the past and the current distribution of local currency denominated external liabilities. Then, what kind of fundamentals show a significant relationship with the distributions of the equity external liabilities and LC external debts? In the companion paper Han (2021), I show that the depth of capital markets — stock and bonds markets — are correlated with the external liabilities of LC equities and bonds. That is, EMEs having larger stock markets tend to borrow more abroad in the form of equity and similarly, EMEs having larger bond markets tend to borrow more in LC bonds. In a section in the appendix, I suggest a simple model to explain the empirical regularities as the model in the appendix is actually a simple extension of the the theoretical model in this paper. The interpretation of the theoretical results will also be given in the appendix. I will interpret the facts as results of the risk sharing desires of the global investors, not securities issuers in EMEs. Summary of empirical findings Before I move on to the model section, I summarize the empir- ical findings that guide me to build a new model 1. Higher LC debt to GDP ratios are associated with higher sensitivity of nominal exchange rate to global financial shocks, changes in VIX 2. Similarly, higher foreign investor shares in stock markets in EMEs are associated with higher sensitivity of the stock indices to global financial shocks. 3. In both exchange rates and stock indices, no significant relationship is found between foreign currency debts and the fragility to global financial shocks. 2.3 Model Having documented that the local currency liabilities are associated with higher fragility to the risk appetite shocks in opposition to conventional wisdom, I now suggest a model to reveal the mechanisms by which risk appetite shocks to global investors result in large fluctuations in financial 61 markets and the real economy in EMEs, through equity external liabilities and LC debts. To be more specific, the two main purposes of the model are 1) to capture the uncovered empirical regularities in the model, and 2) to study how the impact on the financial markets propagate into financial markets and the real economy in EMEs. The key insight from the model is that sell-off from global investors cannot be absorbed by domestic investors and it generates a fire sale mechanism. To obtain the insight, I focus on deriving key analytical results and for the purpose, I maintain the minimum ingredients in the model. In the second subsection, in addition to the theoretical results, I provide evidence from bank level data in Korea, which support the existence of the new channel in the model. 2.3.1 Simple Model The model has three main features, 1) Gertler and Kiyotaki type capital market in that producers issue securities of the claims on the capital like equities in reality, and the securities are purchased by other agents, 2) leverage constraints on domestic banks, and 3) global investors who invest in (LC denominated) domestic capital markets and governments bonds. Other features such as foreign currency debts will be added depending on the purpose. The model is a small open economy model in discrete time with infinite horizon. 2.3.1.1 Environments There are six types of agents in the model: workers, goods producers, capital producers, domestic banks, government and global (foreign) investors. Workers supply labor to the goods producers and save in domestic banks in the form of deposits or invest in government bonds. Goods producers produce consumption goods to be consumed domestically or exported, and they issue securities of claims on capitals, which have to be purchased by either domestic banks or global investors. Capital producers supply (or disinvest) capitals depending on the demands from goods producers. Domestic banks take deposits from workers and supply the funds to the goods producers; buying the securities issued by the producers. Government provides fixed amounts of public services 62 and, to fund the activities, collect taxes or issue government bonds. Global investors invest in the securities or government bonds. The representative household consists of a continuum of bankers and workers with the total population size being normalized to be unity. Each banker member manages a bank (financial intermediary) until he/she retires with probability 1s: retired bankers transfer their remaining net worth as dividend, to the household and are replaced by a given number of workers who become new bankers. New bankers receivex fraction of total asset from the household as start-up funds in total. Bankers will be described in detail later. Workers in the model, as usual in the literature, consume both domestic and imported goods, and supply labors. My purpose in this subsection is to derive some intuitive and analytical results from the simple model. For the purpose, I abstract from the labor supply; there is no disutility of la- bor, and therefore workers supply all the labor endowments. The optimization of the representative household is formulated as follows. max f c d t+ j ;c m t+ j g ¥ j=0 E t h å L t;t+ j U c d t+ j ;c m t+ j i sub ject to c d t +e t c m t + d t + b d t +t t w t L+ R t d t1 R g t b d t1 +p t where c d t is the domestic consumption good, c m t is the imported consumption good, t t is the tax payments, b d t and d t are the government bonds and deposits made at time t, R g t and R t are the returns to the bonds and the deposits respectively, from date t 1 to date t, and e t is the price of imported goods in terms of domestic goods, the terms of trade. Since there is no inflation in the model in this section, the terms of trade is the same as nominal exchange rates. I find it is convenient to take it as a proxy of real exchange rates, whose changes are qualitatively the same as the terms of trade as I implicitly assume foreign price is fixed to 1. Henceforth, I call e t real exchange rate. 63 The per period utility function of the consumption is given by U c d t+ j ;c m t+ j =(1w)ln c d t+ j +wln c m t+ j wherew2(0;1) Notice that every term in the budget constraint of the works is denominated in local (home) currency. I follow the convention that the exchange rate of a country is the price of the foreign currency in units of the domestic currency, so an increase in the exchange rate is a depreciation in the local currency. b is the discount rate. p t is the profits from the capital producers and bankers . The optimality conditions for the workers are characterized by the standard Euler equations. E t " b U c d t+1 U c d t R t+1 # = 1 (2.3) U c d t =e 1 t U c m t (2.4) Producers As noted earlier, there are two types of producers. Before describing the different types of the producers, it is important to clarify that I do not impose any financial frictions on producers: producers can borrow as much as they want. This simplification is consistent with the papers based on Gertler and Kiyotaki (2010) and Gertler and Karadi (2011), which focus on the frictions in financial intermediations. Goods producers operate in perfectly competitive markets. For simplicity, I assume constant returns to scale Cobb-Douglas production with capital and labor as inputs. That is, Y t = A t K a t1 L 1a 64 where K t1 is the total capital stock from the last period and L is the time-invariant labor endow- ments. The optimization conditions for the producers are as follows. A t (1a) K t1 L a = w t (2.5) Capital producers supply new capitals or divest existing capitals using final goods subject to the adjustment cost of investment. The adjustment cost is characterized asF(I t ) whereF(I t ) 0 ¿0 andF(I t ) 00 ¿0 . The capital producers’ problem is defined as max f I t+ jg E t å L t;t+ j Q t+ j I t+ j I t+ j +F I t+ j where L t;t+ j is the stochastic discount factor and Q t+ j is the capital price; of course, it is the Tobin’s Q. For tractability, I use the simplest form of the adjustment cost; actually, the capital producer problem is static. F(I t )= j 2 I t K t1 2 K t1 Domestic banks The banks 19 in this paper purchase capital goods in each period by issuing deposits to households and using own net worth. We can think the purchases as channeling funds from households to firms in all available forms in reality: it includes bank loans, bonds, outside equities, and others. Hence, the value of the capitals purchased by the banks must equal the sum of the banks’ net worth and the deposits. That is, Q t k d t = N t + d t (2.6) where Q t is the capital price, N t is the net worth of the bank, and d t is the deposit. The net worth of the bank evolves in the following way. N t =s (z t + Q t )k d t1 R t d t1 +x(z t + Q t )k d t1 (2.7) 19 The bank here refer to all kinds of financial intermediaries in reality. 65 where z t is the dividends to the capital holdings 20 . For notational convenience, I define R k t+1 = z t+1 +Q t+1 Q t . Then the value of the bank net worth is N t =(s+x)R k t Q t1 k d t1 sR t d t1 (2.8) As already noted, the banks are managed by the bankers who were workers in the past; In each period,x of workers become bankers and in the other way,s bankers retire; retired bankers become workers. This retirement eliminates the possibility that banks accumulate retained earnings so that they will eventually nullify all the financing constraints. Most importantly, domestic banks face leverage constraints. We assume N t f t Q t k d t (2.9) f t is the leverage ratio of the banks and generally, it can be a function of the expected prof- itability and risks in the future. In this section, I letf t be a constantf. More general specification will be used in the next section and I also discuss different specifications in the appendix. Global investors Global investors are international financial intermediaries who purchase local currency denominated equities and bonds in the small open economy. Like other components in the model, I model the global investors in a simple way, but also aim to capture key features in reality. Since this paper studies impacts of risk appetite shocks on global investors, the global investors in the model need to be risk-averse. While there are different ways, I model global investors as international financial intermediaries under “Value at Risk” (VaR) constraint, following Miranda- Agrippino and Rey (2019) and Zigrand et al. (2010). The key idea in their model is that financial intermediaries are risk-neutral in terms of their preference, but act as they are risk-averse as they face a VaR constraint. 20 For simplicity, I set the capital depreciation rate as zero. 66 For detailed steps of the derivation, I refer readers to appendix B. The investments of global investors in the equities and local currency bonds in the small open economy are characterized by the following equations. p k t = Q t k f t e 1 t = 1 Ge v t c 0 k +c 1 k E t e t e t+1 R k t+1 R m t+1 (v t ) (2.10) p b t = b f t e 1 t = 1 Ge v t c 0 b +c 1 b E t e t e t+1 R b t+1 R m t+1 (v t ) (2.11) where bothc 0 k andc 0 b 2(0;1). The terms in brackets,E t h e t e t+1 R k t+1 i R m t+1 are the expected excess returns to the investment in the assets in the small open economy, in which R m t+1 is the return to the global market portfolio denominated in foreign currency, 21 like yields on BAA grade corporate bonds in the US, andE t h e t e t+1 R k t+1 i andE t h e t e t+1 R b t+1 i are expected returns in foreign currency to equities and local currency bonds in the small open economy. The return to the global market portfolio R m t+1 can, of course, reacts to the risk appetite shocks and thus I let it be a function of the risk appetite shocks. c i measures the amounts that the global investors allocate to asset i, regardless of the return, due to the risk management. 22 The constant terms c i are important in deriving sensible quantitative results, but qualitatively do not matter. Hence, depending on the analytical purposes, I assume c 0 i = 0 or c 1 i = 0 in this simple version of the model. 21 The characterizations in equations (10) and (11) are identical to Gabaix and Maggiori (2015) if I let c i = 0 and replace R with R f , the return to the safe asset like return to the US treasury bills. Gabaix and Maggiori (2015) posits an investor who arbitrage between Japanese Yen denominated government bonds and US dollar government bonds, while the investor in my model is arbitraging between different risky assets. Since the benchmark for the investors is an risk asset, the asset, which the global investors in the model compare to the assets in the small open economy should a risky asset. 22 The forms in equations (10) and (11) are approximations from the result of optimal portfolio of the global investors who want to maximize the sharpe ratio of her portfolio. Then is it intuitive that the investors allocate some of her funds to some assets despite low returns if the assets have good risk hedging properties. Another possible interpretation is the constant term reflects some stickiness in the portfolio, due to some informational frictions or gravity in capital flows. 67 1 Ge v t is a measure of the risk appetite of the investors. As one can easily expect, a lower 1 Ge v t indicates lower risk appetite; i.e., higherGe v t indicates lower risk appetite. e v t captures the time- varying risk appetites of the investors. Thus a positive shock to v t means a shrink of the risk appetite, as VIX does so in reality; thus v t is an analogy to VIX. I assume v t follows an AR (1) process as below. v t =r v v t1 +n t (2.12) wheren t s N 0;s 2 n andr v 2(0;1). Henceforth, I calln t > 0 “risk-off” shock andn t < 0 “risk-on” shock. n t is modeled as shocks to the net worth of global investors, who are large international financial intermediaries, adopting the interpretation in Miranda-Agrippino and Rey (2019). Of course, the driving force behind the changes in the risk appetite is not necessarily a shock to the capitals of the global investors. It can be some abrupt changes in the beliefs of the investors or can even be behavioral; for example, changes in the market sentiment. In the context of this paper, adopting a different microfoundation does not alter the specification in this paper or following economic interpretations. To summarize the discussion, investments of the global investors — local currency equity and bond capital flows to the small open economy — are determined by the two factors: the expected excess return and the risk-appetite. Government Government in the small open economy has to make an expenditure at the amount of G every period. To make the expenditure, the government collects taxes from households by the amount oft t . The tax must be not enough to make the expenditure of G, and hence the government issues one period short-term government bonds B t , denominated in local currency. The budget constraint is as below. G=t t + B t R g t B t1 (2.13) Also, I assume that there is no constraint on holding government bonds. This condition imposes that the return on the government bonds, R g t ;must be same as the interest rates on the deposits. 68 That is, R g t = R t by no-arbitrage conditions. In addition, I fix the government bond stock at B. Therefore, B t =B for all t. 2.3.1.2 Market Equilibrium To have market clearing conditions for goods in this model, we need a specification of the exports of the small open economy. I assume that export demand for goods by foreigners, EX t , is a de- creasing function of relative price of the export and an increasing function of foreign income. That is, EX t =e g1 t Y t (2.14) whereg 1> 1. The market clearing condition for the capital market, bond market and foreign exchange market are characterized as K t = k d t + k f t = K t1 + I t (2.15) B t = b d t + b f t (2.16) NX t +CF t = 0 (2.17) where NX t =e g1 t Y t −c m t and CF t =e 1 t h R k t k f t1 R t b f t1 i + p k t + p b t . The other market clearing conditions − deposits market, imported consumption goods market, and labor market − are characterized by equations (2.3), (2.4), and (2.5), respectively. The resource constraint is as usual. Y t = C d t +F(I t ;K t1 )+ I t + G+ Ex t (2.18) 69 2.3.1.3 Inspecting the Mechanism Using the constructed simple model, I illustrate the two different transmission mechanisms by which risk-on/off shocks cause large fluctuations in financial markets and the real economy in EMEs. The first channel is the “capital market channel that changes in domestic capital prices, driven by risk-on/off shocks, impact the asset side of domestic bank balance sheets, and impact the real economy subsequently. The second channel is the rather conventional “exchange rate channel” that local currency depreciation or appreciations impact the liability of the banks. What is new in the exchange rate channel is that the shocks to the foreign exchange market are ignited by LC debt capital outflows and the impacts are amplified by the deleveraging of the domestic banks, if the domestic banks have some net foreign currency debts. The analytical results introduced below match the empirical findings in the last section, and provide ways to “interpret” the correlations observed in the empirical exercise. Mechanism without foreign currency debt: Capital market channel Now I illustrate the mechanism in the model by which risk-on/off shocks cause fluctuations in financial markets and the real economy in small open economies. In particular, I exclude foreign currency debts in the model so that I can separately describe the channel of how capital flows disrupt the economy without currency mismatches. Since the channel has not extensively been pioneered in the liter- ature despite few recent works of similar mechanisms such as Caballero and Simsek (2020) and Devereux and Yu (2019), I name the channel “capital market channel.” To describe the mechanism, at first I need to explicitly solve the market clearing condition for the capital; N t f Q t = k d t + p t e t Q t = k f t = K t1 + I t . Plugging into the first order condition of the capital producer to the market clearing condition, I can solve for the equilibrium price of the capital. Q t = (1j)+ q (1j) 2 + 4j N t f+p k t e t K t1 2 (2.19) 70 where N t =s (z t + Q t )k d t1 R t d t1 +x(z t + Q t )k d t1 . Since the RHS includes Q t , the equilib- rium capital price is the fixed point of equation (2.19). Taking the derivative of Q t with respect to n t , the risk-on/off shocks gives ¶Q t ¶n t j e t = 0 B B B B @ e t s 1j 1 K 1 t1 2 + 4j N t f+p k t e t K 1 t1 d p k t dn t 1 C C C C A | {z } First Foreign Demand Shock 0 B B B B @ 1 (s+x)k d t1 f s 1j 1 K 1 t1 2 + 4j N t f+p k t e t K 1 t1 1 C C C C A | {z } Second Fire Sale 1 < 0 (2.20) If d p k t dn t < 0 as it should be. 23 With other conditions that I impose in the proposition 1, I can show the risk-off (on) shocks result in falls (booms) in the capital market. To understand the mechanism, notice that there are two types of different investors; domestic banks and global investors. Given other states, risk-off shocks derive down the demand for the capitals from the global investors. For the price to be maintained, the other investor, domestic banks should increase their demands, but it is not possible due to the leverage constraint. Hence, for the capital market to be cleared, the capital price must fall: the foreign demand shock in the first term in RHS in equation (2.20). The lower capital price in turn hurts the balance sheet of the domestic banks. To see it, notice the numerator in the second term in RHS in the equation (2.20) is (s+x)k d t1 = dN t dQ t Thus, the term is the marginal impact of capital price changes on the net worth of the bank. The banks whose net worth get damaged are forced to deleverage and therefore the capital price falls even more, as it is revealed in the second term in the RHS in equation (2.2019): the negative effects 23 The risk appetite shocks change the expected returnE t h e t e t+1 R k t+1 i ;but if only consistent change along with the directions in the risk appetite shock is d p k t dn t < 0. For example, if d p k t dn t > 0 due to the changes in the expected return E t h e t e t+1 R k t+1 i ;more foreign capitals will inflow into the domestic equity market, which raises the current capital price; equivalently raise the capital price so lower the expected return, which is a contradiction. 71 of the risk-off shock are amplified through a form of fire sale mechanism. 24 Now I introduce the first proposition in this paper, summarizing the result above along with other analytical results. Proposition 1 (Capital market channel) Assumec 1 k =c 1 b = 0, then we have 1) Risk-off (on) shocks cause falls (booms) in capital markets. That is, dQ t dn t < 0. 2) If(s+x)z t k d t1 <sR t d t1 , then capital demands from domestic banks increase in the cap- ital price. That is, dk d t dQ t > 0, and therefore dk d t dn t < 0 3) Assume j f(s+x)k d t1 + K t1 j 1 1 2 > 4K t1 sR t d t1 (s+x)z t k d t1 . Then impact of risk appetite shock increases in the share of global investors in the capital market b q t = p k t e t N t L+p k t e t , given exchange ratee t . That is, ¶ 2 Q t ¶n t ¶ b q t j e t < 0. Therefore, we have ¶ 2 k d t ¶n t ¶ b q t j e t < 0. The first statement confirms the discussion above. The second statement describes the fire-sale mechanism. Capital price falls lower the net worth, but also increase the amounts of the capitals the domestic banks can purchase given net worth. If the fixed amounts of payments to depositors are more than the net dividends from the capital, the “purchasing power” of the domestic banks always increases in the capital price. In other words, the capital demand curve is upward sloping. Then, any shift in the demands from the global banks generates amplification effects through the capital price. This is depicted in figure 6 below. The left panel in the figure shows the first impact of a risk-off shock on the market. A shrink of foreign investment shifts left the demand curve, putting downward pressures on the capital price. Then, as shown on the right panel, the falling capital price decreases the capital demands from the domestic banks, and so does the capital price, because the capital demand from domestic banks increases in the capital price; the demand curve is upward-sloping. The comparative statics in the third statement matches the empirical results of the cross-country panel regressions. The marginal impacts of risk-on/off shocks on the equity markets increase in the shares of foreign investors in the markets. Intuitively, the risk appetite shocks are the shocks to 24 In a deeper level, a reason why the risk-off shock results in firs sales lies in the bahaviors of the banks. Banks finance by issuing debts and invest in risky assets and therefore any unexpected changes in the risky asset prices impact the net worth of the banks. Bocola and Lorenzoni (2020) pioneered the microfoundation of such a type of contract and its macroeconomic implications. 72 Figure 2.4: Capital Market Channel ∆ ∆ ∆ ∆ ∆ the demand from global investors and then it is straightforward that the magnitude of the shocks depends on how many other domestic investors (banks in my model) exist in the market or how large the demands from foreign parties are compared to domestic investors: when there are more domestic investors compared to foreign investors in the same market, it must be easier for the domestic investors to absorbs the sell-off from global investors. As a result, the impacts of risk- on/off shocks on the capital price and accordingly the capital demands, investments, increase in the share of the global investors in the market. For more detailed and analytical analysis, I refer readers to the appendix. We have analyzed the capital market equilibrium while taking exchange rates as given. The exchange rates will be also heavily affected by the risk appetite shock. Before I illustrate the results, I remind readers that I have not introduced foreign currency debts in the model so that local currency depreciation does not induce negative effects by itself. In addition, in the analysis of exchange rate, I assume c 0 i = 1 for the analytical purpose, in contrast to what I assumed in the proposition 1. Of course, the qualitative results and the underlying intuition are not altered depending on the different assumptions. 73 Recall p k t = 1 G(q i )e v t h E t h e t e t+1 R k t+1 i R m t+1 (v t ) i as I assumedc 0 k =c 0 b = 0. I find it is convenient to formulate p k t and p b t as below. p i t = 1 Ge v t S i t where i2(k;b) and S i t =E t h e t e t+1 R i t+1 R m t+1 i . Then d p k t dn t is as follows. d p k t dn t =p k t + 1 Ge v t dR k t dn t dR m t+1 dn t ! where R k t = e t e t+1 R k t+1 and see dS k t dn t = dR k t dn t dR m t+1 dn t . For the simplicity, I assume as follows. Assumption 1. dR k t dn t > dR m t+1 dn t > dR b t dn t and therefore dS k t dn t > 0> dS b t dn t This assumption allows me to have a nice closed form without concerns about the expectation. Intuitively, although the return to the risky capitals rises following a risk-off shock, if expected returns to all risky assets in the world rise, then global investors will not allocate more funds to the capitals in the small open economy. However, the rise in the global portfolio return is not necessarily lower than the rise in the return to the capital investment in the small open economy. Hence, the assumption is adopted to derive a clean result conveniently, and of course, I will remove the assumptions in a more general model in the next section. From the foreign exchange market clearing condition, we can derive the equilibrium exchange rate as below. e t = e t c m t + R k t k f t1 + R b t b f t1 e t p k t + p b t Y t ! 1 g (2.21) Taking a derivative ofe t with respect ton t and a manipulation gives de t dn t e t = Y t h 1 dS b t dn t S b t h b t + 1 dS k t dn t S b t h k t + e t1 e t dR k t dn t g Y t h k t1 i + dc m t dn t Y t ge g1 t +h k t +h b t c m t (2.22) 74 whereh k t = p k t Y t andh b t = p b t Y t . Unlike the capital price Q t where I take the exchange rate as given, I here take Q t as a function of then t on purpose. I can solve for de t dn t more explicitly and then I can easily show de t dn t > 0 under the assumptions that I impose in the proposition. The mechanism behind it is straightforward. For the risk-off shock, capital outflows driven by the shock reduce foreign currency liquidity in the foreign exchange market so that the price of the foreign currency, the exchange rate rises; local currency depreciates. What is also interesting is de t dn t decreases in Y t , which is the base of foreign demands for exporting goods from the small open economy. Given the same trade openness, Y t proxies the size of the economy, GDP.h k t and h b t measure (or proxies) the equity external liability to GDP ratio and the LC debts to GDP ratio respectively. Note 1 dS b t dn t S b t > 1, but 1 dS k t dn t S b t < 1 because of the assumption 1. That is, the “coeffi- cient” in front ofh b t , 1 dS b t dn t S b t , is larger than 1, while the coefficient ofh k t is smaller than 1. Recall d p b t dn t p b t = 1 dS b t dn t S b t and equivalently for 1 dS k t dn t S b t . Hence, the larger coefficient of h b t means d p b t dn t p b t > d p k t dn t p k t . In other words, during a risk-off event, the bond portfolio investments outflow more than the equity (capital) portfolio investments. It is directly driven by the assumption, but we can interpret the result intuitively. The increases in the return to the LC bond in foreign currency mostly come from local currency depreciations, whereas the increases in the return to the capital come from both capital price falls and local currency depreciations. Then in a risk-off event, the return to the LC bonds (in foreign currency) cannot increase as much as the capital, equity. More intuitively, higher expected return due to equity price falls incentivize the equtiy foreing investors to stay in the market. Now, I can show the rate of marginal depreciation (appreciation) due to risk- off (on) shocks increases in h b t , but it is inconclusive for h k t . Intuitively, higher h b t means more bond portfolio investment capital outflows compared to the size of the economy. To explain more, the local currency sell-off of global investors in the foreign exchange market must be absorbed by the foreign currency suppliers in the market, exporters. If there is too much capital outflows for the exporters to take up, the local currency must depreciate; the price of local currency must fall. In contrast, whether the rate of marginal depreciation (appreciation) increases or decreases inh k t 75 is subtle since the equity portfolio investment capital flows are less sensitive to the risk appetite shocks as I assume dR k t+1 dn t > dR m t+1 dn t ; the falls in the capital market attract more global investors by gifting them higher expected returns. This explains why we cannot see the significance of equity portfolio investment to GDP ratios in the exchange rate regressions. Furthermore, capital price fall also reduces capital outflows; lower capital price lowers amounts of the capital outflows so as to give less local currency depreciation ( dR k t dn t < 0). Ifh k t1 th k t , as it should be in the data, then the equation (2.22) will be de t dn t e t t Y t h 1 dS b t dn t S b t h b t + 1 dS k t dn t S k t + e t1 e t dR k t dn t h k t i + dc m t dn t Y t ge g1 t +h k t +h b t c m t (2.23) Then it is even more clear why the equity-GDP ratio interaction terms are not significant in the exchange rate regression. The “coefficient” in front ofh k t is much smaller thanh b t or it can be even a negative number. As a result, I have shown the transmission mechanism of how the risk appetite shocks change financial markets in EMEs in an environment where the external liabilities of the EMEs are eq- uities or LC debts. I also provided comparative statics matching the results in the regressions. I summarize the theoretical findings in the proposition below. Proposition 2 Assumption 1 holds and assume further ge g1 t 1w > dS b t dn t S b t h b t and p b t 1 dS b t dn t S b t > p k t 1+ dQ t d(p k t e t) w 1w j 1 k f t1 1 dS k t dn t S k t , then we have 1) For dS k t dn t S k t small enough, dQ t dn t < 0. 2) Risk-off (on) shocks depreciate (appreciate) the local currency. That is, de t dn t > 0. 3) Givene t , impact of risk appetite shock increases in LC debts to GDP ratio, but the impact can either increase or decrease in the equity external liability to GDP ratio. In addition, the marginal impact is always larger for LC debts, . Thast is, if I define de t dn t e t h t h b t ;h k t ¶h t ¶h b t > 0> ¶h t ¶h k t or ¶h t ¶h b t > ¶h t ¶h k t > 0 76 The first statement is to assure the finding in proposition 1 and the other statements summarize the impacts of risk appetite shocks on foreign exchange markets. Of course, the impacts on the financial markets propagate into the real economy. Thanks to the simple structure in the model, we can easily characterize the impacts on the real economy in the corollary below. Corollary 1 Assume that dI t dn t + ¶EX t ¶e t de t dn t > 0. Risk-off (on) shocks lower (raise) investments and raise (lower) net exports. That is, dI t dn t < 0 and d(NX t ) dn t > 0 The assumption is made to rule out a peculiar case and it is stronger than necessary. The corollary captures typical reactions of small open economy to risk appetite shocks: risk-off shocks result in falls in investments, while raising net exports. In the case of risk-off shocks, weaker demands from global investors and less financial intermediations of the domestic banks due to the lower asset prices altogether induce less funding from households and foreign investors to the domestic corporate sector, which subsequently diminishes investments of the corporates. The impacts of the risk appetite shocks on net exports are less obvious. The easiest way to see the comparative statics is to look at the foreign exchange market equilibrium condition, NX t +CF t = 0. Also, it is easy to see that higher exchange rates reduce imports through the “intra- substitution” effects and elevate exports. 25 The relationship between exchange rate and net export here is based on the assumption of Producer Currency Pricing (PCP), and a more realistic assump- tion is Local Currency Pricing (LCP) or Dominant Currency Pricing (DCP). However, the impact on net export above is driven though the equality between the capital account balance and current 25 On the contrary, the intertemporal substitution effects and income effects are subtle. Whether the interest rate on the deposit R t+1 will rise or fall is subtle. The capital outflow from the government bond market pushes up the interest rate, while the savings by households given interest rate increase or decrease, depending on the income effects. In terms of income effect, the local currency depreciation and the drop in capital price generate positive income effects by reducing payments to global investors. However, the local currency depreciation and low capital price cause higher expected “rents” for the global investors so as to increase payments to global investors in the future; the mechanism is somehow similar to Fanelli and Straub (2019). On the other hand, low capital stocks in the future also generate negative income effects and the negative effects will be larger as the shock is more persistent, as we can reasonably assume. 77 account balance. In the simple model, I maintain the PCP assumption to focus on the key insights and implications of the “pricing to market” in the context of risk appetite shock transmission will be pioneered in an extended model, which is introduced in appendix, and in a more general model in the next section. Another important observation in corollary 1 is that the risk-on/off shocks cause two opposing effects on GDP. A risk-off shock decreases investments (lower aggregate demands), but at the same time, the shock increases net exports (higher aggregate demands). GDP in this simple model is invariant to the risk-on/off shock because of the absence of nominal rigidity, but the two opposing effects become more clear in a model with nominal rigidity, as I will show in the next section, limiting the impacts of risk-on/off shocks on GDP. This mechanism echos the findings in Blanchard et al. (2016) in that the paper also suggests two opposing effects of capital flows on small open economy. In their paper, non-bond capital inflows generate domestic booms through lower rates on the non-bond assets, but the inflows decrease exports of the small open economy as the inflows appreciate the local currency, 26 dampening the effects through the non-bond assets. It is important to be aware of the two opposing effects to adequately assess the quantitative importance of risk- on/off shocks for the real economy in EMEs. We will back to this point in the next section. Exchange rate channel Now I study the traditional exchange rate transmission channel. In en- vironments where financial corporates have sizable net foreign currency debts, risk-appetite shocks naturally lead to local currency depreciation so as to dampen the balance sheets of the corporates in EMEs. Such exchange rate channel has long been studied in the literature and is at the cores of re- cent influential papers. 27 Although the key mechanism is the same in this paper, the local currency depreciation, and subsequent deleveraging of domestic banks are initiated by capital outflows in the form of equities or LC debts. Also, the exchange rate channel interacts with the capital market channel, forming a negative loop mechanism of the risk appetite shocks. 26 In fact, the model mechanism itself is similar with Blanchard et al. (2016) in that both my model and the model in Blanchard et al. (2016) assume imperfect substitutability between different assets and constrained foreign investors. 27 See Aoki et al. (2018), Akinci and Queralto (2019), and Bocola and Lorenzoni (2019) 78 To add foreign currency debts to the model, I now assume that domestic banks can borrow abroad in the form of foreign currency debt. Let’s denote local currency debt and foreign currency debt by d t and d t respectively. In addition, R t+1 denotes the borrowing rate on foreign currency debts. Then, the bank balance sheet is Q t k d t = N t + d t +e t d t (2.24) For notational convenience, I define D t d t +e t d t e R t+1 (e t+1 ) R t+1 d t d t +e t d t + e t+1 e t R t+1 e t d t d t +e t d t Then, the net worth is N t =(s+x)(z t + Q t )k d t1 s e R t (e t )D t1 +Q e t1 d t1 ;D t1 (2.25) whereQ e t1 d t1 ;D t1 is the management cost of foreign currency debts, which I will describe below. See that e R t (e t ) does increase in the exchange rate. That is, the debt burden after the realization of the exchange rate rises as the local currency depreciates. Then from (2.25), it is easy to see local currency depreciation (highere t ) dampens the net worth of the bank. The marginal impact of risk appetite shock on the capital price is dQ t dn t = p k t sR t+1 d t f X t N t ; p k t e t de t dn t + e t X t N t ; p k t e t d p k t dn t ! | {z } First Foreign Demand Shock 1 (s+x)k d t1 f X t N t ; p k t e t ! | {z } Second Fire Sale 1 < 0 (2.26) where X t ()= r (j 1) 2 + 4j N t f+ K 1 t1 . If p k t (1s)R t+1 d t f < 0, local currency depreciation dampens the net worth of domestic banks so as to expedite capital price fall. 79 Similarly, I can characterize the impacts of risk appetite shocks on the exchange rate in envi- ronments where domestic banks have net foreign currency debts. The equilibrium exchange rate is characterized as follows. e t = 0 @ e t c m t + R k t k f t1 + R t b f t1 e t p k t + p b t e t h d t Q t (v t ) ; v t R t d t1 i Y t 1 A 1 g (2.27) In equation (2.27), it is important to notice that the foreign currency borrowing d t depends on Q t . Intuitively, lower capital price induces deleveraging of the banks and accordingly less borrowing abroad. To see it more clearly, let’s look at the optimal borrowing decision of domestic banks. Because of the leverage constraint, the only optimal decision of the domestic bank is to choose between domestic deposits and foreign currency debts. To characterize the foreign currency bor- rowing explicitly, I borrow an assumption from Aoki et al. (2018). Let’s suppose domestic banks face management cost of foreign currency borrowing. Q(e t d t ;D t )= y 2 x 2 t D t (2.28) Q() is the management cost and x t is the foreign currency debt ratio e t d t d t +e t d t . To make it more tractable, I assume the management cost will be paid next period; Q e t d t ;Q t k d t is paid in time t+1. Then the foreign currency borrowing will be d t = N t (f 1) E t h R t+1 e t+1 e t R t+1 i ye t (2.29) 80 Therefore, the foreign currency borrowing increases in N t . Intuitively, as the bank deleverages due to negative shocks to its own capitals, the bank does not need to borrow from either of de- positors or foreign investors, thereby reducing foreign currency borrowing; less foreign currency supplies to the foreign exchange market. Now we can derive the comparative statics de t dn t . de t =dn t e t = Y t h h b t 1 dS b t dn t S b t +h k t + e t1 e t dR k t dn t g Y t h k t1 i + dc m t dn t ¶d t ¶Q t dQ t dn t Y t ge g1 t +h k t +h b t c m t + d t R t d t1 e t ¶d t ¶n t > 0 (2.30) Since ¶d t ¶Q t > 0 and dQ t dn t < 0, the falls in the capital price due to risk-off shocks amplify local currency depreciation. 28 As a result, the falling capital price and rising exchange rate interact with each other, forming a negative loop mechanism, as illustrated in figure 7. To illustrate the mechanism in figure 7, let’s again think of a risk-off scenario. The risk-off shock induces falls in the capital price and the local currency, higher exchange rates as we saw above. Now think of the equilibrium in the capital market and foreign exchange market separately. In the capital market, the risk-off shocks shift left the capital demand curve, given exchange rate e 0 t . Similarly, the risk-off shock shift left the net capital inflow curve, given the capital price Q 0 t . Given the initial capital price and exchange rate alternatively in the capital and foreign exchange market, the “imaginary” equilibrium in the capital market and foreign exchange market move from e 0 to e 1 . Then, the higher exchange rate obviously raises the real debt burden of the domestic bank, which reduces the net worth of the banks and in turn forces the banks to buy less capitals, as it is in equation in (2.30). In the same way, the lower capital price expedites the deleveraging of the banks; banks are forced to take less deposits and borrow less in foreign currency debts, reducing the capital inflows. As a result, both the capital demand curve and the capital inflow curve shift left further, resulting in even lower capital price and higher exchange rate in equilibrium e 2 in figure 7. I summarize this finding in the second proposition, and the process is illustrated in the figure. 28 The statement should be understood as effects other than through the capital price, the term dQ t dn t k f t1 81 Figure 2.5: Market Crashes from Risk-Off Shocks − , − Proposition 3 Suppose domestic banks have positive net foreign currency debts, that is d t > 0. Then we have 1) Local currency depreciation (appreciation) lowers (raises) net worth of domestic banks. That is, ¶N t ¶e t < 0. 2) Impact on the capital price is amplified through the exchange rate and the impact on the exchange rate is amplified through the capital price. That is, given p k t (1s)R t+1 d t L< 0 and holding other states,j dQ t dn t j increase in de t dn t , and de t dn t increases inj ¶d t ¶Q t dQ t dn t j The statements in the proposition summarize the discussion above. An important note is that although the quantitative effects from the currency mismatch (in an EME whose liabilities are mostly foreign currency debts) may be larger, the capital price fall driven by foreign investors’ disposal of the assets does a role in igniting the negative feedback loop. In other words, outflows in the LC portfolio investments work as a trigger of the negative feedback loop. One discrepancy between the prediction from the model and my empirical results is that there is no statistically significant effect of sizable net foreign currency debts of nonfinancial corporate sectors in the empirical results, while the model predicts these debts should matter. While I cannot completely resolve the discrepancy, I suggest an extended model to explain the insignificance. The 82 extended model borrows some insights from the literature of pricing in international trade. Many of the nonfinancial corporates in EMEs are exporters and the prices in the exporting goods are denominated in key currencies such US dollar, while many of their costs, like wages, are denomi- nated in local currency. Then local currency depreciation boosts the profitability of exporters; costs are given, whereas the revenues from exports in local currency increase. If more foreign currency debts are positively associated with more benefits to exporters from local currency depreciation, higher net foreign currency debts of nonfinancial corporate sectors do not necessarily lead to higher fragility. I refer interested readers to the appendix. 2.3.2 Microlevel Evidence of the Capital Market Channel The capital market channel is newly introduced in this paper although few preceding papers have a similar mechanism in their models. While there are plenty of evidence regarding the exchange rate channel using micro-level data 29 , evidence of the new channel from micro level data has not been reported, to the best of my knowledge. Thus, I provide evidence proving the existence of the capital market channel, using bank balance sheet data in Korea. The model in this paper is a representative agent model and there is only one bank (or numerous identical banks) in the model. However, in reality, there are different banks with different exposures to the shocks to capital markets from global financial shocks. Abusing the implications from the model, the model predicts that, given impacts on the domestic capital markets, banks whose assets are more centered on risky financial securities and leverages are higher should be more impacted than others: for example, a risk-off shock will force almost all domestic banks to reduce their risky asset holdings, but the magnitude should be larger for banks with more risky assets and higher leverages. To test the “hypothesis,” I deploy the balance sheet data of Korean financial intermediaries. Similarly with many countries, certain “investment bank” type financial intermediaries in Korea have important roles in equity and bond markets, capital markets. I can access the data provided 29 See Baskayaa et al. (2017) and Hardy (2018). 83 Table 2.3: Capital Market Channel Regressions (1) (3) (2) (4) (5) (6) (7) Dln A R i;t1 -0.140 † -0.114 † -0.112 † -0.115 -0.110 -0.105 -0.104 [0.089] [0.074] [0.075] [0.078] [0.778] [0.074] [0.074] c i;t 1.701** 1.417** 0.608 0.881 0.955 0.904 [0.819] [0.603] [0.941] [0.674] [0.673] [1.31] A R i;t1 N i;t1 c i;t 1.973** 1.928* 2.047** 2.205* 2.423* 2.515* [0.961] [0.922] [0.976] [1.186] [1.450] [1.472] C i;t1 A i;t1 c i;t 3.706 3.660 3.607 4.009 [3.200] [3.263] [3.190] [3.308] S i;t1 A i;t1 c i;t -2.519 -2.936 -2.949 [4.720] [4.885] [4.890] size i;t1 c i;t -44.276 -59.178 [63.946] [66.958] A R i;t1 N i;t1 -0.041** -0.041** -0.041** -0.037** -0.032** -0.032** [0.016] [0.016] [0.016] [0.015] [0.015] [0.015] C i;t1 A i;t1 -0.029 0.003 -0.023 -0.020 [0.345] [0.346] [0.344] [0.342] S i;t1 A i;t1 -0.119 † -0.181** -0.184** [0.074] [0.081] [0.083] size i;t1 -1.812 -1.814 [1.237] [1.229] Bank Dummyc i;t 3) 2.693* 1.504 Country FE Yes Yes Yes Yes Yes Yes Yes Time FE Yes Yes Yes Yes Yes Yes No Observation # 909 899 899 899 899 899 899 R-squared 0.093 0.123 0.124 0.124 0.128 0.133 0.133 # of banks 35 35 35 35 35 35 35 Note: 1) *** p¡0.01, ** p¡0.05, * p¡0.1, * p¡0.1, † p¡0.15. 2) Sample periods 2005 Q1 ˜ 2016 Q4 23 S: Total financial security holdings, N: Net worth, C: Cash or Cash alike assets, A: Total asset, and A R : the total assets, excluding cashes and tangible assets. 3) Bank Dummy indicates that the interaction of the the gains/losses and dummy variable of investment banks owned by commerical banks. 84 by the regulatory in Korea. The data of course includes basic information of each of the invest- ment bank; e.g., total asset, net worth, and liabilities. Further, the data includes more detailed information on the composition of the asset and liabilities of the investment banks. The data shows how much foreign (mostly USD) currency debts or Korean won debts each investment bank has. More importantly, the data shows the composition of assets of each investment bank; the assets are classified as corporate bonds (in different categories), government bonds, equities, loans, cash or cash-like assets, tangible assets such as buildings, and so on. 30 Deploying the available information and avoiding more complexity, first I estimate how the risk-on/off shock impact the price of different securities in Korea, which are held by investment banks. Then I compute the exposure of each investment bank to risk on/off shocks. After esti- mating the exposures, I finally how the asset growth of each investment bank is affected by the exposure and leverage of each investment bank. I relegate the detailed estimation procedure to the appendix and put the estimation equation below. First I estimate the price elasticities of different securities with respect to unexpected changes in VIX. ln(Q i;t )= c i +d i [ln(VIX t )E t1 [ln(VIX t )]]+e i;t (2.31) Thend i [ln(VIX t )E t1 [ln(VIX t )]] gives me changes in the price if security i due to changes in VIX. Using it, I can estimate the gains or losses of capitals of each investment bank from risk- off/on shocks. I computed as follows. c i;t =q 0 i;t1 d[ln(VIX t )E t1 [ln(VIX t )]] (2.32) where d is the vector of the price elasticities and q i;t1 is the vector of the different securities holdings (denominated by the total asset excluding cash alike assets and tangible assets, A R i;t below) 30 It would be more ideal to have security level information; for example, equity or corporate bond of what firms held by each of the investment bank. Unfortunately, I do not have such detailed information and I am still seeking for more detailed information at the time of this writing. 85 of the investment bank i. 31 Hence, c i;t measures capital gains or losses of the total asset of each bank due to risk-on/off shocks. Then the regression equation is formulated as below. Dln A R i;t =a i +b 0 c i;t +b 1 A R i;t1 N i;t1 c i;t +G 0 0 z i;t1 c i;t +G 0 1 l i;t1 +e i;t where A R i;t is the total assets excluding cash or cash alike assets and tangible assets, A R i;t N i;t1 is the risky assets to net worth ratio (hence the leverage ratio 32 ) of the investment bank i, z i;t1 is the vector of other balance sheet conditions such as the ratio of risky assets to total assets, andl i;t1 is the balance conditions including A R i;t N i;t1 and z i;t1 : The result is reported in table 6. As expected, the coefficientb 1 of the interactive term between the “effective” leverage and the gains/losses due to VIX changes is positive and highly significant. I included another interaction terms, capital gains or losses interacted with z i;t1 jbankcharacteristicsobservableinthebankbalancesheetdata:T heseresultsarerobusttothedi f f erentcontrols: The empirical analysis captures the propagation of global financial shocks in the capital market. Often, macro-financial literature is interested in credit market (loan market) as the credit market is often larger than the capital market even in countries where capital markets are well developed. Appendix E introduces another empirical analysis to see how the global financial shocks are prop- agated from the capital market to the credit market in Korea. A simple illustration of the analysis is that commercial banks in Korea often finance in whole sale funding markets where the investment banks provide short term funds 33 to the commercial banks. When the investment banks cut down their fund supplies in the market due to negative impacts on their net worth, the commercial banks have trouble financing from the whole-sale funding market, and then subsequently reduce their credit supplies to the real sector. 31 The asset holings are the data at the end of last period. 32 Actually, substantial part of the total asset is cash or cash alike assets such as deposits at commercial banks. However, the leverage ratio in the model is more like risky assets to net worth ratio. Hence, I use the financial securities to net worth ratio. 33 It includes call loans, repo transactions, and most importantly short term bank debentures. 86 2.4 Quantitative Exercise In this section, I introduce a medium scale new Keynesian model to conduct quantitative exercises. The purpose of introducing the DSGE model is to quantify the importance of the capital market channel. In other words, in a small open economy where much of external liabilities are equities or LC bonds, how much variation does global financial shock generate in the financial markets and the real sector in the economy? For this purpose, I augment the standard new Keynesian small open economy model with the key features in the simple model; leverage constrained domestic banks and global investors who purchase capitals and LC bonds in the small open economy. Although the model is much more general and richer than the simple model in the last section, I abstract from several important features in reality, which are often important in the analysis of business cycles in EMEs. The model is still designed to study the transmission of global financial shocks to emerging markets, not to study business cycles in EMEs. 2.4.1 Environments Most environments in the model are same as the simple model, except for the nominal rigidity in goods price, more sophisticated bank leverage, and incomplete exchange rate pass-through in export price. For some of the specifications and notations, I follow the influential paper, Aoki et al. (2018). Goods producers Following the standard in the literature, final goods are produced by the re- tailer under perfect competition, and each of the differentiated intermediate goods is produced by an exclusive producer under monopolistic competition. Same as the simple model, the producers use the Cobb-Douglas technology. The difference from the simple model is the producers use imported intermediated inputs. Hence, the production function is y i;t = A t k i;t1 a k a k m i;t a m a ;m l i;t 1a k a m 1a k a m 87 where a k , a m and a k +a m 2(0;1). A t is a TFP in this economy and follows a AR(1) stochastic process. I introduce nominal rigidity following Calvo. In each period, a producer can adjust her price with a probability of 1k. Accordingly, each producer chooses the reset price P t to maximize expected discounted profits subject to the restriction on the adjustment frequency. The first order condition is given by E t "( ¥ å j=0 k j L t;t+ j P t P t+ j h h 1 mc t y i;t+ j )# = 0 whereL t;t+ j is the stochastic discount factor of the representative households, and mc t is the real marginal cost. I skip the steps to derive the equation as I followed the standard in the literature. The real marginal cost is a result of the cost minimization problem and is formulated as below mc t = 1 A t z a k t e a m t w 1a k a m t From the law of large numbers, the aggregate price level is characterized as below. P t = h (1h)(P t ) 1h +h(P t1 ) 1h i 1 1h (2.33) Households The representative households are identical to the simple model in that they con- sume both domestic and imported consumption goods. But, in the medium-scale model, there is disutility of labor so that the labor supply is endogenous. Also, unlike the simple model, house- holds cannot invest in government bonds directly. The investment has to be done through interme- diations by domestic banks. The optimization problem of the representative households is max f c d t+ j ;c m t+ j ;L t+ jg ¥ j=0 E t " ¥ å j=0 b t+ j U C d t+ j ;C m t+ j ;L t+ j # sub ject to C d t+ j +e t+ j C m t+ j + D t+ j +t t+ j w t+ j L t+ j + R t+ j D t+ j1 +p t+ j 88 Now the per period utility function is given by U C d t ;C m t = ln H C d t ;C m t 1 1+z L 1+z t where H C d t ;C m t is the CES composite. H C d t ;C m t = w C d t m1 m +(1w)(C m t ) m1 m ! m m1 w controls the share of imports in consumption while m is the elasticity between domestic goods and imported goods. Note that all the terms in the budget constraint are denominated in the price of domestic goods. For example, w t and R t are the real wages and real interest rates on the deposits respectively, in terms of domestic goods. In the same way,e t is the price of imported goods, but is not necessarily the same as the terms of trade as I deviate from the law of one price. Since the aggregate CPI index is not the same as the price of domestic goods because of the imported goods consumption, e t is not the real exchange rate. However, the movement ofe t is qualitatively the same as the real exchange rate as I fix the foreign price and even quantitatively similar with the true real exchange rate as long as the weight on the imported goods is relatively small. The optimality conditions of the households are identical to the simple model, except for the new labor-leisure condition. w t ¶U t ¶C d t = L z t Capital producers Instead of the adjustment cost of the investment to capital ratio, I used the adjustment cost by which the investment cost varies with the investment growth. The objective of the capital producer is to choose max f I t+ jg E t å L t;t+ j Q t+ j I t+ j 1+F I t+ j I t+ j1 I t+ j 89 with F I t I t1 = j 2 I t I t1 1 2 The optimality condition of the capital producer is as below. Q t = 1+F I t I t1 + I t I t1 F 0 I t I t1 E t " L t;t+1 I t+1 I t 2 F 0 I t+1 I t # (2.34) Long-term bond To model the impacts of sell-off of global investors in bond markets in EMEs, I model the government bond as a perpetual bond. 34 I mostly follow the specification in Gertler and Karadi (2013). The perpetual bond pays one unit of domestic consumption goods every period. Denoting the real bonds price by q t , the return to the bonds including the capital gains/losses is R b t+1 = 1+ q t+1 q t Therefore, the bond investment is risky as its return changes along with the bond price in the next period, q t+1 . Domestic banks Domestic banks, which refer to financial intermediaries in different forms in reality, take deposits from households or borrow abroad, to invest in domestic capital and gov- ernment bonds. What are different from the simple model are 1) domestic banks can invest in government bonds as well as the capitals, and 2) the leverage constraint as a function of the prof- itability in the futures, as in Gertler and Kiyotaki (2010) and Aoki et al. (2018). Before illustrating the banks in the model, I note that I adopt the leverage constraint in Gertler and Kiyotaki (2010) because I want to show how the key insights survive and bring quantitative results in the environment that is widely used in the literature. It does not mean the approach in the paper is more precise than others. As long as banks in the model are leverage constrained, 34 The perpetual bond here is like the inflation-adjusted bond, like TIPs in the US. I can model it as the nominal perpetual bond, but I find the volatility of the bond price in the model simulation is too high, comparing with the observed volatility in data. However, of course, whether bond strip is nominal or real does not meaningfully alter the quantitative results. 90 qualitative results should be the same, but different modeling gives a different quantitative result. Precise specifications of leverage constraints on financial intermediaries are beyond the scope of this paper, and how the impacts of risk-on/off shocks vary with different modeling of the leverage constraint is for future research. 35 Similarly with the banks in the simple model, the banks purchase capitals or government bonds and finance the investments through the deposits of the households, foreign borrowings, and their net worth. Therefore, the balance sheet of a typical bank is Q t k d t + q t b d t = d t +e t d t + n t (2.35) Accordingly, the evolution of the net worth of a bank is n t =(z t + Q t )k d t1 + R b t q t1 b d t1 R t d t1 e t R t d t1 Q x 2 t ;D t (2.36) whereQ x 2 t ;D t = z d 2 x 2 t D t and x t = e t d t d t +e t d t , the foreign currency debt ratio to the total debts and thusQ x 2 t ;D t is the management cost of the foreign currency debt. The evolution of the net worth with the exit of incumbent bankers and the entry of new bankers is 36 N t =(s+x) (z t + Q t )k d t1 + R b t q t1 b d t1 s R t d t1 +e t R t d t1 x 2 t ;D t 35 Another possible approach is Value at Risk (VaR) constraint on financial intermediaries, as I modeled global investors in this paper. Adrian and Shin (2013), Nuno and Thomas (2017), and Coimbra and Rey (2020) model the banks facing a form of VaR constraint. The approaches in the strand of the literature seem to closer to the risk managements of financial intermediaries in reality, and it potentially gives a stronger result to me. However, I follow the approach in Gertler and Kiyotaki (2010) because the goal in this paper is not a precise identification of the leverage constraint. 36 For the convenience, I include the bonds, R b t q t1 b d t1 , in the start-up funds. We can think of the start-up funds as a fraction of the total financial assets held by domestic agents. Of course, exclusion of the bonds from the start-up fund does not meaningfully change any results in this paper. 91 The key idea in Gertler and Kiyotaki (2010) is the continuation value should be larger than the fraction of the total assets that the banks can divert. The continuation value is defined as the sum of present values of the future dividend as V t =E t " ¥ å j=1 L t;t+ j (1s)s j1 n t+ j # The continuation value can be reformulated in a recursive form as V t =E t [L t;t+1 [(1s)n t+1 +sV t+1 ]] (2.37) The leverage constraint arises due to the following moral hazard problem. After raising funds, the banks can decide whether to operate honestly or divert assets for personal use. To divert means to secretly channel funds away from investment in order to consume personally. Specifically, the banker can divert q fractions of the total capitals and4q of the government bond. Government bonds are harder to divert; government bonds are more transparent and have better legal protec- tions. Reflecting those features, I let42(0;1). Then, the bank’s problem is reduced to comparing the continuation value, V t , to the gains from diverting the funds. That means, if the continuation value is less than the gain from diverting funds, the banks cannot raise any outside financing. Therefore the following incentive constraint must be satisfied. V t qQ t k d t +Dqb d t (2.38) The optimization of the banks is to maximize (2.37) subject to (2.36) and (2.38). Since the solution of the maximization problem is well known, I introduce the solutions of the banks as follows. f t = E t L t;t+1 W t;t+1 R t+1 q E t L t;t+1 W t;t+1 R k t+1 R t+1 (2.39) E t L t;t+1 W t;t+1 R k t+1 R t+1 D= E t L t;t+1 W t;t+1 R b t+1 R t+1 (2.40) 92 whereW t;t+1 reflects the shadow value of one unit of net worth to the bank in each state at time t+ 1. Hence,L t;t+1 W t;t+1 is stochastic discount factor of the banks. In the equilibrium, the marginal cost of the foreign currency debt, the expected interest rates on foreign currency debts in local currency and the adjustment cost, discounted by the stochastic dis- count factors of the banks, must be the same as the interest rates on the deposits. The optimization of the bank characterizes the foreign currency debt as follows. d t = D t E t h L t;t+1 W t;t+1 R t+1 e t+1 e t R t+1 (v t ) i ye t (2.41) R t+1 (v t ) is the borrowing rate on the foreign currency debts of the domestic banks. Rather than modeling the determination of the interest rate R t+1 , I assume that the interest rate will react to the risk-appetite v t . More specifically, R t+1 (v t )= 1+ r e c d v t where r e c d v t is the time-varying net interest rate and c d 2(0;¥). v t is the time-varying risk- appetite, same with the simple model. Thus, v t follows an AR(1) process in the equation (2.12). We can think of the interest rate as EMBI spread in reality, which is strongly correlated with VIX index; higher VIX is correlated with higher EMBI spreads. Again, I note that I abstract from the endogenous determination of the interest rate, but recent studies such as Morelli et al. (2019) showed that the interest rates on foreign currency sovereign bonds of EMEs are heavily affected by the bond demands from global banks. Considering the influence of the risk-appetite of global investors on the borrowing rates of EMEs, such a reduced form approach is a way to include necessary ingredients without setting up another optimization problem. 93 Global investors Global investors purchase capitals and LC bonds in the small open economy, and their decisions are made by the equations. p k t = Q t k f t e 1 t = 1 Ge v t c 0 k +c 1 k E t e t e t+1 R k t+1 R m t+1 (v t ) p b t = b f t e 1 t = 1 Ge v t c 0 b +c 0 b E t e t e t+1 R b t+1 R m t+1 (v t ) Now I lift the assumptions thatc 0 k andc 0 b are zero. In addition, it is not necessary that dR k t dn t > dR m t+1 dn t as I estimate the medium-scale model. Similarly with R t+1 (v t ), the return to the global portfolio reacts to the risk-on/off shocks. R m t+1 (v t )= 1+ r m e c ;m v t We can think R m as yields on the BAA grade corporate bonds in the US, and similarly with R , risk-off (on) shock raise (lower) the return to the global portfolio. Government Government is just identical to the simple model. The budget constraint of the government is G=t t + q t B t R g t q t1 B t1 I abstract from the problem of the government and fiscal policy. The supply of government bond is fixed at B. Hence B t = B. Export In the simple model, I adopted the producer currency pricing (PCP) in export pricing for tractability. However, of course, it is counterfactual and the trade literature comes to a consensus that exporters can set a different price in the foreign markets (Local Currency Pricing, LCP) or the price of tradable goods are in general priced in key currencies like USD (Dominant Currency Pricing, DCP). 37 Since the risk-on (off) shocks in my model cause local currency appreciation 37 See Betts and Devereux (2000) for LCP and Gopinath and Stein (2020) for DCP. 94 (depreciation), it is important to model the export pricing in a realistic way to assess the quantitative impacts of the shocks on the small open economy. For the purpose, I make an assumption of the export pricing, following Wang (2018). Denote the export price in the foreign market by p ex t . Then p ex t is p ex t = e 1 t l p ex t 1l wherel2(0;1). Ifl = 1, the export pricing follows a perfect PCP. In contrast, ifl = 0, it indicates a perfect LCP or DCP. p ex t is the exogenously given price of the exports; for example, the price of competitors in the foreign markets. Such a ”reduced form” approach to the export pricing makes the model simple, but also allows tractability. The reality obviously lies in somewhere between LCP and PCP and accordingly I can set a reasonable parameter value for l reflecting empirical evidence. Moreover, I can experiment on how the transmission of GFS varies along with different export pricing policies by setting different values ofl in the DSGE model. The export is EX t =(p ex t ) 1g Y t (2.42) where g > 1 and Y t = Y t e T t . T t is a AR(1) stochastic process, which captures trade shocks, i.e., shocks to the demand for exporting goods from the small open economy. Monetary authority The monetary authority conducts policy using a nominal interest rates rule. The nominal rate i responds to the deviation of inflation from target,p t relative top, which is one in this model. In addition, I assume that the authority tends to avoid drastic changes in the nominal interest rate. As a result, the interest rate rule is characterized as i t = i+(1r i )w p (p t 1)+r i i t1 i + m t (2.43) wherer i 2(0;1) andw p > 1, and m t is the monetary policy shock in this model. 95 Resource constraint The output is divided between consumption, investment, government con- sumption, export and foreign currency debt management cost. The economy-wide resource con- straint is thus given by Y t = C d t + 1+F I t I t1 I t + G+ EX t +Q x 2 t ;D t (2.44) where Y t = R 1 0 y h1 h i;t di h h1 . The net output, GDP of this economy is the total output minus the imported intermediate inputs. Y net t = Y t e t M t (2.45) where M t = R 1 0 m h1 h i;t di h h1 . 2.4.2 Calibration I calibrate the model to the Korean economy because it is an ideal example in the context of the analysis in this paper, as most of the external liabilities in Korea are equities and LC bond portfolio investments. To capture the short-run dynamics, I set one period to a quarter in reality. For most of the parameters in the model, I used standard values in the literature or values reported in well-known preceding studies. For some parameters regarding trade openness or output to capital ratio, I calibrate the parameters to match the observed ratios in Korea. The parameters I newly calibrated in this paper are the parameters about the global investors and global financial shocks, which are new components in this model. Assigned parameters First, I explain the parameters I set externally. For those parameters, I mostly followed Akinci and Queralto (2019), Aoki et al. (2018), Gertler and Karadi (2013), and few others. I set the discount factor, b, to be 0.9925 so that the annual interest rate is 3%. This corresponds to the discount rate used in Akinci and Queralto (2019) for their emerging market bloc in their two country model. It also approximately matches the real interest rate in Korea before 96 the global financial crisis in 2008. For the labor supply parameters, I set the Frisch elasticity to be 0.33 following Gali and Monacelli (2005); therefore the inverse of the Frisch elasticity, z is 3. The elasticity between the domestic goods and imported goods is 1.5 (m = 2), same as Akinci and Queralto (2019). This is in the range of standard values in the literature. I set the parameter of imported goods 1w to 0.225. This value corresponds to the foreign goods and service consumption to GDP ratio in Korea. 38 On the production side, I calibrate the capital share and imported intermediate goods share to Korean economy. The calibrated values of the capital sharea K and imported goods sharea M are 0.25 and 0.225 respectively. The elasticity of demand from the aggregator,h, is 9 following Aoki et al. (2018). For the inverse elasticity of net investment to the capital price, capital depreciation rate, and Calvo parameter (probability of keeping the price constant), I followed the standard values in the parameters. I set the investment adjustment cost parameter to be 2.85, following Akinci and Queralto (2019). This value is in the range of conventional values in the literature. 39 I set the elasticity of export demand,g, to be 2.5 so that the elasticity is similar to the elasticity of domestic demands for imported consumption goods. An important parameter is the exchange rate pass-through,l. I used the value reported in Gopinath and Burstein (2014). The exchange rate pass through from local currency to USD is 0.2 Reflecting on the consensus that most of tradable goods are in fact denominated in USD, I setl to be 0.2. For the parameters of the domestic bank, I mostly follow Aoki et al. (2018) and Gertler and Karadi (2013). The bank survival rate s and the fraction of the total financial assets to the new bankersx are set to 0.94 and 0.462 respectively, following Aoki et al. (2018). The proportion of divertible capital to the total capital, q 0 is set to 0.34. This value is close to Gertler and Karadi (2013). 40 With the parameter values, the spread between the return to the capital and deposit rates 38 The consumption of imported goods and service includes the expenditures made abroad by residents in Korea, such as traveling abroad or tuitions for students studying abroad. 39 This is little higher than the value in Gertler and Karadi (2013), 1.728. It is to capture more realistic volatilities of the capital price and investments so that the generated second moments are closer the volatilities of stock index and investment in Korea. 40 In Aoki et al., the propotion of divertable assets depends on the ratio of foreign currency debts to the total assets. However, I have no reasoning or empirical evidence to support it. 97 in deterministic steady state is very close to 0.02 annually. This is the target used in the calibration in Gertler and Kiyotaki (2010) and Gertler and Karadi (2013). In this paper, it is important to have realistic capital to GDP ratios in the model as the impact of risk-on/off shocks depends on the foreign investors’ share in the capital market. The capital to GDP (annual GDP) ratio in the steady state is close to 2 and it is close to tangible assets to GDP ratio in Korea, after excluding residential real estates from the tangible assets. The parameter of the advantage of government bond in terms of leverage,D, is set to 0.5, following Gertler and Karadi (2013). Besides the global financial shocks, there are three exogenous shocks in the model, TFP shock, export shock and monetary policy shock. For both TFP and export shock, I set the autocorrelation parameter to be 0.9, which lies in the range of the values used in the literature. I set the standard deviation of the TFP shock to be 0.004 and 0.01 for the export shock. These are smaller than the values used in the emerging market literature. There are two reasons I use relatively low values. First, the financial amplification mechanism in my model amplifies the impacts of TFP and export shocks and therefore feeding standard values into the model will make the model economy more volatile than the real economy in reality. More importantly, I calibrate the model to Korean economy, which is stable compared to typical example countries in the emerging market literature, such as Mexico or Brazil. The standard deviation of the calibrated model economy is 0.015, which is a little higher than the true standard deviation of quarterly GDP in Korea, 0.012, in the sample period 2001-12. The standard deviation of monetary policy shock is 0.001 so that the unexpected changes in policy rate is 0.4% annualized rate. The monetary policy shock is serially uncorrelated as the shock will be persistent by the Taylor rule in equation (2.43) I set the parameters in the Taylor rule, following Aoki et al. (2018). Regarding the government, I set government consumption to GDP and government debt to GDP ratio to be 0.2 and 0.45 respectively, same as Gertler and Karadi (2013). These ratios are also close to the observed ratios in Korea. 41 I summarize the assigned parameters in table 8. 41 The government debt to GDP ratios are higher than government debt to GDP ratios in Korea, but it is close once we inlude the monetary stabilization bonds issued by the central banks in Korea, Bank of Korea. 98 Table 2.4: Assigned Parameters Parameter Symbol Value Discount factor b 0.925 Inverse Frisch elasticity of labor supply z 3.000 Trade elasticity m 2.000 Share of imported goods 1w 0.225 Capital share a K 0.250 Imported intermediate goods share a M 0.240 Elasticity of demand from the aggregator h 9.000 Capital depreciation rate d 0.025 Inverse elasticity of net investment to the capital price j 1.728 Probability of keeping the price constant k 0.779 Government consumption to GDP ratio in steady state G Y net 2.000 Government bond to GDP ratio in steady state qB Y net 0.450 Inflation coefficient in the Taylor rule w p 1.500 Persistence coefficient in the Taylor rule r i 0.800 Elasticity of export demand g 1.500 Exchange rate pass-through l 0.200 Fraction of capital that can be diverted q 0.340 Leverage advantage in seizure rate of government bond 4 0.500 Transfer to the entering bankers x 0.046 Survival rate of the bankers s 0.940 Management cost for foreign currency debt y 0.111 Estimated parameters I need to estimate the parameters of the global investors and the related global financial shock because these are novel components in the model in this paper. Recall the equity and local currency bond investment of the global investor and the determination of foreign currency debt. p k t = Q t k f t e 1 t = 1 G k e v t c 0 k +c 1 k E t e t e t+1 R k t+1 R m t+1 (v t ) (2.46) p b t = b f t e 1 t = 1 G b e v t c 0 b +c 1 b E t e t e t+1 R b t+1 R m t+1 (v t ) (2.47) 99 d t = D t E t h L t;t+1 W t;t+1 R t+1 e t+1 e t R t+1 (v t ) i ye t (2.48) where R m t+1 (v t ), R t+1 (v t ) and v t are as follows. R m t+1 (v t )= 1+ r m e c ;m v t R t+1 (v t )= 1+ r e c v t v t =r v v t1 +n t I let R m and R be the 5 years BAA corporate bond yields in the US and JP Morgan Emerging Market Bond Index (EMBI Index). I estimate c ;m and c d by regressing those interest rates on Cboe VIX index. I relegate details of the estimations to the appendix. The estimated values ofc m and c are 1.04 and 1.37 respectively. 42 I used the values in quarterly data and also considered the standard deviations of VIX and global financial shock process in my model, which I describe below. I estimate the parameters in equation (50) and (51), using GMM. Notice that once I get the ratio of c 0 j to c 1 j , I can easily computeG j based on the observed equity liability to GDP and LC bond to GDP ratios. Target moment in the GMM estimation is the growth of the portfolio investments. I relegate the detail of the estimation to the appendix. Estimated c 1 k c 0 k and c 1 b c 0 b are 4.621 and 9.785. These values reflect that global investors do not strongly respond to the arbitrage opportunity. Then it is easy to computeG k andG b from the data. I setG k andG b to match the equity portfolio investment to GDP ratio (0.28, on average in 2012 - 18) and Korean won bond portfolio investment to GDP (0.08, on average in 2012 - 18). 43 42 The regression might suffer from autocorrelation in R t and v t . I discuss this issue in a separate section in appendix, and show the results do not change much in another empirical identification, which is relatively free from the concern of autocorrelation. 43 This includes local currency deposits and I counted as the deposits are mostly held by foreign investors in Korean won bond market. The reason I excluded it in the regressions is that in some countries like India, the deposits are much held by residents abroad, who are actually citizens of the emerging market country. 100 Table 2.5: Estimated Parameters Parameter Symbol Value Target Stickness of equtiy portfolio investment c 1 k c 0 k 4.645 g p k Stickness of LC bond portfolio investment c 1 b c 0 b 9.363 g p b Inverse of funds allocated to the capital market G k 0.053 LCE Y net Inverse of funds allocated to the bond market G b 0.180 LCB Y net Elasticity of global portfolio return to risk-on/off shock c m 1.046 BAA Elasticity of foreign borrowing rates to risk-on/off shock c 1.372 EMBI Standard deviation of risk-on/off shock s n 0.090 s p k Autocorrelation of the risk-appetite r n 0.850 r p k There are handful papers, which estimate global financial shocks from different risk assets over the real world. The approach in those papers is beyond the scope of this paper. Instead, I calibrate related parameters in the simplest way. In the model simulations, global financial shocks matter through capital flows. Motivated by this, I set the standard deviation and autocorrelation to match the observed standard deviation and autocorrelation of the foreign equity portfolio investments in Korea. The observed standard deviation of the equity portfolio investment (in the US dollar) is 0.2 and the autocorrelation is 0.834. To capture the autocorrelation, I setr v to bo 0.85. With Th valuer v = 0.85, the autocorrelation of simulated p k is 0.84, which is close to the data. To have the standard deviation of 0.2, I need the standard deviation ofn t at 0.104. However, this might yield a too high standard deviation since the observed equity portfolio investment includes changes in stock price, which is more volatile than the Tobin-Q in the model. Therefore, I take a litter lower value: I set the standard deviation of the risk-appetite shock to be 0.09. There are four shocks in the model, global financial shock, TFP shock, export demand shock, and monetary policy shock. Except for monetary policy shock, the other three shocks are certainly correlated in reality. The correlation among the shocks are not crucial in my analysis, but I let the shocks are correlated, depending on the purpose of the model simulation. If the shocks are correlated, I set the correlation between global financial shock and export demand shock to be 0.6 and global financial shock and TFP shock to be 0.3. 101 2.4.3 Results 2.4.3.1 Transmission of Global Financial Shocks in Korea I simulated the model, using standard techniques. 44 First, I illustrate how much fluctuations in financial markets and the real economy can be generated by global financial shocks in the model economy calibrated to Korea, where most of the external liabilities are equities and Korean won denominated debts. I opened the discussion in the paper with the comovements of stock indices and exchange rates with VIX. Furthermore, proposition 1 predicts a risk-off (on) shock causes a fall (rise) in stock price, rise (fall) in exchange rate, fall (rise) in investment, and rise (fall) in export. To see whether the model can generate such patterns in the simulated data, I simulate the model with four ”uncorrelated” shocks. Figure 7 below confirms that the model can generate dynamics corresponding to the theoretical prediction. As it is clear in the figure, a risk-off shock causes a fall in Tobin-Q and a rise in the real exchange rate (Korean won depreciation). Some discrepancies between the simulated path and the data are Tobin Q seems to be stable compared with the stock index in Korea. Surely, Tobin Q in this DSGE model cannot seriously replicate the volatility of the stock index in reality. The relative stability of simulated Tobin Q is also attributable to the features in the model in that the global risk-appetite process in the model probably misses sudden big falls in the risk-appetite in reality. 45 In terms of impacts on the real economy, it is observable in the figure that a risk-off shock causes a fall in investment and a rise in export. As it will be clear in the impulse response analysis, simulated VIX lags behind the simulated investment due to the features in adjustment cost. Export varies with VIX, but the exports do not react to global financial shock strongly as the exchange rate pass-through is low in the calibration (l=0.2). 44 I solved the model using dynare in third order approximation. I also used the pruning technique built in dynare. As discussed in Brunnermeier and Sannikov (2014), such pertubation techniques miss some of the non-linear dynamics in a model of financial amplification mechanism. However, I limit my attention to a normal business cycle, not a big crisis event. Use of the pruning technique can create some inaccuracy in the estimation, but it is unavoidable to prevent the spurious explosive path as discussed in Dou et al. (2017). 45 Perhaps, it is ideal to include some jump process in the risk appetite. But, this is beyond the scope in this paper. 102 Figure 2.6: Simulated Capital Prices and Exchange Rates with VIX 0 20 40 60 80 100 120 140 160 180 200 -10 -5 0 5 10 % -40 -20 0 20 40 % Tobin Q (Left) Real Exchange Rate (Left) VIX (Right) 0 20 40 60 80 100 120 140 160 180 200 -10 -5 0 5 10 % -40 -20 0 20 40 % Investment (Left) Export (Left) VIX (Right) Next, I examine how the key macro variables react to a risk-off shock. I give a risk-off shock of one standard deviation to the model economy. In figure 8 below, I show the response of the key financial variables such as capital price, bond price and real exchange rate, and the key real variables such as consumption, investment, export and GDP. I also show responses of important endogenous variables, which are important to understand the mechanism. Those variables are capitals and bonds held by global investors and net worth and leverage of the domestic banks. To highlight the importance of the leverage constrained banks in the model, I also simulate another model in which there is no domestic financial friction, but otherwise is identical to the baseline model. In figure 8, the ”baseline” indicates the results from the model with the leverage constrained domestic banks and the ”frictionless” indicates the results from the model without banking sectors. First, I describe the impulse response functions from the baseline model. The responses of the variables are all as expected. Tobin Q, capital price, falls by nearly 0.8%, and the bond price fall is slightly lower (0.5%). The capital outflows obviously depreciate the local currency and 103 the depreciation rate is close to 0.6%. The impulse response of other variables illustrates the mechanism behind the falls in the capital and bond price. The risk-off shocks induce sell-off of the global investors in domestic financial markets, as shown in the impulse responses in the response functions of capital and bond held by global investors. The sell-off lowers the capital price, which in turn lowers the net worth of the domestic bank and raises the return to the capital investment; the gross return increases by 0.3%. The higher expected return raises the leverage, but it is not enough because of the leverage constraint rooted in the agency problem of the banker. 46 As a result, the demand from domestic banks cannot increase enough so that the capital and bond price fall significantly. The fall in the capital price simultaneously happens with the fall in investment; investment falls by nearly 0.5%. Because of the technological features in the adjustment cost, the fall in the investment peaks one period after the shock. 47 Consumption falls, and it is mainly due to the falls in the consumption of imported goods . On the other hand, the local currency depreciation increases the exports despite the low exchange rate pass-through, and accordingly, the falls in GDP is relatively mild; it falls by 0.2%. Comparison of the results from the baseline model to the frictionless model highlights the importance of leverage constraint on domestic banks. First, one can easily notice that the negative impacts on domestic financial markets, falls in the capital price and bond price, are much smaller than the baseline model. Accordingly, falls in investment are much smaller as well. This is because, in the frictionless environment, the expected returns to capital and government bond are dictated by the household Euler equations. That is, if capital price and bond price falls and thus the expected return to the investments rises, then it immediately results in more saving from households as they expect higher returns. Despite higher borrowing rates on the foreign currency debts, the households are incentivized to borrow more in foreign currency to invest in domestic assets as the local currency is expected to appreciate. As a result, increased investments by households 46 The rate of leverage increase is slightly lower than the rate of net worth decrease, as predicted in proposition 2. 47 It is a typical observations in a DSGE model with “investment” adjustment cost. Once I replace the investment adjustment cost with the capital adjustment cost, the hump shape response disappears. 104 Figure 2.7: Impulse Response Functions 0 10 20 -1 -0.5 0 % Tobin Q 0 10 20 -0.4 -0.2 0 % Bond q 0 10 20 0 0.2 0.4 0.6 % Real Exchange Rate Baseline Frictionless 0 10 20 0 0.1 0.2 % Return to Capital 0 10 20 -6 -4 -2 0 % K f 0 10 20 -6 -4 -2 0 % B f 0 10 20 -4 -2 0 % Net Worth 0 10 20 0 2 4 % leverage 0 10 20 -0.4 -0.2 0 0.2 % Consumption 0 10 20 -1 -0.5 0 0.5 % Investment 0 10 20 -0.1 0 0.1 0.2 % Export 0 10 20 -0.2 -0.1 0 0.1 % GDP 105 significantly offset the decrease in investments by global investors, preventing large falls in the asset prices and the investment. In contrast, the impact on the FX market is almost the same in the two models, and similarly for export. Falls in consumption are slightly higher for the frictionless model since the households in the model are incentivized to save and invest more. As a result, GDP falls are slightly higher for the baseline model. The purpose of the quantitative analysis is to evaluate the quantitative importance of the capital market channel. To quantify the importance in a different way, I compare the impulse response functions above to the other small open economy, in which everything is identical, but most of the external liabilities are foreign currency debts. 48 Suppose another economy where foreign currency debt to GDP ratio is 28% and equity liability to GDP, LC bond to GDP ratios are 5% each. Again two economies are almost identical, except for the composition of external liabilities. For nota- tional convenience, we call the economy of equity and LC bond external liability “LC” economy and call the other economy “FC” economy. The different responses of the two different economies are introduced in figure 9. Surprisingly, it turns out that the two economies show quite similar responses to the one stan- dard deviation risk-off shock. My model is designed to evaluate the impacts of different types of capital flows on the small open economy, and therefore I abstracted from several features in EMEs, by which can amplify potential risks from foreign currency debt, such as sovereign default risk or country level collateral risk, as Bianchi (2011). Despite the limit, figure 9 shows that a large sell-off of global investors in domestic financial market can generate sizable falls in small open economies. To understand the similarity, notice that the financial amplification mechanisms in the two different economies are similar. In both economies, the negative impacts of risk-off shock are amplified through the negative balance sheet effects; i.e., negative pecuniary externality on the net worth of domestic bank. In LC economy, the sell-off of global investors directly causes the capital price fall and of course, it reduces the net worth accordingly. In FC economy, the higher exchange 48 The size of foreign currency debts is affected byf. To change the parameter value makes it hard to compare the results from the two different model economies. To avoid the confusion, I adjust the management cost. In the “FC debt dominated economy,” I change the management cost toQ(e t d t ;D t )= y 2 (x c) 2 t D t . I adjust c so that the marginal management costs in the two different economies are almost identical. 106 Figure 2.8: Impulse Response Functions, LC vs. FC 0 10 20 -1 -0.5 0 % Tobin Q 0 10 20 -0.4 -0.2 0 % Bond q 0 10 20 0 0.5 1 % Real Exchange Rate LC FC 0 10 20 0 0.1 0.2 % Return to Capital 0 10 20 -6 -4 -2 0 % K f 0 10 20 -6 -4 -2 0 % B f 0 10 20 -4 -2 0 % Net Worth 0 10 20 0 2 4 % leverage 0 10 20 -0.4 -0.2 0 % Consumption 0 10 20 -1 -0.5 0 0.5 % Investment 0 10 20 0 0.1 0.2 % Export 0 10 20 -0.2 -0.1 0 % GDP 107 rate caused by capital outflows raises the real debt burden and reduces the net worth. The lower net worth again put downward pressure on capital price. Regardless of whether the negative impacts are on the liability side or asset side, in both economies, financial market prices (capital price or exchange rate) hurt the net worth of the domestic banks. Despite the similarity, one can notice that the falls in the real economy are larger in FC. There are two reasons why the fall is larger for FC economy. First, in FC economy the capital price and exchange rate both work in a way of reducing net worth, while in LC economy, only capital price falls lower the net worth. Hence, the drops in net worth are larger in FC economy, and so the falls in investments in FC economy. Second, falls in the prices of capital and bond held by foreign (global) investors create positive income effects for the residents in the LC economy. The lower asset prices hurt the balance sheets of domestic banks, but at the same time, the lower asset prices (in local currency) and the local currency depreciation ”inflate away” the liabilities in the sense that values of the liabilities decline, when measuring the values in terms of export price or imported goods price. 49 This positive income effect upholds the aggregate consumption during a recession caused by the risk-off shock. 2.4.3.2 Quantitative evaluation of the importance of the capital market channel One of the important questions in the literature of Global Financial Cycle is “How important is the global financial cycle to peripheral economies, small open economy like Korea?” The answer to the question should vary country by country. Different countries have different features in their financial markets and the real economies, and thus the transmission and propagation mechanism in each of the economies should be different from others. Because of the difficulty, I limit my focus to the model economy calibrated to Korean economy and then evaluate the quantitative importance of GFS in financial and business cycles in Korea. I gave four different shocks to the model economy as I described in the calibration section. I set the four shocks to be all independent from each other 49 More precisely, the values of the liabilities decline in terms of tradable goods. If all the tradable goods are denom- inated in US dollar, as it is in DCP hypothesis, local currency depreciation devalue the local currency denominated assets in terms of tradable good. See Fanelli (2018) for more sophisticated analysis. 108 Table 2.6: Variance Decomposition Q t q t e t R k t R b t C t I t Ex t Y net t Global Financial Shock 52.9 30.4 36.6 59.3 42.8 18.0 27.6 2.9 10.6 Export Shock 1.0 2.4 30.8 1.2 1.6 4.6 1.0 94.6 9.2 TFP Shock 20.5 50.2 26.8 7.2 7.2 72.9 58.6 2.1 79.0 Monetary Policy Shock 25.6 16.9 5.9 33.3 48.4 5.5 12.8 0.5 1.2 for the accurate assessment. I simulated the model for 20,000 periods and dropped first 2,000 periods. The result of the variance decomposition is in table 10. The first observation from the variance decomposition is risk-on/off shocks account for large parts of the fluctuations in financial markets in the model economy. The risk-on/off shock accounts for more than half of the variations in Tobin Q (52.9%) and the return to the capital (59.3%), and 36.6% of the real exchange rate variation. Similarly for the bond market, the risk-appetite shocks explain 30.4% of bond price variations and 42.8% of the variations in return to the bond. Comparing to the financial variables, relatively small parts of the variations in real sector variables are attributable to the risk-on/off shocks. For the investment and the consumption (including both domestic goods and imported goods), 27.6%t and 18.0% are attributable to the risk-on/off shock respectively. For the export and GDP, only 2.9% of export variations and 10.6% of GDP variations are attributable to the risk-on/off shocks. The variance decomposition analysis provides reasonable results, but we should not take each of the point estimate too seriously. The small open economy model is designed to study the impact of global risk-appetite shock on EMEs through different capital flows. I added some necessary fea- tures, which are potentially important for quantitative analysis, but the DSGE model is still short of accommodating all the important ingredients. Different parameter values about the global financial shocks and the global investors can substantially change the result of the variance decomposition analysis. Nevertheless, the result is comparable to the recent study of global financial shocks and usual belief in financial market practitioners. Among the traders and commentators in Korean financial markets, a pervasive view is that a dominant factor in the market is the movements of 109 global financial market, especially the markets in the US. More importantly, a recent paper, Acalin and Rebucci (2020) empirically analyzed the importance of global financial shock in explaining the stock market movements and business cycles in Korea. They showed that approximately 50% of stock market variations are attributable to global financial shocks (GFS) and 10% of GDP vari- ations are attributable to the same GFS. 50 If their estimation is precise enough, then the calibrated model in this paper has generated realistic quantitative results. Despite the results quantitatively similar to Acalin and Rebucci (2020), I interpret the variance decomposition results in a different way. In their paper, the authors stated that the importance of the global risk-appetite shocks in the business cycles in Korea is limited as only 10% of GDP forecast errors are attributable to the shocks. However, the low number for GDP is because the falls in investments are largely offset by the rises of net exports. The parts of the variations attributable to the risk-on/off shocks are much larger for investments and consumption. Considering consumption is the variable most relevant to welfare, taking GDP as a criterion for the impacts of the global financial shocks on the real economy results in an underestimation of the importance of the shocks. A heuristic way to mute the increase in export is to let the export shock be negatively correlated with the risk-appetite shock, as it should be in the real world. Once the correlation between the export shock and the risk-appetite shock is set to -0.6, 18.7% of GDP variations are attributable to the risk-appetite shocks. 2.5 Concluding Remarks In this paper, I explored the channel through which global financial shocks, i.e., global financial shocks, are transmitted to small open economies, in particular EMEs. Motivated by the fact that nowadays, substantial parts of the external liabilities of many EMEs are actually local currency denominated portfolio investments, such as LC equities or LC bonds, I proposed the capital market channel. In an environment where domestic financial intermediaries are leverage constrained, 50 I note that their approach is much different from this paper. Besides the different methodologies, they computed how much of the forecast errors can be attributable to each of different shocks. Hence, their results is not directly comparable to the variance decomposition. 110 shifts in demands for domestic financial assets from global investors can cause drastic changes in asset prices, which in turn affect domestic financial intermediaries through the changes in their net worth. Risk-off (on) shocks lower (raise) the asset prices: the lower (higher) the asset prices, the weaker (stronger) the intermediations of the domestic financial intermediaries. The impacts on the financial markets and financial intermediations propagate into the real economy, mainly through capital investments. I also studied the conventional exchange rate channel. Here, I used a different approach wherein the local currency depreciation is sparked by capital outflows from the domestic bond market and the exchange rate channel interacts with the capital market channel, thereby producing more devastating effects of a risk-off shock. However, the traditional exchange rate channel seems to be weakened in the current states of EMEs. Financial intermediaries in EMEs have little exposure to changes in exchange rates as they are balanced between foreign currency debts and assets. Meanwhile, nonfinancial corporations have sizable net foreign currency debts in some EMEs, but they have foreign currency revenues from exports. A pervasive view of export pricing in the trade literature predicts the mark-up from exports to rise in domestic currency depreciation. The theoretical findings in this paper are supported by evidence of different layers. The cross- country panel regressions indicate that financial variables in an EME, namely, stock indices and exchange rates, tend to be affected by the global financial shocks more when the EME received more equity and LC bond portfolio investments. Using the model calibrated to the Korean econ- omy, quantitative studies have shown that global financial shocks are the dominant factor in the financial markets and also important for business cycles in Korea. Moreover, empirical analysis using bank balance sheet data in Korea evidenced the validity of the capital market channel in realistic environments. To conclude, all theoretical and empirical findings in this paper reveal that to a substantial extent, the risk-appetite shocks to global investors are transmitted to EMEs via fickle portfolio capital flows to equity and local currency bond markets in EMEs. More broadly, EMEs have a lesser concern about foreign currency debts, the previous cause of crises, as their borrowing ability 111 in equities and LC debts has improved. However, they simultaneously face a new risk from the new sources of external financing. I abstracted from several important features, in reality, to focus on the key question. The exter- nal assets by residents in EMEs were not added to the model. The existence of foreign currency assets abroad held by residents in EMEs can insulate the residents from exchange rate fluctuations, but it can stabilize or destabilize the economy through different channels. An important factor missed in this study’s analysis is the evolution of beliefs of both domestic and foreign investors in financial markets. Presumably, financial market booms (falls) caused by risk-on (off) shocks generate optimistic (pessimistic) beliefs among the domestic market participants. The interaction between changing beliefs and financial amplification mechanism will significantly amplify the quantitative impacts of the capital market channel. Meanwhile, another deep question related to this paper is “What is behind the original sin dissipation?” In this paper and the companion paper, I suggested related empirical regularities and theoretical explanations for the facts. However, that is far short of answering the deep question of the causes of the original sin dissipation. I believe all the issues above give us hard, but interesting questions unanswered in this paper. I leave these issues to future research. 112 References [1] Acalin, J. and Rebucci, A. 2020. “Global Business and Financial Cycles: A Tale of Two Capital Account Regimes.” Seoul Journal of Economics, 33 (3), 395-435. [2] Adrian, T. and Shin, H. S. 2014. “Procyclical Leverage and Value-at-Risk.” The Review of Financial Studies, 27 (2): 373–403 [3] Aghion, P., Bacchetta, P. and Banerjee, A. 2000. “ A Simple Model of Monetary Policy and Currency Crises.” European Economic Review, 44 (4-6): 728–738. [4] Aizenman, J. 2018. “A Modern Reincarnation of Mundell-Fleming’s Trilemma.” Economic Modelling, 1:1–11. [5] Aizenman, J., Binchi, M., and Hutchison, M. M. 2016. “The Transmission of Federal Reserve Tapering News to Emerging Financial Markets.” International Journal of Central Banking, 12 (2): 318–356. [6] Arellano, C., Bai, Y . and Mihalache, G. P. 2020 “Monetary Policy and Sovereign Risk in Emerging Economies (NK-Default).” NBER Working Paper, No. 26671. [7] Arslanalp, S. and Tsuda, T. 2014 “Tracking Global Demand for Emerging Market Sovereign Debt,” IMF Working Paper, WP/14/39. [8] Akinci, O. and Queralto, A. 2019. “Balance Sheets, Exchange Rates, and International Mon- etary Spillovers.” FRB of New York Staff Report, No. 849. [9] Alfaro, L., and Kanczuk, F. 2013. “Debt Redemption and Reserve Accumulation.” NBER Working PaperNo. 849. No. 19098. [10] Aoki, K., G. Beningno, and N. Kiyotaki. 2018. “Monetary and Financial Policies in Emerging Markets.” Manuscript, Princeton University. [11] Avdjiev, S., Binder, S., and Sousa, R. 2017. “External Debt Composition and Domestic Credit Cycles.” BIS Working Papers, No 627. [12] Avdjiev, S. and Hale, G., 2019. “U.S. Monetary Policy and Fluctuations of International Bank Lending” Journal of International Money and Finance, 95: 251–268 113 [13] Baskaya, Y . S, Giovanni, J., Kalemli-Ozcan, S., Peydro, J-L. and Ulu, M. F. 2017. “ In- ternational Spillovers and Local Credit Cycles.” Journal of International Economics, 108: S15-S22. [14] B´ en´ etrix, A. S., Lane, P. R., and Shambaugh, J. C. 2015. “International currency exposures, valuation effects and the global financial crisis.” Journal of International Economics, 96 (S1): S98–S109. [15] Beningo, G., Chen, H., Otrok, C., Rebucci, A., and Young, E. R. 2016. “Optimal Capital Controls and Real Exchange Rate Policies: A Pecuniary Externality Perspective.” Journal of Monetary Economics 84: 147–165. [16] Betts, C. and Devereux, M. 2000. “Exchange rate dynamics in a model of pricing-to-market.” Journal of International Economics 50 (1): 215–244. [17] Bianchi, J. 2011. “Overborrowing and Systemic Externalities in the Business Cycle.” Ameri- can Economic Review, 101 (7): 3400–3426. [18] Blanchard, O., Ostry, J. D., Ghosh, A. R. and Chamon, M. 2016. “Capital Flows: Expansion- ary or Contractionary?.” American Economic Review 106 (5): 565–69. [19] Bocola, L. and Lorenzoni, G. 2018. “Financial Crisis, Dollarization, and Lending of Last Resort in Open Economie.” Manuscript, Northwestern University. [20] Bruno, V . and Shin, H. S. 2015a. “Capital Flows and the Risk-taking Channel of Monetary Policy.” Journal of Monetary Economics, 71: 119–132. [21] — — . 2015b. “Cross-Border Banking and Global Liquidity.” The Review of Economic Stud- ies, 82 (2): 535–564. [22] — — . 2018. “Currency Depreciation and Emerging Market Corporate Distress.” BIS Working Papers, No. 753. [23] Burstein, A. and Gopinath. G. 2014. “International Prices and Exchange Rates.” Handbook of International Economics, 4th ed., 4: 391-451 [24] Caballero, R. J. and Krishnamurthy, A. 2003. “Excessive Dollar Debt: Financial Develop- ment and Underinsurance.” Journal of Finance, 58 (2):867–894. [25] Caballero, R. J., Farhi, E. and Gourichas, P-O. 2008. “An Equilibrium Model of “ Global Imbalances” and Low Interest Rates.” American Economic Review, 98 (1): 358–393 [26] Caballero, R. J., and Simsek, A. 2020. “A Model of Fickle Capital Flows and Retrenchment.” Journal of Political Economy, 128 (6): 2288-2328 [27] Calvo. G. 1998. “Capital Flows and Capital-market Crises: the Simple Economics of Sudden Stops.” Journal of Applied Economics, 1 (1): 33–54. 114 [28] Calvo, G., Izquierdo, A., and Talvi, E. 2006. “Phoenix Miracles in Emerging Markets: Recov- ering Without Credit From Systemic Financial Crises.” Inter-American Development Bank Research Department Working Paper, No. 570 [29] Calvo, G. and C. Reinhart. 2002. “Fear of Floating.” Quarterly Journal of Economics 117 (2), 379–408. [30] Cavallino, P. and Sandri, D. 2019. “The Expansionary Lower Bound: Contractionary Mone- tary Easing and the Trilemma.” Manuscript, Bank for International Settlements. [31] Cesa-Bianchi. A., Andrea Ferrero, A., and Rebucci, A. 2018. “International Credit Supply Shocks.” Journal of International Economics, 112: 219–237. [32] Cerutti, E., Claessens, S., and Rose, A.K., 2017a. “How Important is the Global Financial Cycle? Evidence from Capital Flows.” NBER Working Paper, No. 23699. [33] Christiano, L., Dalgic, H. C. and Nurbekyan, A. 2020. “Financial Dollarization in Emerging Markets: Efficient Risk Sharing or Prescription for Disaster?” Manuscript, Northwestern University. [34] Dalgic, H. C., 2020. “Corporate Dollar Debt in Emerging Markets.” Manuscript, University of Mannheim. [35] Devereux, M. B. and Yu, C. 2020. “International Financial Integration and Crisis Contagion,” The Review of Economic Studies, 87 (3): 1174–1212. [36] Du, W. and Schreger, J. 2016. “Local Currency Sovereign Risk.” Journal of Finance, 71 (3): 1027–1070. [37] Du, W.and Schreger, J. 2016. “Sovereign Risk, Currency Risk, and Corporate Balance Sheets.” Manuscript, Columbia Business School. [38] Du, W., Pflueger, C. E., and Schreger, J. 2016. “Sovereign Debt Portfolios, Bond Risks, and the Credibility of Monetary Policy.” NBER Working Paper, No. 22592. [39] Edwards, S. 2015. “Monetary Policy Independence under Flexible Exchange Rates: An Illu- sion?” NBER Working Paper, No. 20893. [40] Eichengreen, B. and Gupta, P. 2014. “Tapering Talk: The Impact of Expectations of Re- duced Federal Reserve Security Purchases on Emerging Markets.” World Bank Policy Re- search Working Paper, No. 6754. [41] Eichengreen, B., Hausmann, R., and Panizza, U. 2002. “Original Sin: the Pain, the Mystery, and the Road to Redemption.” Inter-American Development Bank Conference Paper. [42] Engel, C. and Park, J. J. 2018. “Debauchery and Original Sin: The Currency Composition of Sovereign Debt.” NBER Working Paper, No. 24671. [43] Fanelli, S. 2019. “Monetary Policy Capital Controls, and International Portfolios.” Manuscript, Princeton University 115 [44] Farhi, E. and Werning, I. 2014. “Dilemma not Trilemma? Capital Controls and Exchange Rates with V olatile Capital Dlows.” IMF Economic Review, 62:569–605. [45] Gabaix, X. and Maggiori, M. 2015. “International Liquidity and Exchange Rate Dynamics.” Quarterly Journal of Economics, 130 (3): 1369–1420. [46] Gal´ ı, J., and Monacelli, T. 2005 “Monetary Policy and Exchange Rate V olatility in a Small Open Economy.” The Review of Economic Studies, 72 (3): 707–734 [47] Gertler, M. and P. Karadi (2011). “A model of unconventional monetary policy,” Journal of Monetary Economics, 58, 17–34. [48] — — . (2013). “QE 1 vs. 2 vs. 3...: a framework for analyzing large-scale asset purchases as a monetary policy tool,” International Journal of Central Banking, 9: 5–53 [49] Gertler, M and Kiyotaki, N. 2010. “Financial Intermediation and Credit Policy in Busi- ness Cycle Analysis,” Handbook of Monetary Economics - Volume 3A, Elsevier Ltd, 3 (11): 547–599. [50] Gopinath, G. and Stein, J. C. 2020. “Banking, Trade, and the Making of a Dominant Cur- rency,” Quarterly Journal of Economics, qjaa036. [51] Han, X. and Wei, S. 2018. “International Transmissions of Monetary Shocks: Between a Trilemma and a Dilemma.” Journal of International Economics, 110: 205–219. [52] Hardy, B. 2018. “Foreign Currency Borrowing, Balance Sheet Shocks and Real Outcomes.” BIS Working Paper, No. 758. [53] Jeanne, O. and Korinek, A. 2010b. “Managing Credit Booms and Busts: A Pigouvian Taxa- tion Approach.” NBER Working Paper, No. 16377. [54] Jeanne, O., and D. Sandri. 2020. “Global Financial Cycle and Liquidity Management.” NBER Working Paper, No. 27682. [55] Jiang, Z., A. Krishnamurthy, and H. Lustig (2019a). “Dollar safety and the global financial cycle.” Stanford University Graduate School of Business Research Paper No. 19–16. [56] Levy- Yeyati, E. and Zu˜ niga, J. 2015. “ Varieties of Capital Flows: What Do We Know.” Harvard University, John F . Kennedy School of Government Working Paper, RWP15–025 [57] Maggiori, M. 2017. “Financial Intermediation, International Risk Sharing, and Reserve Cur- rencies.” American Economic Review 107 (10): 3038-3071. [58] Mendoza, E. G. 2002. “Credit, Prices, and Crashes: Business Cycles with a Sudden Stop.” In Preventing Currency Crises in Emerging Markets, ed. Jeffrey A. Frankel and Sebastian Edwards, 335–92. Chicago: University of Chicago Press. [59] Mendoza, E. G. 2010. “Sudden Stops, Financial Crises, and Leverage.” American Economic Review, 100 (5):1941–1966. 116 [60] Mendoza, E. G. and Bianchi, J. 2018. “Optimal Time-Consistent Macroprudential Policy.” Forthcoming, Journal of Political Economy. [61] Mendoza, E. G., Quadrini, V . and Rios-Rull, Jose-Victor. 2009. “Financial Integration, Finan- cial Development, and Global Imbalances.” Journal of Political Economy, 117 (3):371–416. [62] Miranda-Agrippino, S. and Rey, H. 2020. “ U.S. Monetary Policy and the Global Financial Cycle,” The Review of Economic Studies, rdaa019. [63] Morelli, J. M., Ottonello, P., and Perez, D. J. 2018. “Global Banks and Systemic Debt Crises.” Manuscript, New York University. [64] Nu˜ no, G. and Thomas, C. 2017. “Bank Leverage Cycles.” American Economic Journal: Macroeconomics, 9 (2): 32–72. [65] Nuno, C. and Rey, H. 2019. “Financial Cycles with Heterogeneous Intermediaries.” NBER Working Paper No. 23245 [66] Obstfeld, M. 2015. “Trilemmas and Trade-offs: Living with Financial Globalization.” In Global Liquidity, Spillovers to Emerging Markets and Policy Responses, edited by C. Rad- datz, D. Saravia and J. Ventura. Central Bank of Chile. [67] Ottonello, P. and Perez, D., 2019 “The Currency Composition of Sovereign Debt.” American Economic Journal: Macroeconomics, 11 (3): 174–208. [68] Rey, H., 2013. “Dilemma Not Trilemma: the Global Financial Cycle and Monetary Policy Independence.” Jackson Hole Economic Symposium. [69] Rey, H. 2016. “International Channels of Transmission of Monetary Policy and the Mundel- lian Trilemma.” IMF Economic Review, 64: 6 [70] Shin, H. S. and Shin, K. .2011. “Procyclicality and Monetary Aggregates.” NBER Working Paper, No. 16836. [71] Shin, H. S., Bertaut, C., and Bruno, V . 2020. “Original Sin Redux.” Manuscript, Bank for International Settlements. [72] Wang, O. 2018. “Exchange Rate Pass-Through, Capital Flows, and Monetary Autonomy.” Manuscript, New York University. 117 Chapter 3 International Reserve Accumulation: Balancing Private Inflows with Public Outflows (with Dongwook Kim) The massive international reserve 1 accumulation of Emerging Market Economies (EMEs) since the crises in the mid 1990s 2 has been one of the most debatable issues in the international macroe- conomics literature and among policy makers. While there are a few different explanations about the motivations of such large amounts of reserve holdings, one of the widely accepted views is that EMEs accumulate reserves as a buffer against sudden stops 3 in the future: this is the view so-called “Precautionary view.” Especially in the policy circle, holding adequate levels of reserve is regarded as the most important policy tool against sudden stops; IMF (2012). However, the academic litera- ture is still struggling to find the mechanism of how accumulating and holding reserves help EMEs with dealing with sudden stop. In the absence of an adequate theory of reserve accumulation in the precautionary view, we cannot find the answers to the following related important questions. “What is the adequate level of reserve holding for an EME?,” “How is it related to other policies such as monetary policy?” and “Is the role of international reserve substitutable by other policy tools such as capital controls?” 1 Throughout this paper, “reserve” means international reserves and we mostly use reserves instead of international reserve. 2 Mexico Tequila Crisis in 1994 and East Asian Crisis in 1997. 3 Sudden stop in the context of cross-border capital flows is defined as sudden reversal or stop of capital inflow to an Emerging Market Economy (EME), and the subsequent currency depreciation, which results in a serious balance sheet deterioration in the country. 118 In this paper, we suggest a novel mechanism of how reserve accumulation serves the goal in the precautionary view. Our view of reserve accumulation is that it is the capital outflow by public sectors corresponding to capital inflows. To understand it, let’s think of the following scenario. Imagine that capital inflows into an EME in the forms of foreign direct investment or equity port- folio investment suddenly increase for some exogenous reasons. Given all other states, it worsens the capital account and hence the current account as well. But, we usually do not think in such a way because we normally expect that in equilibrium, there may be corresponding capital outflows or the increased capital inflows may crowd out other forms of capital inflows. Here let’s assume further that the increased direct investment flows are so excessive that the total capital inflows will be larger even after the direct investment inflows crowd out other capital inflows. Then the EME needs to make capital outflows to maintain its macroeconomic balance. It can be done by private sectors if capital outflows from the economy are frictionless. However, if private sectors cannot invest abroad enough or do it inefficiently because of some reasons, for example, low expertise in the domestic financial sector, parts of overseas investments of the EME must be done by public sectors. The capital outflows by public sectors instead of private sectors are reserve outflows: the direct investment inflows cause a sort of “floods of foreign currency liquidity” and then central banks in EMEs accumulate reserves to “pump out” the liquidity. The main contribution of this paper is to suggest a new theoretical model, which formally illustrates the ideas in the scenario above. Although our main contribution lies in the new theory, our model is motivated by the empirical regularities we uncover in this paper. Our strategy to found empirical regularities is to look at the evolution of the structures of external liabilities and assets. As it is documented in Lane and Milesi-Ferretti (2007) and B´ en´ etrix et al. (2015), there have been structural changes in the size and the composition of external liabilities and assets of EMEs for the last 30 years. We link these changes to reserve accumulations. Similarly, we analyze the gross flows rather than focus on the net flows following the lessons from some influential papers such as Forbes and Warnock (2012) or Bruno and Shin (2015). Broadly speaking, the empirical regularities we found are 1) both of “inflows” and “stocks” of direct investment and equity portfolio investment 119 are strongly and positively correlated with both of “outflows” and “stocks” of international reserve, 2) reserve outflows (accumulations) are also strongly and positively correlated to current account surplus, and 3) reserve outflows are positively correlated to private sector capital outflows. The patterns described so far hold in both cross-country data and individual country time-series data. To summarize our empirical findings more simply, both reserve outflows and private sector capital outflows increase when EMEs receive more direct investment or equity portfolio investment capital inflows, or have higher current account surpluses. Based on the empirical regularities, we construct a three-period model to explain the facts qualitatively. Our modeling strategy is to put minimal ingredients into the baseline model in the capital control literature. The baseline model is the Fisherian deflation model developed in Bianchi (2011), Korinek (2018), and Jeanne and Korinek (2019). The class of the models captures the features of sudden stop in a simple way, and more importantly, private saving and borrowing in the model are inefficient due to the pecuniary externality, with which we can design a planning problem of the optimal reserve accumulation. The two ingredients added to the baseline model are 1) imperfect capital mobility for both debt inflows and debt outflows, and 2) gross capital flows including direct investments. We model the imperfect capital mobility in the idea of limited participation following Fanelli and Straub (2019) for debt capital inflows, and then model the imperfect mobility for the private sector debt outflows in a similar fashion. Our specification is an EME needs to rely on international financial inter- mediaries (IFIs) and the IFIs ask for more fees for the intermediations facing higher demands of overseas investment intermediation from the EME. 4 In this environment, more overseas invest- ments by private sectors lead to lower returns to the economy, but the private sector agents do not take account of it. Direct investments in our model are similar to equity portfolio investments or merger and acquisition in the reality. Foreign direct investors purchase claims on the capital. 4 We model in such a way to have a clean result. Our view is there are different sources of inefficiencies in private sector capital outflows from EMEs. We discuss this in section 3 and suggest a different microfoundation in the appendix. 120 In such environments, receiving direct investments instead of debt inflows makes the EMEs more robust to possible sudden stops in the future. However, when the direct investment inflows are beyond a certain level, the EME needs to save abroad since more direct investments mean more capital returns to the foreign investors in the future; it is better than external debt, but still a differ- ent form of external liability. The problem is private sectors cannot make enough capital outflows (overseas investments) by themselves and it is done inefficiently. Furthermore, the economy may be subject to insufficient net foreign assets to prepare for sudden stops, and the frictions on the capital outflows make it worse. As a result, in the absence of reserve accumulation, the excessive direct investment flows cause inefficient domestic currency appreciations and equivalently domes- tic consumption booms. Therefore, the social planner in the EME is incentivized to accumulate reserves to generate capital outflows as reactions to the massive capital inflows. In our model, there is no particular structure except for the two new ingredients. Unlike pre- ceding papers, we do not need any particular structures in the model such as longer maturity in the external debt than reserves, the existence of long-term project 5 or constantly binding credit con- straint. Furthermore, reserve accumulation in our model is a reaction to direct investment inflows and hence it is a function of direct investment and equity portfolio capital inflows. Hence, we can easily explain our empirical regularities of the relationship between direct investment external liability and reserve accumulation; once we posit direct investment inflows to EMEs as given, we can explain much of the increases in reserve holdings of EMEs since the mid-1990s and why some EMEs hold more reserves than others. In addition, the role of reserve accumulation in our paper is unique because the function of reserve accumulation is to supplement insufficient capital outflows by private sectors. In contrast to a few other papers, the reserve management policy — accumulat- ing and depleting reserves — in our model cannot be perfectly replaced by other policy tools such as capital controls. Our model also provides important implications about the debate of “currency manipulation”: Do some EMEs depreciate their currencies to boost their exports? While we cannot fully answer 5 This is required in the papers where the reserve is modeled in the structures of Diamond and Dybvig banking model 121 the question, our model implies that the amount of reserve holding is not a good litmus for the test of currency manipulation. Reserve accumulation in our model is a passive reaction to direct investment type capital inflows. In terms of currency valuation, the interventions to accumulate reserves prevent domestic currency appreciation so as to keep the current account balance from getting worse due to the capital inflows: but, there is no intention to depreciate the currency to boost exports. Moreover, reserve accumulation in our model looks like a managed floating exchange rate regime in the sense that the policy limits the currency appreciations although it is not an object itself. Hence we link the reserve accumulation to the managed floating literature initiated by Calvo and Reinhart (2002). Related Literature Our paper is related to several strands of literature in international macroe- conomics. First and foremost, our paper relates to the literature that studies reserve accumulation of EMEs. The main objective in the literature is to find why EMEs hold a huge amount of costly reserves. While there are a few different approaches, the literature broadly falls into two differ- ent views: Mercantilist view and Precautionary view. Articles in the mercantilist view argue that reserve accumulation is a byproduct of exchange rate policies aiming at boosting exports by de- preciating local currencies. Early works in the literature include Dooley et al. (2004), and there have been a few notable recent papers of similar ideas (Korinek and Serven, 2016; Benigno and Fornaro, 2014; and Choi and Taylor, 2017). On the other hand, the literature of the precautionary view pays attention to the historical fact that most EMEs began building their stocks of reserves after experiencing financial crises, in particular after the East Asian crisis in 1997. The papers in this view argue that policy makers in the EMEs have accumulated reserves in the belief that the reserves will protect them against sudden stops. Relatively early works worth listing in this fashion are Aizenman and Lee (2007), which analyzed the macroprudential role of reserves in the framework of the Diamond-Dybvig model, and Jeanne and Ranciere (2011), which tried to quantify the optimal reserve holding assuming reserve is a sort of Arrow-Debreu security. More recently, Bianchi et al. (2018) show how reserves can help EMEs with reducing roll-over risks 122 of external sovereign debts. 6 A recent work that shares similar insights with ours is Jeanne and Sandri (2018). The paper introduces a model where EME policy authorities accumulate reserves to complement insufficient capital outflows by private sectors and generating enough capital out- flows (having more liquid foreign assets) can stabilize domestic debt prices from volatile capital flows. We contribute to this literature by proposing a novel theory of reserve accumulation in the precautionary view. 7 We show that in environments where there are frictions on capital outflows, EME policy authorities facing large capital inflows are incentivized to accumulate reserves and it is not based on any mechanism in preceding papers. Although we share some insights with Jeanne and Sandri (2018), we construct our model using different microfoundations of frictions on capital outflows and we link reserve accumulation to the types of direct investment or equity portfolio inflows, which we will discuss more below. Our paper is also related to the papers that pioneered the positive correlations between FDI external liabilities and official reserve assets. The positive correlation between direct investment inflows and reserve accumulation is documented in Dooley et al. (2004). The paper insists that EMEs depreciate their currencies to attract direct investments so as to utilize otherwise wasteful resources in the economy, such as labor: EMEs accumulate reserves to attract direct investments. Matsumoto (2019) and Wang (2019) share similar ideas with Dooley et al. (2004) and the papers introduce a small open economy model where EMEs accumulate reserves to have more FDI to the economy. While these papers interpret the observed correlations in an idea of the mercantilism, we provide another way of looking at the fact from the point of the precautionary view. 8 Furthermore, 6 Other papers that studied reserve accumulation in the precautionary view are Durdu et al. (2008), Obstfeld et al. (2010), Aizenman (2011), Hur and Kondo (2016), Shousha (2017), and Bocola and Lorenzoni (2019). 7 Throughout this paper, we do not discuss the mercantilism view or include related ingredients in our model. The motivation of boosting exports certainly matters for reserve accumulations of some EMEs. However, to formalize the mercantilism view, it is often required to model some opaque externalities in export sectors, which we cannot easily verify. Also, the question of why EMEs rely on exchange rate manipulations to help their export sectors remains. On the empirical side, what crucially matters is the historical fact that the rally for the reserve accumulation began or was expedited after the crises in the mid-’90s. Moreover, it is often documented that the correlation between reserve accumulation and export growth is low for many EMEs, which contradicts the prediction from the mercantilism view; BIS (2019). However, our framework can be incorporated into the mercantilism view and our sense is we do not need any externality from export sectors if we add the mercantilism motivation to our model. We will discuss it at the end of this paper. 8 However, we also showed that reserve accumulation attracts more direct investments in an extension of the base- line model. 123 our model can naturally explain empirical regularities, which papers listed above might have diffi- culties explaining; positive correlations between FDI inflows and reserve outflows in a short run, 9 and positive correlations between equity portfolio investments and reserve accumulation, which resembles the correlations between direct investments and reserve accumulation. In our model, the reserve accumulation is an almost linear function of capital inflows in the form of direct invest- ment or equity portfolio investment 10 and thus those empirical facts are naturally explained by the model. The third literature related to ours is the nascent literature that studies the effectiveness of foreign exchange market interventions under the assumption of imperfect capital mobility. It is Gabaix and Maggiori (2015), which shows how limits to the arbitrage in global asset markets — UIP violation — can explain some important puzzles in the exchange rate literature, such as the forward premium puzzle. As an extension, they also show interventions in foreign exchange market should be effective and can be welfare-improving. Cavallino (2018) built a continuous-time New Keynesian general equilibrium model that analyzes foreign exchange market intervention, using almost the same specifications as Gabaix and Maggiori (2015). From a slightly different microfoundation, Fanelli and Straub (2019) derive general principles of foreign exchange market interventions. 11 To model limited arbitrage in a foreign exchange market, we mostly followed Fanelli and Straub (2019), but extended their modeling technique to private capital outflows from EMEs. Our new insights are that EME policy authorities might intervene not just to manage spreads on the borrowing rates, but also to supplement insufficient overseas investments by private sectors in the EME. We also incorporate the idea of imperfect capital mobility into a sudden stop 9 In the next section, we will show the positive correlation between FDI and international reserve is also clear in the flow data in the frequency in quarter or year. This is problematic in the explanation of the mercantilism view since the policy makers do not necessarily react to capital flows in such short runs; it usually takes several years for direct investors to decide which country to invest. 10 In addition, the approaches in Matsumoto (2019) and Wang (2019) cannot clearly explain why EMEs rely on reserve accumulation to attract more foreign direct investments rather than use other seemingly more efficient tools. For instance, simple tax cuts on the profits of foreign direct investors probably are more efficient than reserve accu- mulation. Another difficulty in the approaches in those papers is some exogenous factors such as geographic location might be dominant in determining FDI so that we will not see such clear positive correlations between FDI liabilities and reserve assets although EME policy authorities accumulate reserve to attract more FDIs. 11 For other papers that analyzed foreign exchange market intervention, see Basu et al. (2017), Chang (2017) and Amadar et al. (2019). 124 model and show how policy authorities can use foreign exchange market intervention to prepare for sudden stops in the future. Besides, our paper is related to capital control literature. After the Global Financial Crisis, there has been an eruption of the capital control literature; notable works are Bianchi (2011), Farhi and Werning (2014), and Bianchi and Mendoza (2018). Recently, a few papers analyze the rela- tionship between capital control and reserve accumulation. Acharya and Krishnamuthy (2018) and Jeanne (2015) argue that capital control is a complement to reserve accumulation policy in terms of financial stability since the capital control eliminates the possible moral hazard of private agents when central banks hold reserves. Arce et al. (2019) and Davis et al. (2019) show that reserve accumulation and capital control are perfect substitutes with each other in environments where collateral constraints on private sectors are always binding and yields on the reserves are the same as the borrowing rates. We do not precisely identify the relation of capital control to reserve ac- cumulation, but our analytical results imply the role of the reserve cannot be perfectly replaced by capital control since policy authorities accumulate reserves to generate additional capital outflows. Lastly, we contribute to the fear of floating literature. Since the famous paper Calvo and Rein- hart (2002), it has been widely known that many EMEs in fact manage their exchange rates rather than let the exchange rates float. More recently, Levy-Yeyati et al. (2013) document the pattern of foreign exchange market interventions in EMEs, which suggests the EMEs intervene in the market to prevent their currencies from appreciating rather than depreciate the currencies. The authors interpret the pattern as evidence that EMEs manage their exchange rates to expedite cap- ital accumulation in their countries. However, our model shows that reserve accumulation amid substantial capital inflows would look like interventions to avoid appreciation of the currency, but the purpose of the intervention is at preventing sudden stops and more efficient foreign liquid asset accumulation. Layout The rest of the paper is organized as follows. Section 2 illustrates stylized facts about reserve accumulation, which motivate our model. Section 3 introduces the model. The model explains the stylized facts in section 2 and provides new insights related to reserve accumulation. 125 In Section 4, we suggest three extensions of the baseline model in section 3. We confirm that our key insights will survive or even become stronger in the more general environments. Section 5 concludes and discusses avenues for future researchers. 3.1 Empirical Regularities As a first step, we find empirical regularities that will provide a guide to build our model of reserve accumulation. We aim at finding out what macro variables co-move with reserve accumulation and what macroeconomic conditions can explain different reserve holdings across EMEs. We first document the general facts regarding the reserve accumulation of EMEs since the mid-1990s. Then, we focus on how different types of capital flows are associated with reserve accumulation. 3.1.1 General Facts This subsection introduces general facts about international reserves. Readers familiar with the literature or readers who might want to focus on the key insight may skip this subsection. Reserve accumulation and the evolution of national wealth of EMEs As it is widely known, EMEs began accumulating massive amounts of reserves since the mid-90s. By the end of the ’80s, the levels of reserve holdings of most EMEs were around 5% of GDP. As in Figure 3.1, the level began rising from the mid-90s and the accumulation was escalated from the late 90s, which approximately corresponds to the East Asian crisis in 1997. The “rally” for having more reserves had been kept until the Global Financial Crisis in 2008 and after the crisis, the pace of reserve accumulation has become slow down. The pattern of increasing reserve holdings from the late 90s to the beginning of the crisis in 2008 is very consistent across different EMEs, but nevertheless the level of reserve holding varies along with a region and a country. Looking at Figure 1, it seems that Asian EMEs tend to hold more reserves; particularly Thailand and Malaysia holding reserves by the amounts of 30-40% 126 of GDP. Most EMEs in Latin America and East Europe, as of 2017, are holding reserves by the amounts of 10-20% of GDP, but some EMEs there hold reserves comparable to or even more than the levels of the East Asian EMEs; Peru, Russia and Bulgaria. Figure 3.1: Reserve Accumulation of EMEs Source: IMF BOP/IIP Now, we look at such rapid increases of the reserve holdings along with the evolution of ex- ternal liabilities and assets of EMEs; hence, along with the national “wealth” of the EMEs in the terminology in Lane and Milesi-Ferretti (2007). In terms of the Net Foreign Assets (NFAs), how reserve accumulation changes NFAs of EMEs is not clear as shown in Figure 3.2 below. Clearly, EMEs holding more reserves have more external liabilities and hence here we have no support- ive evidence of positive effects of reserve accumulation on NFAs. In the earlier literature, such ambiguous effects of reserve accumulation on NFA and the low returns to reserve compared with the borrowing rates of EMEs altogether raised a doubt on the precautionary motivation of reserve accumulation. For example, Rodrik (2006) stated that reserve accumulations seem to expedite the borrowing of private agents in EMEs, and therefore the NFAs of the EMEs might not increase enough to compensate the cost of reserves. More recently, Korinek (2018) insinuated that the Ricardian equivalence mechanism might work in a way of undoing the effects of reserve accumu- lation. 12 However, while there have been no vivid changes in the net position, there have been notable changes in the gross positions. First of all, as it was documented in Lane and Milesi-Ferretti 12 Another important criticism on the precautionary view of reserve accumulation is Alfaro and Kanczuk (2009). 127 Figure 3.2: NFA ex-IR and International Reserves Note: Average ratios of NFA ex-Reserves and Reserves to GDP, 2011-17 Source : IMF BOP/IIP (2007), Prasad (2011) and B´ en´ etrix et al.(2015), there was a shift of a way of external financing of EMEs: from debt financing to foreign direct investment (FDI) or equity portfolio investment; hence to the equity type financing. To put it another way, the high growth of the gross liabilities of EMEs was mostly attributable to direct investments and equity portfolio investments. As Figure 3.3-a indicates, the external debts scaled by GDP had slowly declined from the late 90s to 2006 and have remained at similar levels since then. On the other hand, FDI and portfolio equity liabilities have increased sharply and become the main component of external liability. It is noteworthy that the periods external debt to GDP ratios decreased roughly match the periods when the pace of reserve accumulation of EMEs was at its peak. Second, as in Figure 3.3-b, external assets held by private sectors in EMEs have also increased and its trend is very similar to the reserve assets. In the aggregation over EMEs in the sample, reserve assets of public sectors and external assets of private sectors closely co-moved until the early 2010s and interestingly, the two different external assets have diverged since then; external assets of private sectors, denominated by GDP, have rapidly increased while the official reserve holdings have been stable. In the appendix, we discuss possible 128 Figure 3.3: External Liability and Asset of EMEs (a) External Liability (by type) (b) External Asset (by agent) Note: All values are the summation across the sample countries. (a) Debt liability includes debt portfolio investment liability and other investment liability. Source : IMF BOP/IIP, External Wealth of Nations Database reasons for the divergence between private external assets and public external assets — official international reserves —, but the focus in this paper is on the periods EMEs rapidly increased the stocks of their reserve holdings. During the periods EMEs increased their reserve holdings, external assets of private sectors also increased. Weakly positive correlations between external assets of private sectors and official reserve assets As we saw in Figure 3-b, external assets held by private sectors in EMEs increased even during the time official reserve holdings were increasing. Here we look at the relationship between reserve assets and the private sector external assets in the cross-country data. Figure 4 presents a positive correlation between external assets of private sectors and reserve assets of public sectors. The correlation is low, but even such a weakly positive correlation is opposite to a prediction from the Ricardian equivalence mechanism; private agents expect reserve assets of public sectors will be ultimately given to them in the future, and thus the reserve accumulation incentivizes the private agents to have less foreign assets. In the next section, we provide a rationale behind the positive correlations. Also, in the appendix, we discuss other related facts such as the recent divergence between private sector external assets and official reserve assets in Figure 3.3-a. 129 Figure 3.4: Private External Assets and Reserves Note: All values are averaged over 2013-2017. Source : IMF BOP/IIP Positive correlations between current account surplus and reserve outflows Another impor- tant macroeconomic variable that comoves with reserve accumulation is the current account. What we found is that EMEs tend to accumulate more reserves facing larger current account surpluses, as evidenced in Figure 5 below. Looking at Figure 3.5, one might wonder whether the reserve accumulation causes the current account surplus or the current account surplus causes the reserve accumulation. Careful empirical analysis may find which causality results in the correlations in Figure 5 and few papers documented they found reserve accumulations improve the current ac- counts 13 . Empirical identification of precise causalities is beyond the scope of this paper, but our model shows the causality in either direction: reserve accumulation improves the current accounts and more (less) current account surplus (deficit) also causes more reserve accumulation. 13 See Bayoumi et al. (2015) and Choi and Taylor (2017). 130 Figure 3.5: Current Account and Reserve Accumulation Note: All values are averaged over 1998-2007. Source: IMF BOP/IIP 3.1.2 Reserve Accumulation and ”Extra” Capital Inflows In this subsection, we provide key empirical facts on the relationship between reserve outflows and equity-type capital inflows. Based on this relationship, we define ”extra” capital inflows and present how it is related to reserve accumulation. Positive correlations between reserve outflows and FDI & equity portfolio investment inflows As evidenced in Figure 3.3-a, the FDI liabilities of the EMEs are positively associated with reserve holdings of the EMEs. 14 Equity portfolio investment liability is a relatively small part of the total external liability. But, we can still see clear positive correlations between the total equity type liabilities — the summation of FDI liabilities and equity portfolio investment liabilities — and the reserve assets in Figure 3.3-b. 14 As we discussed in the last section, Matsumoto (2018) and Wang (2019) also reported the strong positive correla- tions between the FDI external liabilities and reserve assets. 131 Figure 3.6: Reserves and FDI & Equity Liability (a) Reserves and FDI (b) Reserves and FDI & Equity Note: All values are averaged over 2013-2017. Source : IMF BOP/IIP In addition to the close relationship in the stock variables, another interest is to see how different types of capital inflows are related to reserve outflows. We run a simple panel regression 15 and relegate the details of this regression to the appendix. The key results are that the equity type capital inflows, particularly equity portfolio investment inflows, is more correlated to reserve outflows. Furthermore, the positive correlations between the equity type inflows — FDI and equity portfolio — and reserve outflows become larger in the sample period of 2003-07 when EMEs were rapidly increasing their reserve holdings. Highly positive correlations between “extra” capital inflows and reserve outflows Finally, we define a measure of extra capital flows that summarizes the information from the empirical findings above. The extra capital inflow we define is seemingly unnatural, but interesting correla- tions between the external inflows and reserve outflows are observed in the data. We denote the extra capital inflows by ECI. Then we define ECI= Net Inflows of FDI & Equity Port f olio +Current Account Balance 15 Our regression analysis must be plagued by possible endogeneity and cross-sectional dependence in the data. However, we emphasize that this is a step to find suggestive evidence to help us with building a new model. 132 Then surprisingly, as in Figure 3.7 below, the crude measure of the extra capital inflows exhibits strong comovements with reserve outflows. For the selected six EMEs, the reserve accumulations move “in tandem” with the extra capital inflows. Figures of other EMEs are relegated to the appendix. Figure 3.7: Reserve Outflows and Extra Capital Inflows on Selected EMEs Note: All values are scaled by GDP. Source : IMF BOP/IIP To test the comovements in a different way, we regress reserve outflows on the extra capital inflows. Table 1 reports the results from the regressions. table reaffirms the relationship in Figure 3.7. A one percentage point increase in the extra capital inflows-to-GDP is associated with a 0.27- 0.62 percentage point increase in reserve outflows-to-GDP, depending on the sample periods and specifications. However, the positive coefficients may just result from the accounting identity; the current account balance must be almost identical to the capital account balance. To check it, we define the extra capital inflows differently; we replace the net FDI & equity portfolio investment inflows with the net debt inflows, naming them ECI2. Then it turns out that the measure with 133 the equity type flows has larger coefficients and more explanatory power — higher R 2 — than the measure with debt flows. As we expect, the difference between the two measures becomes larger in the sample period of 2003-07. Table 3.1: (a) Extra capital inflows Reserve outflows-to-GDP (Whole sample) (2003-2007) V ARIABLES (1) (2) (3) (4) ECI to GDP 0.33*** 0.27*** 0.62*** 0.42*** (0.071) (0.052) (0.113) (0.066) Constant 0.01*** 0.01 0.01*** 0.02*** (0.001) (0.005) (0.003) (0.005) Observations 560 560 140 140 R-squared 0.163 0.245 0.334 0.363 Number of Country 28 28 28 28 Country FE YES YES YES YES Year FE NO YES NO YES Robust standard errors in parentheses *** p<0.01, ** p<0.05, * p<0.1 (b) Alternative extra capital inflows Reserve outflows-to-GDP (Whole sample) (2003-2007) V ARIABLES (1) (2) (3) (4) ECI2 to GDP 0.21*** 0.15*** 0.14 0.12 (0.075) (0.047) (0.152) (0.072) Constant 0.02*** 0.01** 0.03*** 0.03*** (0.000) (0.005) (0.001) (0.006) Observations 560 560 140 140 R-squared 0.077 0.160 0.025 0.098 Number of Country 28 28 28 28 Country FE YES YES YES YES Year FE NO YES NO YES Robust standard errors in parentheses *** p<0.01, ** p<0.05, * p<0.1 Note: In each table, column (1) and (2) show the results with the whole sample (1998-2017) and column (3) and (4) with 2003-2007. Source: IMF BOP/IIP 134 What kind of rationale behind the strong comovements between the two series? Let’s imagine an EME receiving large amounts of capital inflows in the forms of direct investment or equity portfolio. Without any changes in the EME, the large capital inflows will cause domestic currency appreciations, which might incentivize private agents to invest abroad. However, if some frictions prevent the agents from investing abroad and it is desirable for the policy authorities to suppress the currency appreciation, the policy authorities do intervene and generate capital outflows by themselves. Of course, the outflows by the policy authorities are in the corresponding amounts of the excessive capital inflows. Another way of thinking of the relationship is that reserve outflows move as it would be done by private sectors. 16 The figures and the regression exercise imply that reserve outflows are made when EMEs receive capital inflows. Those tremendous capital inflows cause excessive resources and it obviously incentivizes the decentralized agents to invest abroad. In other words, if there is no friction on private capital outflows, the capital outflows by private agents will look like reserve outflows in Figure 3.7. The ideas of the discussion here will be materialized in the model of the following sections. Summary of the empirical regularities Now we summarize and list our empirical findings as below. 1. During the time EMEs rapidly increased their reserve holdings, both of external liabilities and external assets of the EMEs had risen. On the liability side, most of the increment of the external liabilities was foreign direct investments and equity portfolio investments. On the asset side, both assets of private sectors and official reserve assets increased. 2. EMEs holding more official reserves tend to have more external assets held by private sec- tors. 3. Current account balances are positively correlated to reserve outflows. 16 Such comovements only exist for EMEs. In most AEs, central banks do nothing with reserves. Some EMEs such as Japan frequently intervene in the market, such tight comovements are not observed. 135 4. Over the whole sample period, 1998-2017, equity portfolio investment inflows show the highest correlations with reserve outflows, while direct investment and debt inflows are also significantly and positively correlated to reserve outflows. In the sample period of 2003-07 when reserve accumulation of EMEs was at its peak, direct investment inflows are much more correlated to reserve outflows than debt inflows are correlated to reserve outflows. 5. The extra capital inflows, which we define as the sum of net inflows of FDI and equity and the current account surplus, exhibit strong comovements with reserve outflows. 3.2 Model In this section, we present our model to explain why EMEs accumulate reserves. Of course, we also aim at capturing the empirical patterns that we highlighted in the last section. Since the mechanism in the model is novel, we adopt a “minimum ingredients” strategy so that we can derive a few “pen and paper” analytical results that map the model to the empirical regularities. The minimum ingredients are 1) capital inflows in the form of direct investment, and 2) imperfect capital mobility for both capital inflows and outflows. We plug in these two into the framework of the small economy version of Fisherian deflation model, which was developed to model currency crises in the context of pecuniary externality. 17 Though we discuss key assumptions below, some readers may wonder how our key results change as we bring the model to a more general environment. In the next section, we will introduce three extensions of the baseline model, in which our key results and insights survive, and the key mechanism can be even more powerful. 3.2.1 Model Setup We consider a small open economy that last three periods, t = 0, 1, 2. There are two domestic agents in our model: households and the social planner. On the other side, there are two international 17 For more details of Fisherian deflation model, see Mendoza (2010) and Mendoza and Korinek (2014). 136 investors: international financial intermediations who intermediate capital inflows (outflows) to (from) the small open economy and direct investors who purchase the domestic capitals in the small open economy in period 0. The small open economy faces a credit constraint only in period 1, 18 which may or may not bind depending on the states. Thus, for precautionary purposes, the social planner accumulates reserves in period 0 when there is no concern over the binding credit constraint. 19 Production To model direct investments, while keeping the simplicity, we assume that both tradable and nontradable goods are produced by AK technology. That is, there are two different capitals, namely K T ;K N . K T ;K N are capitals to produce tradable goods and nontradable goods respectively. Two different kinds of capital cannot be converted from one to another and, more importantly, there is no capital accumulation or depreciation. Such a “semi-production” economy is a useful modeling technique to allow direct investments in the model while keeping the only necessary ingredients. We denote output of sector j at time t by y j t ; therefore y j t = A j t K j where A j t is the TFP of sector j at time t. Throughout this section, we do not allow direct investors to purchase the nontradable goods sector capitals. Then it is convenient to drop the upperscript j and let y T t = A t K. 20 We set y N 0 = y N 1 = y N 2 , but y T 0 < y T 1 < y T 2 (hence A 0 < A 1 < A 2 ) so that households need to borrow against higher outputs in the future. It is to investigate how reserve accumulation in our model is linked to the precautionary motivation. 21 18 This is same as Korinek and Sandri (2016). We can put the credit constraint in period 0. But, it makes it harder to solve the model, without providing any extra insight. 19 Remember the level of reserves of EMEs soared from the late 1990s to the mid 2000s. Considering many EMEs experienced or witnessed the currency crises around the mid 1990s, it is reasonable to think that policy authorities in EMEs during the time had a fear of currency crisis in the future, which is captured by the occasionally binding credit constraint in our model. 20 In fact, in terms of nontradable goods the economy is same as an endowment economy. In the next section, we will look at an extension where direct investors purchase capitals in nontradable goods sector. 21 However, we note that even with a decreasing output stream of y T 0 > y T 1 > y T 2 , the model generates positive amounts of reserve accumulation and the amounts of reserve accumulation can potentially be larger. 137 Households The overall utility of the representative household is given by U = u(c T 0 ;c N 0 )+E 0 bu(c T 1 ;c N 1 )+b 2 u(c T 2 ;c N 2 ) where the utility function u(c T t ; c N t )= ln c T t a c N t 1−a . Following the tradition, a and 1−a are the shares of tradable goods and nontradable goods respectively, andb is the discount rate. For simplicity, we use log-utility function. The households enter period 0 with some legacy debts d 0 , d 1 , d 2 , which the households have to repay in each period respectively. This is not necessary for us to derive desired results. The seemingly ad-hoc assumption is to match our model to the empirical findings in the last section: although the total outstanding external debt to GDP ratio has declined, surely EMEs have had substantial external debts. 22 Given the output streams and the legacy debts, households determine their borrowing or saving of tradable goods. Since there is no saving technology domestically, for the households to save they must invest abroad. The households’ saving decision will be discussed in detail in the paragraphs of overseas investments. Throughout this paper, we use households’ saving, lending abroad and overseas investments interchangeably. We denote borrowing (saving) of households at time t by b t+1 ; b t+1 < 0 corresponds to a borrowing and naturally b t+1 > 0 corresponds to a saving. Direct investments Direct investment inflows into the economy are novel and one of the most important elements in our model. A group of papers in the international trade literature analyzed foreign direct investments. 23 However, we analyze direct investments from a different research question and therefore we model direct investments in a different structure rather than follow the traditions in the international trade literature. In other words, we take some aspects of foreign direct investments depending on our purpose although the related environments look too barren or 22 Such drawback arises because our model is a representative agent model. In the next section, we will examine an extension to heterogeneous agents setup. It turns out that the heterogeneous agent extension matches the empirical findings better and prove our key insight survives in the new environment. However, as one expects it makes it harder to solve the model and poses few challenges. 23 See the excellent survey by Antras and Yeaple (2013). 138 unnatural. We will discuss more details below and here we just explain the direct investments in the model rather than discuss it. There is a direct investor interested in the tradable goods sector capitals in the economy. At the beginning of period 0, the direct investor determines how much capital she will purchase depending on technological features from which we abstract. Here purchasing capitals means to buy claims on the returns to the capital like equities in the reality. One can imagine direct investments in our model as a merger and acquisition process. Regarding the price of the capital, we assume the price is determined through a bargaining process, again similarly as like a merger and acquisition process in the reality. We assume that direct investors value the capitals higher than domestic households, which imply the equilibrium price of capital is higher than the valuation of households. Let q be the share of capitals sold to the direct investor among the measure 1 unit of capital: thus among the total returns of A t K,qA t K belongs to the direct investors and remaining returns for households. Also, let’s denote the price of capital by Q 0 , which is determined through the bargaining pro- cess. Then we have Q 0 > 2 å t=0 M h 0;t A t where M h 0;t is the stochastic discount factor of households; hence, M h 0;t =b t E u T t u T 0 and u T t is the marginal utility of tradable goods consumption in period t. Then direct investment capital inflows in period 0 is Q 0 qK: When we simulate our model for the purpose of the illustration of the model outcomes, we may assume that some direct investors buy the claims on the capital returns in period 0 and 1. In such a case, the “average” capital price will be Q 0 > 2 å t=0 M h 0;t A t (1s1 t=2 ) where s is the measure of the direct investors who purchase the capital claims only in period 0 and 1, and 1 is an identity function. This minor change gives a little more realistic description and 139 accordingly more realistic numerical results. Of course, this does not seriously change our results or significantly affect our insights. 24 In illustrating our analytical results, we mostly lets be zero, and otherwise will be noted in figures of numerical results. International Financial Intermediation and household overseas investment We haven’t ex- plicitly solved the model, but can easily envision that large amounts of direct investment inflows induce households to save. Since there is no available saving technology, households must invest abroad to transfer their incomes to the future. In the reality, there should be different sources of frictions on overseas investments, which we cannot include all in our model. Regarding the im- portance and simplicity, we focus on frictions about intermediations done by global investment banks. Below we will separately discuss how we can derive similar results with other similar specifications but under similar spirits. Reflecting on the fact that much of overseas investments are intermediated by international financial intermediaries (IFIs), we assume that any overseas investment by households must be intermediated by IFIs. There is a continuum of IFIs and they have a heterogeneous operation cost to intermediate. 25 Following Fanelli and Straub (2019), let the fixed cost be uniformly distributed on IFIs. That is, if there exists a continuum of IFIs, labeled by j2[0;¥), then IFI j pays a cost of j. Further, we assume that IFIs of measure c can manage the assets by the amount of gc. If the equilibrium intermediation fee is determined by the marginal cost of the intermediation, then it implies gc = b 1 (3.1) 24 Another way of understanding the minor modification is some direct investors quit the EME at the end of period 1. 25 Global banks in this context indicate that international investment banks operating in multiple countries. In fact, almost all renowned global banks such as JP Morgan, HSBC, and etc. have branches in major cities in most EMEs and do important roles in both capital inflows and outflows. In this context, the cost can be understood to run these branches or offices in EMEs; e.g., costs to hire and train new people in the EMEs. Also, the participation of a new IFI can be both of entrance of new IFI (extensive margin) or more operation of an incumbent IFI (intensive margin). 140 And furthermore, the return facing households, say r t+1 , is r t+1 = r G s b t+1 (3.2) whereG s =g 1 and r is direct return to the overseas investment before paying the intermediation fee 26 . As a result, r t+1 < r since b t+1 > 0. Furthermore, the gross return to the overseas investment b t+1 will be decomposed as follows. b t+1 (1+ r )= b t+1 (1+ r t+1 ) | {z } returns to households 1 2 G s b 2 t+1 | {z } returns to IFIs 1 2 G s b 2 t+1 | {z } total f ixed costs (3.3) The return to financial experts 1 2 G s b 2 t+1 is a sort of rent to IFIs and it represents a cost of overseas investment, which is not taken account by households. International Financial Intermediation and household borrowing The friction on capital outflows described above is a version of imperfect capital mobility, which has been popularized since the influential work by Gabaix and Maggiori (2015). 27 The feature in the paper is the break-down of interest parities so that exchange rates are determined by capital flows and accordingly foreign exchange market intervention becomes effective by changing the spreads. For the borrowing of households and the borrowing rates determination, we adopt results in Fanelli and Straub (2019). The key idea is that limited participation arises due to participation 26 We implicitly make two assumptions. First, the heterogeneous fixed cost is the cost in terms of tradable goods; each expert must pay tradable goods to participate. This is for simplicity and tractability. However, much of such cost following overseas investment in reality is the costs denominated by foreign currencies; for example, costs to manage branches in foreign countries. Second, the only available option to invest abroad is a fixed income security that pays a net return at the rate of r . This is surely counterfactual, but incorporating portfolio decision problem into our model will be overly complicated. 27 Other papers sharing a similar framework are Basu et al. (2016) and Cavallino (2018) 141 costs. If an EME borrows more, the EME attracts more IFIs into the bond market by paying higher rates; IFIs with higher participation costs need to join the market. That is, −b t+1 = 1 G b (r t+1 −r ) (3.4) Hence, the amount of capital inflow in the debt form linearly increases in the spread r t+1 −r . Rearranging the equation (3.4) yields a similar form with (3.2) r t+1 = r G b b t+1 (3.5) Credit constraint Households face a credit constraint in period 1. That is, b 2 f y T 1 (1q)+ p 1 y N 1 (3.6) The credit constraint is just the same as the collateral constraint in the recent capital control liter- ature (e.g., Bianchi, 2011; Korinek, 2018) where the collateral is GDP. 28 The idea of the standard specification of credit constraint in EMEs is that the borrowers in an EME may default on their external borrowing, but in such cases, international investors can take some properties in the coun- try, which prevents a default of the EME. Since the international investors cannot fully utilize the properties in the EME, the international investors should discount the values of the collateral. We add one property to this standard form. In our model, f is stochastic. 29 More formally, f(w) depends on the realized state w. It is frequently argued that an important driver of sudden stops is a change in the amount of funds that international investors are willing to provide for a 28 Some readers might wonder how the credit constraint can coexist with the limited participation. To handle the issue, we can conceptually decompose the borrowing process. In the first step, whether the credit constraint binds or not is determined by the realized states. If it turns out that the constraint binds, the amount of borrowing and the borrowing rates by the equation (3.6) and the credit constraint in (3.6). 29 The exogenous change of the f mostly reflects “global financial shocks” changes in global financial market conditions. Like Shin (2012), Bruno and Shin (2015), and Agrippino and Rey (2019), the change in global financial market conditions, which may stem from center economics, would cause changes in risk appetites of international investors. For example, when the conditions in global financial market become worsen, the risk appetite of the investors will be lower (risk-off), and therefore the investors will ask EMEs to provide more collaterals. 142 given amount of collateral, i.e. changes in the leverage parameter f. A few theoretical works in macro-finance, such as Geanakopolos (2008) document such pro-cyclical leverage ratios as a general feature of financial markets, and more recently Arce et al. (2019) have a similar feature with ours. It isn’t necessary, but for convenience and tractability, we assumef has a support of an interval, and further its CDF and PDF are both continuous. An additional note about the credit constraint in (3.6) is that it captures the infimum of the cost of sudden stop. Notice that disruption of consumption smoothing by drops of the consumption is the only cost of sudden stop. This is counter-factual; many of sudden stops accompany significant falls in output. 30 Furthermore, Nakamura et al. (2013) document that the negative impacts of a financial crisis on output may last much longer than expected in a standard model. In this regard, the conditions under which the social planner in our model is incentivized to accumulate reserves can be interpreted as “sufficient” conditions in the sense that it is socially desirable to accumulate reserves with the infimum cost of sudden stop. Social planner The social planner accumulates international reserves in period 0 when there is no concern over the binding credit constraint. To finance the accumulation of reserves, the planner imposes lump-sum taxes by the amount of T units of tradable goods. With the revenue from the tax, the planner purchases foreign bonds in period 0, which will earn 1+r units of tradable goods in period 1 per one unit of the bond. Accordingly, the dynamics of international reserves holding is given by T(1+ r)= R 1 3.2.1.1 Discussion of assumptions Direct investments The description of FDI in our model is much different from international trade literature. This is because the features we want to look at in this paper are different from the literature. In the classical small open economy model where net capital flows are captured by debt flows, external borrowing or lending is determined by the Euler equation of the representative 30 See Basu et al. (2017). 143 agent. The final decision is made by the agents in the small open economy. In contrast, it is not clear who makes the final decision of direct investment or equity portfolio investment; the final decision can be made by foreign investors. If there is no capital control, the capital inflows or equity portfolio investments are not necessarily aligned with the optimality conditions of domestic agents: capital inflows can be much more than what the small open economy needs to finance current deficits and pay back external debts on the maturity date. We model direct investment inflows as an almost exogenous variable, but we can endogenize the direct investment along with the capital price that direct investors pay. In the next section, we introduce an extension where the amounts of direct investments and the price of the capital are endogenous. We found that the social planner will accumulate more reserves in the new en- vironment. Intuitively, more reserve accumulation derives down the capital price through lower marginal utilities in the future. In our model, the direct investors never undo their investment during sudden stops. This reflects the idea that the direct investments are more stable than other capital flows. However, few papers such as Ostry et al. (2011) document that certain FDI can outflow during sudden stops. Similarly, it is also argued that irreversible ’greenfield’ FDI is actually small parts of the total. Considering all these points, it is more realistic to think that some FDI is partly reversible. But, throughout this paper, we maintain the common view that FDIs are more stable than others. Letting FDI outflow during sudden stops will give us a similar result with Jeanne and Sandri (2018). Overseas investment To the best of our knowledge, the only preceding paper in the international reserve literature that includes capital outflows of private sectors is Jeanne and Sandri (2018). In fact, almost all papers in the emerging market economy literature focus on capital inflows (or net inflows while assuming that EMEs have negative net foreign asset positions) 31 . However, thanks to the few influential empirical papers such as Forbes and Warnock (2012) and Broner et al. (2013), now the importance of analysis of gross flows is widely recognized. We also believe 31 An exception is a strand of papers that interpreted the global imbalance as a result heterogeneous financial devel- opment between developed and developing countries. See Caballero et al. (2008), Mendoza et al. (2009) and Maggiori (2017) 144 capital outflows of EMEs are really important, but unfortunately we still have a scant theory about private capital outflows from EMEs. Since we nearly have no giant on which we can stand, we borrow some modeling tools from recent papers assuming imperfect capital mobility and describe frictions of private overseas investment in the simplest way that gives us a clean result. To model frictions on capital outflows of private sectors, we assume that all the overseas in- vestments are intermediated by global investment banks. Of course, it is extreme, but also reflects reality. It is true that domestic banks in EMEs often need assistance from global investment banks when they invest abroad, probably due to the lack of expertise. For instance, sovereign wealth funds or national pension funds are often advised by global investment banks for overseas invest- ments. In the appendix, we documents that the profit growth of branches of global banks in Korea is positively correlated with the growth of overseas investment in Korea, which suggests that much of the overseas investments are directly or indirectly intermediated by the global banks. However, we do not want to limit the frictions on capital outflows to rent extractions to foreign banks. Actually, another important cost about capital outflows might be capital “migrations” for the purpose of tax evasion or improper concealment of assets, especially for EMEs with low quality institutions. The fear of such capital migration would have been a reason behind strict capital control measures on outflows, as documented in Fernandez et al. (2016). In the appendix, we document the relative strict controls on the outflows than inflows, and illustrate the evolution of the regulations in Korea, as an example. 32 Accordingly, we also provide another setup where domestic financial experts extract rents from households and conceal some of them. The result is a little different, but we can still maintain our key insights. 33 We also abstract from the portfolio decision of overseas investment; the only available option is a short-term fixed income security. Allowing different securities of different risk profiles does not 32 If the controls on capital outflows strictly limit the physical amounts of capital, to reflect such features in the model, we can set an arbitrary limit on b s or it can be understood as a case ofG s =¥ when the limit on b s is zero. 33 The concealment in the appendix is still the concealment by asset managers. Another important case for EMEs is the concealment by customers, not managers. Our sense is it should be almost identical to the case of asset managers, but simply the negative marginal impacts not in the consideration by other people are much larger. In our model, it can be understood as a case of a largeG s . 145 alter our qualitative result, but may strengthen the quantitative performance of the model because households may underestimate the cost of risk-taking in the portfolio decision. 3.2.2 Solving the Model Now we solve the model. First, we will solve the household problem in decentralized equilibrium, and then solve for the solution of the social planner who accumulate reserves or control capital flows. 3.2.2.1 Decentralized Equilibrium First we illustrate the household decisions. Then, a simple comparative statics is introduced to highlight how decisions of the households are related to direct investment inflows and reserve ac- cumulation, and how the relationship changes along with the measure of imperfect capital mobility, G. Next, we introduce our first analytical result, which proves the existence of pecuniary externality in our model and shows how the externality depends on the frictions on capital outflows. Finally, based on the comparative statics and the analytical result, we illustrate how the benefit of direct investment inflows is greatly reduced. Utility maximization of households We derive the solution of households via backward induc- tion. In the last period, by construction, there is no dynamic decision of households. Households consume all the available tradable and non-tradable goods after paying back all the debts and re- ceiving remaining reserves from the planner. As we will show later, in our three period model, the social planner depletes all the reserves in period 1 regardless of the realization of f, which corresponds to our intuition. However, the social planner might not deplete all the reserves once we make the model more dynamic; the number of periods is equal to or larger than 4. It is more formally stated and discussed in the subsection of the planning problem. Of course, the main conclusion and insights from the model are not affected by the reserve depletion decision. Back 146 to the household problem, since all the reserves are depleted in period 1, the consumption of the households is as below. c T 2 =(1q)y T 2 + b 2 (1+ r 2 ) d 2 ; c N 2 = y N 2 In period 1, the households take the states of the economy as given and solve the utility maximiza- tion problem. It is important to note that the states include f. As we emphasized earlier, f is a random variable whose value is determined at the beginning of period 1. A difficult question is “what would be a good distribution of f that resembles the reality?” To materialize an idea of a disaster, it seems a positively skewed distribution would be good, i.e., its pdf has a left-leaning curve. Here, we only assume that the distribution has a support on an interval of positive real numbers. That is, the support of f is h f;f i where f > 0 and ¯ f <¥. Also, we suppose it is nicely well-defined so that we don’t have any trouble in using calculus. The utility maximization is formally defined as below. max c T t ; c N t ; b t u c T 0 ; c N 0 +E bu c T 1 ; c N 1 +b 2 u c T 2 ; c N 2 sub ject to (3.7) (3.8) c T 0 = (1−q)y T 0 + p 0 y N 0 −p 0 c N 0 −b 1 −T+ Q 0 qk d 0 (3.9) c T 1 = (1−q)y T 1 + p 1 y N 1 + b 1 (1+ r 1 )+ R−p 1 c N 1 −b 2 d 1 (3.10) c T 2 = (1−q)y T 2 + p 2 y N 2 + b 2 (1+ r 2 )−p 2 c N 2 d 2 (3.11) −b 2 f(y T 1 (1−q)+ p 1 y N 1 ) (3.12) where u(c T t ; c N t )= ln c T t a c N t 1−a and R= T(1+ ¯ r). Market clearing conditions will be given by the pricing functions. p t = 1−a a c T t c N t and r t =G j b t + r where j= b;s 147 If the credit constraint doesn’t bind, i.e., the realized f is high enough, then the household determines its consumption of tradable goods according to her Euler equation. The amount of borrowing in period 1 is determined by −b 2 = −b(1+ r 2 ) (1q)y T 1 + b 1 (1+ r 1 )+ R d 1 +(1q)y T 2 d 2 (1+ r 2 )(1+b) (3.13) The interest rate r 2 is accordingly determined by r 2 = −G b b 2 +r . Plugging in the pricing function into (3.13) yields −b 2 = −z 1 + q z 2 1 −4(1+b)G b(1+ r ) (1−q)y T 1 + b 1 (1+ r 1 )+ R d 1 −(1−q)y T 2 d 2 2(1+b)G (3.14) wherez 1 =(1+ r )(1+b)+bG (1−q)y T 1 + b 1 (1+ r 1 )+ R d 1 . If the credit constraint binds, the consumption of tradable goods will be determined by the credit constraint. Plugging in the budget constraint into the credit constraint equation, we can derive 3435 b 2 = 1 1−f 1−a a ! f (1q)y T 1 + 1−a a (1q)y T 1 + b 1 (1+ r 1 ) d 1 + R (3.15) r 2 = r + G b 1−f 1−a a ! f (1q)y T 1 + 1−a a (1q)y T 1 + b 1 (1+ r 1 ) d 1 + R (3.16) 34 To have a unique equilibrium, we need a condition f 1−a a < 1. Here, it is easily satisfied since the credit constraint only binds for low values of f. The parametric restriction is imposed to make sure households cannot increase the limit in the credit constraint just by consuming more. Such parametric restrictions are usual in a small open economy model with a credit constraint including asset price (Korinek, 2018). 35 The equation (10) implies that the borrowing rate r 2 is low during sudden stop. This is a little unsatisfactory since the spread often soars during sudden stop. This can be corrected by lettingG b as a function off andG 0 b < 0 or similarly assuming r is a decreasing function off. However, since the counterfactual does not seriously affect our key insight and the modifications make the model complicated without providing extra insights, we keepG b and r as constants. 148 Sincef has a support of an interval, we can derive a formula of the cut-off off, below which the credit constraint binds, given other states and the reserve. We can obtain f c = b 2 1 a (1q)y T 1 + 1a a (b 1 (1+ r 1 )+ R d 1 b 2 ) (3.17) whereb 2 is determined by (3.14). We now turn to the period 0 optimization problem. Since there is no credit constraint, trivially the solution is the Euler equation. By denoting the marginal utility of good x in period t as u x t , u T 0 (b 1 ; ;)=b(1+ r 1 ) Z f c f u T 1 (b 1 ; b 2;c ; ;)dF f + Z f f c u T 1 (b 1 ; b 2;u ; ;)dF f ! (3.18) where b 2;c is determined by (3.15), while b 2;u is determined by (3.14). Comparative statics of the households decision Although we cannot solve for b 1 more explic- itly, it is easy to see the key characteristics of the b 1 as a function of q and R. We can easily show ¶b 1 ¶q > 0; ¶b 1 ¶R 1 < 0 It is straightforward that b 1 increases inq since more capital purchase by direct investors in period 0 induces more excess tradable goods to save for the future 36 . Similarly, b 1 decreases in the amounts of reserve accumulation because households borrow more or save less as resources are transferred from the current period to future periods through the reserve accumulation. In addition to these comparative statics, it is important to notice that the marginal impacts of direct investments and reserve accumulation on the borrowing (saving) vary with the value of G b (G s ): the absolute values of ¶b 1 ¶q and ¶b 1 ¶R depend onG. We first look at the dependence of ¶b 1 ¶q on G. If b 1 > 0; more direct investments increase the saving and therefore lower the interest rates facing the households; recall r t =G s b t + r if 36 Recall the budget constraint in period 0, c T 0 + p 0 c N 0 =(1−q)y T 0 + p 0 y N 0 −b 1 −T+ Q 0 qk d 0 149 b 1 > 0 37 . Hence, the “second” effect works in a way of reducing the saving. To see it more clearly, let’s invoke the implicit function theorem to the Euler equation. Then we can see db 1 dq = u 00 0 y T 0 + Q 0 qk +b(1+ r 1 )y T 1 E 0 h u 00 1 i u 00 0 −bG s E 0 u 0 1 −b(1+ r 1 ) E 0 h u 00 1 (1+ r −2G s b 1 ) 1− ¶b 2 ¶q i (3.19) where u 00 t = ¶u t ¶c T t : It is little cumbersome to show it formally, but we can easily seej db 1 dq j decrease in G s . To an extreme, as G s !¥, db 1 dq ! 0. Obviously, for G s !¥, any positive saving derives down the gross yield to zero, and therefore households don’t save in any amount; they consume all the tradable goods. To summarize, massive direct investments in period 0 incentivize households to save so much for the consumption smoothing, but severe friction on capital outflows (highG s ) hinder the households from investing abroad to transfer extra resources to the future periods. The friction on capital outflows by households is the key to understanding the motivation of reserve accumulation in our model. Direct investments or equity portfolio investments, which are not necessarily in line with optimal consumption smoothing of the households, give households more resource than needed to consume now and if households cannot save (invest abroad) to reallocate the resources enough or it is done inefficiently, the social planner needs to supplement (or substitute for) the insufficient savings by households. Next we look at the relationship betweenG and ¶b 1 ¶R . As with direct investments, reserve ac- cumulation affects the borrowing (saving) rates through the changes in borrowing (saving) itself. When households borrow, the households face higher interest rates as they borrow more respond- ing to reserve accumulation; therefore the households will borrow less. On the contrary, when households save, the less savings due to reserve accumulation drive up the yields facing the house- holds, which results in more savings. To summarize, the changes in the interest rates also make 37 The effect is opposite for b 1 < 0 since direct investments reduce borrowing and it lowers the borrowing rates. Here, we only look at b 1 > 0 because it is a more empirically relevant case and that is the case in which central banks are likely to accumulate reserves. 150 the borrowing and saving less sensitive to reserve accumulation: hence, the Ricardian equivalence is broken. This is illustrated in Figure 3.2 and the equation (3.20) below. db 1 dR 1 = u 00 0 +b(1+ r 1 )E 0 h u 00 1 (1+ ¯ r)− ¶b 2 ¶R i u 00 0 −b(1+ r 1 )E 0 h u 00 1 1+ r −2G j b 1 − ¶b 2 ¶b 1 i −E 0 bu 0 1 G j (3.20) where u 00 t = ¶u t ¶c T t : Obviously, higherG suppresses the responsiveness of the borrowing (saving) to the reserve accumulation. These comparative statics are important in deriving our key results, which we will derive in the next subsection. Considering its importance, we highlight the comparative statics in the following remark. Remark 1 More frictions on capital flows, higherG, make the saving and borrowing by house- holds less responsive to reserve accumulation and to direct investments as well if households save. In analytical forms, ¶b 1 ¶q > 0; ¶b 1 ¶R < 0 ¶ 2 b 1 ¶q¶G s < 0; ¶ 2 b 1 ¶R¶G > 0 Pecuniary externalities Before we move on to the subsection of the planning problem, we in- troduce our first analytical result. We can easily notice the amount of borrowing and saving in the Euler equation (3.18) is not necessarily socially optimal. b 1 will impact the interest rates r 1 , r 2 , and p 1 , and households do not take account of it. The externalities arise because agents do not take account of the impact of their action on the prices and such externalities are often named as pecu- niary externality 38 . We here note that there are two different sources of pecuniary externalities; one through the real exchange rate in period 1 (p 1 ), and the other one through interest rates in period 0, 38 For a more detailed discussion of pecuniary externality, see Davila and Korinek (2018) 151 Figure 3.8: Comparative statics of the households decision 1 (r 1 , r 2 ). Interestingly, when households borrow, both externalities work in the same way, whereas the two different externalities work in the opposite way from each other when households save. In the same way as the models in the capital control literature, households do not consider the effects of their decision on the real exchange rate in the future, which creates overborrowing (Bianchi, 2011). At the same time, more borrowing by the households also creates another source of over- borrowing; more borrowing by households pushes up the interest rate on the borrowing. On the contrary, when households save more (higher b 1 > 0), it pushes down the yields on the saving, but it also makes the economy better prepared for a possible crisis in the next period. Consequently, higher saving by households creates two different externalities working in the opposite ways in terms of the distance between the saving by households and the socially optimal saving. Whether the saving decision by households will be “undersaving” or “oversaving” in the eye of the social planner is itself indeterministic and depends on the state of the economy and values of important parameters; the distribution off andG s . We formally state these results in the following lemma. The formal proof is relegated to the appendix B. 152 Lemma 1 Assumef c >f andG s ;G b > 0. b h t be the solution of (3.18), and b sp t be the solution by a social planner. 1) If b h t < 0, then b h t 0, then there always existsg 0 2(0;¥] such that forG s 2(0;g 0 ) b h t b sp t Proof) See the Appendix C. Lemma 1 illustrates two (or three) different externalities from capital flows and when we are more likely to have undersaving or oversaving by households. First, as long as prob(f <f c ) > 0 39 , any borrowing by households has an externality of tightening the credit constraint in period 1. The borrowing of households will lower the real exchange rate in period 1, p 1 , which is included in the credit constraintb 2 f (1q)y T 1 + p 1 y N 1 ; obviously, more borrowing will reduce the tradable consumption in period 1, and thereby resulting in lower real exchange rates. Second, any borrowing or lending abroad results in extra costs of capital flows, of which households do not take account. Recall that r t =G j b t + r and one unit of borrowing or lending gives b 1 (1+ r 1 ) in period 1. Differentiating b 1 (1+ r 1 (b 1 )) with respect to b 1 gives d(b 1 (1+ r 1 (b 1 ))) db 1 = 1+ r 2G j b 1 = 1+ r 1 (b 1 )G j b 1 Obviously, households are price takers and hence they only consider 1+ r 1 as a cost (return) to their borrowing (lending). Thus, any borrowing or lending by households leaves a term that is not in the calculation of the households: the second externality through borrowing rates or returns to overseas investments. Here it is important to note that the termG j b 1 is positive for b 1 < 0 while it is negative for b 1 > 0. These different signs show reasons why we always have b h t b sp t for b h t > 0 only ifG s is large enough. If households borrow abroad, additional borrowing raises the borrowing rate. Hence less borrowing, which leaves more resources for period 39 We assumed that the support off is 0;f , the condition is equivalent to b 2 < 0. 153 1, is desirable in terms of both the borrowing rates and the preparation for sudden stops in period 1. In contrast, once the borrowing alters to lending due to direct investment inflows, additional lending increases the marginal cost of overseas investments, equivalently lowers returns to the lending. This makes more lending by households undesirable for the social planner, while it still makes the economy better prepared for possible sudden stops in the next period. Thus the saving creates two externalities that work in the opposite direction from each other; the negative externality of lower returns, and the positive externality of less probable and less severe sudden stops. Of course, the negative externality increases in the measure of the frictions on oversea lending. Therefore the socially optimal overseas investments are lower than the investments by households forG s large enough and vice versa. Direct investment inflows and decentralized equilibrium The comparative statics above im- ply how the equilibrium without intervention by the social planner might change. To discuss the change, we need to see the problem facing the economy. Since we assumed y T 2 > y T 1 > y T 0 , the optimum for the households is to borrow against larger outputs in the future. However, borrowing creates two externalities as discussed above. Those externalities hinder households from “transfer- ring” resources from the future to the present. The direct investment inflows provide a better way of external financing for the households with the problems described above. Capital inflows in the form of direct investment are free from concerns about the externalities, or at least are better in terms of the externalities as we discussed in the last section; for example, the required return is less sensitive to the amount of capital inflows than debt inflows. However, the friction imposed on capital outflows, G s , gives a new challenge for the households. For direct investments inflows above a certain level, households need to save; they need to lend tradable goods abroad. Previously, they needed to bring the resources from the future to the present for the purpose of consumption smoothing, but now they need to reallocate the extra resources in the present to the future, again for the consumption smoothing. However, ifG s is non-negligible more saving by households leads to lower returns, which in turn lead to insufficient saving and inefficient consumption boom accordingly. 154 Figure 3.9 below shows how the decentralized equilibrium changes along with direct invest- ment inflows, parameterq in the model. As one can easily envision from the comparative statics, more capital inflows cause higher tradable goods consumption in period 0, so higher real exchange rates. Despite the inefficient consumption booms in period 0, the direct investment inflows make the economy more robust to sudden stop: lower probability of binding credit constraint and less tight constraint for givenf. Hence, as it is commonly argued, external financing in the forms of di- rect investments or equity portfolio investments is better in terms of macro-prudence in our model. However, the gain is strictly diminishing in the friction of capital outflows. Figure vividly indi- cates that magnitudes of the decline of the sudden stop probability prob(f <f c ) rapidly decrease inG s . For an extremely largeG s , it is observed that the probability of sudden stop against direct investment inflows exhibits “U-shaped” curve: For direct investment flows above a certain level, the economy becomes more vulnerable to sudden stop as more direct investment capitals inflow. 40 Such an efficiency arises because households cannot save enough by themselves due to the friction underlying overseas investments. To summarize, the economy suffers from the inefficiency of overseas lending by households, which generates extra costs of the lending,G j b 1 . In another aspect, the low returns significantly dampen the benefit of direct investment flows in terms of the macro-prudence. It will be introduced and analyzed in details, but the analytics so far already hint what would be the role of the social planner: Since agents in private sectors cannot save enough and it is done inefficiently, the planner invests instead of the private agents. 3.2.2.2 Equilibrium with Social Planner Now we solve for the solution of the planning problem. First, we will derive the solution of a social planner without capital controls and then derive the optimal reserve accumulation with 40 We can see the relationship between sudden stop and direct investment more explicitly through ¶b 2 ¶q j f<f c. ¶b 2 ¶q j f<f c=f 1 a y T 1 +( 1a a )(12G s ) db 1 dq l1f( 1a a ) . The first term in the derivative 1 a y T 1 is less collateral in period 1; since the claim on the capital was sold to foreigners, it cannot be used as pledged collaterals. The second term is effect from more saving from period 0. More tradable goods saved for period 1 raise the real exchange rate so as to ease the credit constraint. 155 Figure 3.9: Decentralized Equilibrium Note: As a benchmark, the parameter values for our numerical results are as follows: lower bound of borrowing rate(r )=0.05, interest rate on reserves(¯ r)=0.01, weight on tradables(a)=0.35, discount factor (b)= 0.94, distribution of credit coefficient=Beta(1.5,5), wedges in UIP(G s = 0.2, G b =0.25), endowment stream( y T 0 =0.8, y 1 T = 1; y 2 T = 1:5; y NT =1), legacy debts(d 0 =0.1, d 1 =0.2, d 2 =0.1). 156 capital controls. Next, we will introduce a proposition that gives an implication of “currency manipulation.” Reserve accumulation without capital control To solve for the reserve accumulation in period 0, we first need to solve for the reserve depletion in period 1. As one easily expects, the social planner, regardless of whether the credit constraint binds or not, depletes all the reserves. It ap- parently looks natural, but it is actually a little unsatisfactory considering the fact that many EMEs during sudden stops did not deplete much of their reserves: fear of losing reserve (Aizenman and Sun, 2012). A few recent theoretical papers explicitly showed that in their environments a social planner never depletes all the reserves during a sudden stop or other crises looking alike; Bocola and Lorenzoni (2019). However, once we extend our model to a more dynamic version, i.e., the number of periods larger than 3, we can have similar results. The intuition is as follows. If the social planner facing an ongoing sudden stop expects that the crisis might be recurrent in the near future, then the social planner leaves a part of the reserves for a possible crisis in the future. A more detailed analysis is in the next section. Given that the social planner will deplete all the reserves in period 1, we can formulate the planning problem as below. max R V = u c T 0 ; y N 0 +E 0 bu c T 1 ; y N 1 +b 2 u c T 2 ; y N 2 sub ject to −b 1 = the solution of(3:18) −b 2 = 8 > > < > > : f((1q)y T 1 + 1−a a ((1q)y T 1 +b 1 (1+r 1 )d 1 +R)) 1−f( 1−a a ) i f f2 h f;f c i −z 1 + q z 2 1 −4(1+b)G(b(1+r )((1−q)y T 1 +b 1 (1+r 1 )+Rd 1)−(1−q)y T 2 d 2) 2(1+b)G i f f2 f c ;f c T 0 = (1−q)y T 0 + p 0 y N 0 −p 0 c N 0 −b 1 −T+ Q 0 qk d 0 c T 1 = (1−q)y T 1 + p 1 y N 1 + b 1 (1+ r 1 )+ R−p 1 c N 1 −b 2 d 1 c T 2 = (1−q)y T 2 + p 2 y N 2 + b 2 (1+ r 2 )−p 2 c N 2 d 2 r t = −Gb t + r 157 where u(c T t ; c N t )= ln c T t a c N t 1−a and R= T(1+ ¯ r): Then deriving the first-order condition and rearranging the equation yields the proposition 1. Proposition 4 The optimal reserve accumulation at t=0 is characterized by If b h 1 < 0 bG b E u T 1 b 1 db 1 dR 1 | {z } Higher r 1 + u T 0 b(1+ r)E u T 1 | {z } ConsumptionWedge at r = b dw 1 dR 1 2 6 6 6 4 Z f c f d(−b 2 ) dw 1 u T 1 b(1+ r 2 )u T 2 dF f | {z } Marginal Value o f Borrowing −bG b E u T 2 b 2 db 2 dw 1 | {z } Lower r 2 3 7 7 7 5 (3.21) If b h 1 > 0 u T 0 b(1+ r)E u T 1 | {z } ConsumptionWedge at r = b dw 1 dR 1 2 6 6 6 4 Z f c f d(−b 2 ) dw 1 u T 1 b(1+ r 2 )u T 2 dF f | {z } Marginal Value o f Borrowing −bG b E u T 2 b 2 db 2 dw 1 | {z } Lower r 2 3 7 7 7 5 bG s E u T 1 b 1 db s 1 dR 1 | {z } Higher r 1 (3.22) where w 1 = R 1 + b 1 (1+ r 1 ) and dw 1 dR 1 = ¶w 1 ¶R + ¶w 1 ¶b 1 db dR = 1+ 1+ r 2G j b 1 db 1 dR Proof) See the Appendix C. The proposition above well illustrates what determines the optimal reserve accumulation. The terms in the LHS are the marginal costs of reserve accumulation and the terms in the RHS are the benefits. The term u T 0 b(1+ r)E u T 1 in the LHS indicates the cost of reserve accumulation due to low returns to reserve, in terms of utility. 41 The two terms in the RHS are the benefits from higher wealth in period 1 42 . Because of the imperfect capital mobility in both of capital inflows 41 As long as r< r 1 , u T 0 b(1+ r)E u T 1 is positive; hence positive marginal cost. 42 Please notice that this excludes the direct investments. The term wealth can be understood as net foreign currency liquidity. 158 and outflows, reserve accumulation raises the wealth in the future, which in turn helps with sudden stops and drives down expected borrowing rates in period 1. 43 An interesting term isbG j E u T 1 b 1 db 1 dR 1 . The term is in LHS in equation (3.21) whereas it is in RHS in equation (3.22): the term is a marginal cost when b 1 < 0; but it is a marginal benefit when b 1 > 0. This is related to the mechanism of how reserve accumulation works. When households borrow, the Ricardian equivalence breaks down due to higher borrowing rates; hence db 1 dR >1. On the contrary, the Ricardian equivalence breaks down due to higher yields for households when the households save. Thus the changes in the interest rates driven by reserve accumulation are the costs when households borrow, but are benefits when households save. It naturally implies that we are more likely to have an interior solution, i.e., positive reserve accumulation when households save rather than borrow. Figure 10 below illustrates this point. As we stated in the last subsection, b 1 increases in direct investment which is measured byq, and as b 1 increases, the optimal reserve accumulation increases as well. Moreover, forq such that b 1 < 0; the optimal reserve accumulation is not to accumulate reserve. Of course, our model is stylized and thus we should not think of the results quantitatively. However, the results in propo- sition 1 and Figure 10 highlight deficiencies of reserve accumulation as a macroprudential policy tool in an environment where agents borrow from outside of the economy. As noted in lemma 1, when households borrow, the economy suffers from the overborrowing problem. Since reserve accumulation always incentivizes households to borrow more if the households borrow in the ab- sence of reserve accumulation, the reserve accumulation calls a side effect: the economy suffers even more from the overborrowing. This is similar to a few preceding papers that documented the moral hazard from reserve accumulation (Acharya and Krishnamurthy, 2018). 44 Furthermore, once we extend the model to a more dynamic version as we will see in the next section, the reserve 43 Throughout this paper, we implicitly assume b 2 j f>f c< 0. That is, households still want to borrow in period 1. However, if direct investment inflows in period 0 is so overwhelming; i.e., q is large enough, it is possible to have b 2 j f>f c> 0 under the reserve accumulation. This might be relevant to some EMEs with large amounts of external assets comparing with the external debts. However, it is a little difficult to interpret the results in such a case. We relegate the analysis of the case of b 2 j f>f c> 0 to the appendix C. Also notice that we likely to have b 2 j f>f c> 0 when the planner optimally accumulates reserves. Otherwise, very lowG s and largeq are required to generate b 2 j f>f c> 0. 44 Same as our paper, the moral hazard in those papers is slightly different from the traditional concept of moral hazard because of the absence of agency problem. 159 Figure 3.10: Reserve accumulation without capital control Note: All the parameter values are the same as the benchmark except for distribution of credit coefficient(=Beta(1.2,5)). accumulation policy is not time-consistent. Altogether it implies that reserve accumulation in our model and similar models using imperfect capital mobility suffer from the usual side effects of ex-ante bail-out policies: Moral hazard and Time-inconsistency. However, once we model reserve accumulation as a policy to supplement insufficient overseas investments by private sectors the deficiencies will be lessened because less saving by households is desirable at least in terms of the returns to the saving. Reserve accumulation with capital control Now we analyze the reserve accumulation in an en- vironment where the social planner has control over the capital flows. Since there are two different types of capital flows - direct investment and debt instrument - in our model, we separately look at the two different capital controls. 45 45 Throughout this subsection, we assume that the capital control is perfect; the social planner can compute the optimal tax or subsidy to the different capital flows and can impose it perfectly. This is obviously counterfactual. In reality, the capital control is imperfect due to incomplete information or any other reasons and furthermore cannot 160 Capital controls over debt flows Suppose the social planner can tax or subsidize (hence negative tax) debt instruments type capital flows, i.e., b t+1 . As it is well known in the capital con- trol literature, the optimal tax can achieve the same equilibrium where the social planner directly chooses the capital flow. Because this is well known and analyzed in the literature, we do not discuss it further and introduce the optimal tax on the debt type capital flows. The optimal tax is characterized as below. t b = 1 u T 0 b[ Z f c f db 2 db 1 u T 1 −b(1+ r 2 )u T 2 dF f − Z f f Gb 1 u T 1 −bGb 2 db 2 db 1 u T 2 dF f ] (3.23) As one can easily see, the optimal tax increases in the externalities from borrowing or saving of households. 46 Now we present the two equations that characterize the equilibrium where the social planner optimally accumulates reserve and tax (or choose borrowing or saving). Assuming an interior solution, which is not always, the equilibrium is characterized by the two equations. u T 0 +bu T 1 1+ r 2G s b sp 1 +b Z f c f u T 1 +bu T 2 (1+ r 2 ) db 2 dw 1 1+ r 2G s b sp 1 = 0 (3.24) u T 0 +bu T 1 (1+ r)+b Z f c f u T 1 +bu T 2 (1+ r 2 ) db 2 dw 1 (1+ r)= 0 (3.25) where w 1 = b 1 (1+ r 1 )+R 1 and b 2 is characterized in equation (3.14), (3.15). It is straightforward that we can solve for b sp 1 once we assume an interior solution for both of b sp 1 and R 1 . Equation (3.24) and (3.25) give us b sp 1 = r r 2G s (3.26) be imposed perfectly. To cover those realistic features is obviously beyond the coverage of this paper. For a formal analysis of the leakage of capital control, see Bengui and Bianchi (2018). 46 Some papers named such taxation ’Pigouvian taxation’. See Jeanne and Korinek (2019). Also, please notice that there are multiple policy instruments that achieve the same equilibrium. For deeper discussions, see Benigno et al. (2016). 161 See in equation (3.26), for b sp 1 > 0, b sp 1 ! 0 as G s !¥. On the contrary, b sp 1 !¥ as G s ! 0, which is a contradiction. The contradiction implies that we cannot have an interior solution forG s small enough. More broadly, the reserve accumulation increases inG s , while overseas investments are chosen by the social planner decrease in G s . This is intuitive and corresponds to our main message throughout this paper. The social planner accumulates reserves since households cannot lend abroad enough by themselves.G s is the measure of the friction on the private capital outflows and therefore the results above are straightforward. b sp 1 > 0 in equation (3.26) since r > r. This implies that R = 0 if b sp 1 < 0. The result of no reserve accumulation when the social planner chooses to borrow is also important and interest- ing analytics. In this paper, we aim at explaining why EMEs choose reserve accumulation as a macroprudential policy tool against sudden stop. For this purpose, suggesting mechanisms of how reserve accumulation works against sudden stop is not enough: It is a necessary condition, but not a sufficient condition. If there are multiple policy options, we need to explain why reserve accumu- lation is chosen over others and what is the unique role of it. Equation (3.26) already answers those questions. Although the social planner optimally subsidizes investments by households, which is unrealistic since such taxation or subsidy is never perfect in the reality, the social planner would like to accumulate reserves for G s large enough. If G s is large, the marginal benefit of overseas investment rapidly falls, and therefore beyond a certain level, the planner accumulates reserves to have more foreign assets. On the contrary, when the social planner optimally chooses to borrow, the capital control always dominates the reserve accumulation. Recall that the function of reserve accumulation is to raise net foreign currency external assets. In this regard, the taxation on external borrowing is always better than the reserve accumulation since the reserve accumulation works by raising the borrowing rates; it gives a penalty to the borrowers. Then it is obvious to see why taxing external borrowings 162 is better than the reserve accumulation. The capital control lowers the borrowing rates whereas the reserve accumulation raises the rates 47 . We summarize these findings in the second proposition. Proposition 5 Suppose the planner optimally taxes or subsidies debt capital flows, then 1. b sp t+1 = b t+1 (t t ) 2. If b sp t+1 < 0 and b sp 1 = b t+1 then R = 0 3. If b sp t+1 > 0 and b sp 1 = b t+1 then we may have R > 0 Proof) See the Appendix C. We can formulate the proposition in a different way, which gives us a corollary. It restates that the capital control on debt flows always dominates reserve accumulation for the borrowing, but reserve accumulation can be more efficient than the saving. 48 Corollary 2 Let V 0 (R) is the value function of the planner with the optimal reserve accumulation at t=0. Similarly, define V 0 (t). Then, V 0 (t)> V 0 (R) if b sp 1 < 0: On the contrary, if b sp 1 > 0; then we may have V 0 (R)> V 0 (t) for r high enough orG s high enough. Proof) See the Appendix C. Capital controls over direct invesment flows If reserve accumulation is a reaction to direct investment flows beyond a level above which households are forced to save in an inefficient way, one easy solution would be to limit the direct investments themselves. For example, EME govern- ments may ban foreign investors from buying domestic assets or set a cap, above which foreign 47 Few theoretical papers (Davis et al., 2019; and Arce et al., 2019) documented the equivalence between reserve accumulation and capital controls based on the assumption that the credit constraints are always binding. However, these papers also assumed that the returns to reserve are the same as the borrowing rates, which is counterfactual. 48 The proposition and corollary also explain another puzzle about international reserve accumulation. One puzzle about international reserve is why central banks in EMEs limit the compositions of their reserves to certain safe assets such as US treasury bonds. Because of the low returns to these assets, sovereign wealth funds, whose portfolios accommodate more risky assets were expected to replace international reserves by central banks to a substantial extent. However, most of the external assets held by the public sector in EMEs are still international reserve. Proposition 2 suggests that the sovereign wealth funds are probably subject to the same frictions with private sectors. If sovereign wealth funds need to rely on foreign banks to make overseas investments, they must be subject to the same restriction (G s ). Also, the inefficiencies of domestic financial sectors or low quality institutions may matter in a similar way; for example, corruption in the sovereign wealth funds. In other words, although the social planner chooses how much to save and invest abroad, the social planner still faces the same constraint and thereby being incentivized to accumulate reserves. 163 investors cannot buy more. Regardless of the difficulties in implementing these regulations, it is hard to analyze the optimal control over direct investment flows in this paper. Direct investment flows are viewed better than debt flows not just because it is more stable, but also there might be some technological spill-over effects. Such unobservable positive effects might be the same for equity portfolio investments; for example, it promotes the development of domestic stock markets. Those externalities may exist or not, and moreover, the quantitative importance is hard to measure. Since these are beyond the scope of our paper, we mute all these channels, through which the capital inflows positively impact EMEs. Interestingly, even without those externalities, it turns out that the social planner in our model economy wants to receive direct investments to some extent, which provides incentives to accumulate reserves. To make the problem simple, let’s suppose that the social planner can choose q. Further, assume that the capital is priced by the stochastic discount factors of households. That is Q 0 = 2 å t=0 b t A t u T t u T 0 Hence there is no extra gain from direct investments in terms of price. Letq beq chosen by the social planner. Then FOC ofq is as follows. E −bG j b 1 u T 1 db 1 dq −b 2 G b b 2 u T 2 db 2 dq | {z } Changes in r 1 ;r 2 Z f c f db 2 dq u T 1 b(1+ r 2 )u T 2 | {z } Marginal Value o f Borrowing + dQ 0 dq u T 0 | {z } Change in Q 0 = 0 (3.27) We interpret the equation (3.27) after introducing our second lemma driven from the equation (3.27). Lemma 2 Suppose ¶b 2 ¶q j f<f c< 0 but E h u T 2 b 2 ¶b 2 ¶q i < 0. Then withq =q ;b h 1 > 0. Proof) It it obvious that all the three terms in equation (3.27) are positive under b h 1 < 0. If b 1 > 0, then the first termE h −bG j b 1 u T 1 db 1 dq i is negative. Therefore, for the equation (3.27) to hold, b h 1 must be positive. 164 To give some intuition to equation (3.27) and lemma 2, first notice that all the terms in equation (3.27) are related to the externalities from borrowing or saving by households. Any borrowing or saving by households affects interest rates and real exchange rates in the future, of which the households do not take account. Since direct investments change b 1 , it generates the same exter- nalities as well through the changes in b 1 . The first and second term are related to changes in r 1 and r 2 respectively, and the last term is the changes in borrowing under sudden stops through real exchange rates in the collateral constraint 49 . As stated in lemma 1 and the comparative statics in the last section, it is always b h 1 0. Then it is straightforward that more direct investment is beneficial since it lessens the overborrowing problem: Receiving more direct investment flows is desirable for the social planner as long as the economy is in the state of overborrowing. To give more economic interpretation, our model EME “transfers” resources from the future to the present for the con- sumption smoothing. However, because of the imperfect capital mobilities and credit constraint, the economy has trouble having transfer resources from the future. Direct investments do provide another way of transferring the resources while circumventing the frictions in the external borrow- ings. As a result, it is optimal to allow direct investors to purchase domestic capitals at least until the borrowing in the form of debt alters to a saving. How is it related to reserve accumulation? The social planner in our model does not limit direct investments until households begin lending abroad. Also, we know that we may have the optimal reserve accumulation at positive amounts when households save. Hence if the social planner can do both reserve accumulation and control of direct investments, the social planner is likely to accumulate reserves. It is more likely if we add some positive externalities from direct investments, which we do not consider in this paper. 49 Full characterization of ¶b 2 ¶q j f<f c and conditions to guarantee the assumptions in lemma 1 are in appendix 165 3.2.2.3 Reserve accumulation and currency manipulation In this subsection, we characterize the optimal reserve accumulation more specifically as a func- tion of direct investment capital inflows. It uncovers the nature of the reserve accumulation as a macroprudential policy tool against sudden stop and vividly maps our model to the empirical find- ing in section 2. In addition, the new proposition will provide important implications for the policy debate of currency manipulation. As a first step, we denote b 1 = b 1 (q;R). That is, b 1 is a function of q and R. From Remark 1, we know ¶b 1 ¶q > 0 and ¶b 1 ¶R < 0. And recall the FOC of reserve accumulation without capital control. u T 0 b(1+ r)E u T 1 = b dw 1 dR 1 " Z f c f d(−b 2 ) dw 1 u T 1 b(1+ r 2 )u T 2 dF f −bG b E u T 2 b 2 db 2 dw 1 # bG s E u T 1 b 1 db 1 dR 1 (3.28) Rearranging the equation (3.28) and using u T 0 b(1+ r 1 )E u T 1 = 0 give us r −rG s b 1 (q) 1 db 1 (q) dR 1 E u T 1 = dw 1 dR 1 " Z f c f d(−b 2 ) dw 1 u T 1 b(1+ r 2 )u T 2 dF f −bG b E u T 2 b 2 db 2 dw 1 # (3.29) See the LHS in the equation (3.29) decreases inq. However, the RHS is always positive as long as b 2 < 0. 50 We can think of the LHS as adjusted marginal costs and RHS as adjusted marginal benefits accordingly. Now defineq c such that the LHS in equation (3.29) is zero. That is r −rG s b 1 (q c ) 1 db 1 (q c ) dR 1 = 0 (3.30) 50 If b 2 is negative in any state, then it is negative in all the states. The only stochastic variable is f. Once the decision is to borrow, lowerf only yields lower borrowing. 166 At q =q c , the social planner must accumulate reserves because the marginal cost is zero while the marginal benefit is positive. Thus q c is a cut-off of direct investment above which the social planner must accumulate reserves. Along with a few other related analytical results, we introduce our third proposition. Proposition 6 (Passive Reserve Accumulation) Let Q 0 = Q and define q c such that r r = G s b 1 (q c ) 1 db 1 (q c ) dR 1 . Then we have 1) There existsd 0 such that forq >q c d 0 , R 1 > 0 and R 1 increases inq 2) R 1 =(Q 0 A 0 )(qq c )K+b 1 (q c ;0) b 1 (q;R 1 )+ c T 0 (q c ;0) c T 0 (q;R 1 ) 3) b 1 (q c ;0)> b 1 (q;R 1 ) 4) There exists d 1 such that c T 0 (q c ;0) c T 0 (q;R 1 ) > 0 for q2(q;q+d 1 ). Therefore R 1 > (Q 0 A 0 )(qq c )K forq2(q c ;q c +d 1 ). Proof) See the Appendix C. Broadly speaking, the proposition 3 illustrates the existence of the cut-off and the amounts of reserve accumulation are bounded below by (Q 0 A 0 )(qq c )K. Now assume that d is small and b 1 (q c ;0) b 1 (q;R )' 0 then we have R 1 '(Q 0 A 0 )(qq c )K (3.31) More importantly, the real exchange rate in period 0 becomes almost invariant toq. More formally, define p 0 = 1a a y T 0 +(Q 0 A 0 )qK d 0 + b 1 (q) R(q) y N 0 (3.32) ' 1a a y T 0 +(Q 0 A 0 )q c K d 0 + b 1 (q c ) y N 0 Then p 0 is almost invariant to q2(q c ;1). In other words, because the social planner absorbs the extra liquidity from direct investments by accumulating more reserves, the real exchange rate turns out to be almost “invariant” to direct investment capital inflows. Figure 3.11 illustrates such 167 “passive” reserve accumulations and invariant real exchange rates under the passive reserve accu- mulation. To explain more, the social planner facing direct investment flows above a certain level “passively” absorbs the extra inflows beyond the level: The reserve accumulation increases by the almost same amount of the increase in direct investments. We explain more about such a passive reserve accumulation and following invariant real exchange rates in the discussion of currency ma- nipulation below. Before we move on to the discussion, we restate the findings in the remark 2. Remark 2 For direct investments inflows beyond a certain level, 1) the social planner must accu- mulate reserves, and 2) for direct investments more than the level, the planner absorbs the extra inflows by accumulating reserves and thereby making the real exchange rates almost invariant to the capital inflows. Remark 2 and Figure 11 map our model to the empirical regularities in section 2. The key finding in section 2 is the close relationship between reserve accumulation and extra capital inflows that we defined as a summation of current account surplus and net inflows of direct investment and equity portfolio investment. The equation (3.31) is the theoretical counterpart to the empirical regularity. Furthermore, once we posit the direct investment capital flows and equity portfolio flows as given, we can explain the evolution of international reserve holding of EMEs for the last two decades: as more direct investment and equity portfolio capitals flow into EMEs, many of the EMEs absorb the capital inflows by accumulating reserves. Also, we can explain much of the cross country difference of reserve holdings: The more direct investments or equity portfolio investments EMEs receive, the more reserves the EMEs accumulate. The model predicts that EMEs with more direct investment or equity portfolio external liabilities accumulate more reserves as we saw in the data in section3.2. 51 Another important parameter to explain the reserve accumulation isG s . The model predicts lowerG s induces more reserve accumulation. Unfortunately, we have no clear idea of what determinesG s or the empirical counterpart of the parameter. Despite such difficulty, 51 The EMEs holding reserves more than 40% of GDP, such as Malaysia, Thailand or Bulgaria also have sizable direct investment and equity portfolio investment external liabilities. 168 Figure 3.11: Passive Reserve Accumulation Note: A 0.1 unit ofq approximately corresponds to the inflow in the amount of the 9.5% of GDP. All the parameter values are the same as the benchmark except for distribution of credit coefficient(=Beta(1.2,5)): it seems that the recent divergence between private sector external assets and official reserves in EMEs, which we showed in section 3.2, is a result of less friction on overseas investment by the private sectors: that is, lowerG s . For more discussion and related empirical findings, we relegate it to the appendix. Discussion of currency manipulation Reserve accumulation is often viewed as evidence that some EMEs depreciate their currencies to boost their exports, and such one side interventions to depreciate domestic currencies are often called “currency manipulation.” For example, one of the criteria that the US treasury examines to judge whether an EME is manipulating their currency value is the amounts of reserve accumulation; detailed information is in table 2 below. 169 Table 3.2: The U.S Treasury’s Foreign Exchange Report Each country reports shall contain: (1) country's bilateral trade balance with the United States (2) country's current account balance as a percentage of its GDP ... (4) country's foreign exchange reserves as a percentage of its short-term debt (5) country's foreign exchange reserves as a percentage of its GDP Enhanced analysis shall include: (1) a significant bilateral trade surplus with the United States (2) a material current account surplus (3) engaged in persistent one-sided intervention in the foreign exchange market Source: Section 701 of the Trade Facilitation and Trade Enforcement Act of 2015 In the literature, it has been often argued that international reserve holdings at extraordinary large amounts cannot be justified by the precautionary view and therefore the large amounts of reserves are a byproduct of the export-oriented growth strategy; for example, some East Asian countries such as China, Thailand, or Malaysia. The underlying idea is that reserve holdings in such countries seem to be much more than needed for a precautionary purpose. Here we argue that the simple international reserve to GDP ratio is not a correct measure to identify a currency manipulation. If the apparent excessive reserve accumulation is caused by massive capital inflows, then the purpose of the reserve accumulation can be precautionary. Of course, we do not analyze or discuss the currency manipulation itself or try to examine whether certain EMEs manipulate their currencies in the spirit of mercantilism. We only aim at explaining why amounts of accumulated reserves are not a good “litmus” to test currency manipulation. 52 The empirical facts about reserve accumulation that we find in section 2 and proposition 3 altogether imply that EMEs facing massive capital inflows make corresponding capital outflows in the form of reserves to maintain the macroeconomic balance. The purpose of reserve accumulation is to reallocate the resources from capital inflows so as to minimize the inefficiencies of capital outflows by private sectors and be better prepared for possible sudden stops in the future. All the 52 For the theoretical exploration of currency manipulation, see Hassan et al. (2019). For empirical studies, see Dominguez (2019). 170 motivation of reserve accumulation lies in the precautionary purpose. The social planner in our model does not have an idea of mercantilism since we do not put any ingredients related to the mercantilism, such as positive externalities from export sectors. Figure 3.12 below well illustrates it. First, in the absence of reserve accumulation, the direct investment flows generate consumption booms. Asq increases, the households’ saving increases, but not enough due to falling returns for the households, which appreciates the real exchange rate through more tradable goods consumption. Therefore, although the direct investments lower the probability of sudden stop, it falls slowly as the households cannot invest abroad enough. On the contrary, when the social planner accumulates reserves, more capital outflows are made and it keeps tradable goods consumption constant and hence real exchange as well. More importantly, thanks to the more overseas investments, the sudden stop probability falls much faster. As a result, the equilibrium with reserve accumulation is similar to little or no friction on capital outflows; that is, a very smallG s . Figure 3.12: Reserve Accumulation and Currency Manipulation Note: All the parameter values are the same as the benchmark except for distribution of credit coefficient(=Beta(1.2,5)). 171 In this regard, reserve accumulation is a way to “restore” the macroeconomic balance under little direct investment inflows or to achieve the balance under frictionless capital outflows. The real exchange rate might be a target for the planner since it measures over-consumption in the present, but the purpose of the intervention is to prevent “appreciations” of the currency, not to “depreciate” the currency. The current account ( y T 0 c T 0 ) neither improves or becomes worsen as more reserves are accumulated, as opposed to a common prediction from the mercantilism view. Such patterns of reserve accumulation in our model correspond to observations in Levy-Yeyati (2008) and Levy-Yeyati et al. (2013), which document that foreign exchange market interventions seem to aim at limiting domestic currency appreciations. Also, the patterns are in the same line with the famous finding in Calvo and Reinhart (2002) so-called “Fear of Floating” in a sense that real exchange rates in our model economy may look somehow managed in the eyes of outsiders to the economy. However, the motivation for reserve accumulation in our model lies in the precautionary purpose as it is designed so. Therefore, our model reconciles the precautionary view of reserve accumulation with empirical findings in the managed float literature. Back to our discussion of currency manipulation, our theory implies that evidence supporting the managed float exchange regime does not necessarily support the mercantilist view and the evidence can be aligned with the precautionary view. 3.3 Extensions In the last section, we maintained the simplest structures in the models to make the model tractable. In this section, we introduce three extensions of the baseline model. In each extension, strong assumptions in the baseline model are alleviated and it turns out that our key insights and results survive in the more general environments. 172 3.3.1 The Model with Heterogeneous Agents We introduce a heterogeneous agents model. Our goal in the model is to show how our main results in the representative agent model can survive in the new heterogeneous agents model rather than solve the heterogeneous model fully. We borrow some features from Korinek and Sandri (2016). The environments of productions, international financial intermediations (IFIs), and direct in- vestors are same as the baseline model. However, there are two heterogeneous agents in the small open economy: Borrower and Saver. We may think that savers are the group who receives direct investments; hence they sold their tradable goods capitals to foreign direct investors. The savers need to lend their tradable goods to the borrowers or invest abroad. Similarly, the borrowers can borrow from either IFIs or the savers. We denote the total borrowing of the borrowers by b b t+1 and similarly the total saving of the savers by b s t+1 . To differentiate the borrowings from IFIs from the total borrowings, we let b b t+1 be the borrowing from IFIs. Similarly, we define the overseas investment by the savers as b s t+1 . The social planner issues their own bonds to accumulate reserves. Let’s denote it by b g t+1 : Similarly with b b t+1 , define b g t+1 . The market clearing condition of the domestic funds market is as below. b b t+1 b b t+1 + b g t+1 b g t+1 + b s t+1 b s t+1 = 0 (3.33) That is, the total demand for the tradable goods borrowing in the domestic market is b b t+1 b b t+1 + b g t+1 b g t+1 , while b s t+1 b s t+1 is the supply from the savers in the domestic market. The description of flows of funds is provided in Figure 13 below. To clear the market, we need one more market clearing condition, by which borrowers (savers) are indifferent between borrowing from IFIs and savers (lending abroad and lending to the borrow- ers). To have the condition, let’s assume that the yields from investing abroad without the fee to IFIs are higher than the borrowing rates. The net returns to the savers are characterized in the same way, but now r t+1 =G s b s t+1 + r (3.34) 173 Figure 3.13: Flow of Funds Similarly, for the borrowing rates from IFIs, r t+1 =G b b b t+1 + b g t+1 + r (3.35) Since b b t+1 < 0 and b s t+1 > 0, for the market to be cleared, we need r > r 53 . In period 1, the credit constraint can bind and it is as follows. −b b t+1 f t (1q t )y T t + p t y N t (3.36) That is, the amount of the total external debt by the borrowers is constrained by the aggregate GDP. The credit constraint does not include b g t+1 since the social planner does not borrow abroad when the credit constraint binds 54 . Furthermore, during a sudden stop, nontradable goods are re- ally cheap so that savers dispose of all their assets abroad to consume more nontradable good; the 53 The r is the required rate for an economy who have nearly do external debt in terms of gross. Hence it must be low. 54 It is different in a more dynamic version of the model as it is explained in the next subsection. 174 retrenchment in Forbes and Warnock (2012). 55 Hence, we can imagine that from period 1, the dif- ferent groups of the agents merge into one big family so that the model backs to the representative agent model. The return to reserves, b g t+1 is low at r< r . Thus, the planner has to collect tax from the agents to pay the extra interest rates, r t+1 r. We assume the tax is imposed optimally so that it does not make any extra terms in the first order condition of the reserve accumulation, which we introduce below. It turns out to be optimal to impose the tax on the savers under reasonable parameter values. Taking all the changes, we derive the first-order condition for the optimal reserve accumulation in period 0 56 . bG b E h u b 1;T i b b 1 d(b b 1 + b g 1 ) dR 1 | {z } Higher r 1 + u s 0;T b(1+ r)E u s 1;T | {z } ConsumptionWedge at r = dw 1 dR 1 2 6 6 6 4 å i=b;s Z f c f b d −b i 2 dw 1 u i 1;T b(1+ r 2 )u i 2;T dF f | {z } Marginal Value o f Borrowing −bG b E u i 2;T b i 2 db i 2 dw 1 | {z } Lower r 2 3 7 7 7 5 bG s E u s 1;T b s 1 db s 1 dR 1 | {z } Higher r 1 (3.37) where w 1 =b g 1 (1+ r)+ b b t+1 + b g t+1 + b s t+1 (1+ r 1 ), R 1 =b g 1 . and dw 1 dR 1 =(r 1 r) db g 1 dR 1 + (1+ r 1 ) db b 1 dR 1 + db s 1 dR 1 b b t+1 + b g t+1 + b s t+1 G b db b 1 dR 1 + db g 1 dR 1 . From (27), we see b b t+1 + b g t+1 + b s t+1 = b b t+1 + b g t+1 + b s t+1 . Then dw 1 dR 1 =(1+ r)+(1+ r 1 ) 1+ db b 1 dR 1 + db s 1 dR 1 b b t+1 + b g t+1 + b s t+1 G b db b 1 dR 1 + db g 1 dR 1 !! (3.38) See db j 1 dR 1 > 0 for j = s;b. Therefore, for G b ;G s large enough, dw 1 dR 1 > 0. The mechanism of how reserve accumulation improves the net foreign asset position is analogous to the baseline model. 55 Such a nice retrenchment does not strictly hold in the reality due to risk hedging motives or low confidence of the savers about the economy. 56 We assume the planner assign equal weights to each of the borrower and the saver. 175 Since it raises the borrowing rates and therefore returns to the savers as well, the interventions to accumulate reserves discourage the borrowing, but encourage the saving. As one might expect, the mechanism in proposition 1 and proposition 3 applies. As direct in- vestment capitals inflows,b b 1 decreases, butb s 1 increases. It reduces marginal costs of reserve accumulation while raising the benefits. Therefore, we have the following proposition. Proposition 7 The optimal reserve accumulation in the heterogeneous agents model is character- ized as follows. 1) It is characterized by the first-order condition " r −rG s b s 1 (q)+G s b s 1 db s 1 (q) dR 1 # E u s 1;T +G b b b 1 (q) d(b b 1 (q)+ b g 1 ) dR 1 E h u b 1;T i = dw 1 dR 1 " å i=b;s Z f c f b d −b i 2 dw 1 u i 1;T b(1+ r 2 )u i 2;T dF f −bG b E u i 2;T b i 2 db i 2 dw 1 # (3.39) 2) There existd such that forq >q c d, R 1 > 0 and R 1 increases inq 3) There existsd 1 such that R 1 >(Q 0 A 0 )(qq c )K forq2(q c ;q c +d 1 ). Proof) See the discussion above. Proposition 4 is analogous to proposition 3. Therefore, our analytics and insights in the baseline model survive and hold in the heterogeneous agents model. As one might expect, more direct investment capital inflows — higherq — make more borrowers switch to savers and the increase in the share of savers in the economy force the planner to accumulate more reserves. 3.3.2 The Infinite Horizon Model We illustrate an infinite horizon model as an extension of the baseline model. Same as the het- erogeneous model, we aim at proving that our key insights survive in the new environment. In addition, we introduce analytical results about reserve depletions during sudden stop and the time- inconsistency of reserve management policy. 176 The economy is populated by a continuum of identical, infinitely-lived households of measure unity with preferences given by: E 0 ( ¥ å t=0 b t u(c t ) ) (3.40) where c t = f c T t ;c N t . Vector of endowments given by y y y T ;y N 2 Y R 2 ++ follows a first- order Markov process. The social planner taxes households by the amount T to accumulate re- serves; of course, the planner grants subsidies when decumulating reserves. Using notations in the baseline model, the budget constraint of household is as below. b t+1 + c T t + p t c N t + T t = b t (1+ r t )+ Q t (q t q t1 )K T +(1q t )y T t + p t y N t (3.41) whereq t is the share of foreign investors in domestic capital markets. For simplicity, we assume that the capital price is determined through the same bargaining process in the baseline model. We assume thatq t is time-variant, but deterministic. Same as the three period model, the household faces the credit constraint, but now in every period. That is, 57 b t+1 f y t (1q t )y T t + p t y N t +f R R t+1 (3.42) where f t follows a stochastic process. 58 Following Shousha (2017), we assume that holding re- serves of R t+1 can relax the credit constraint by the amount off R R t+1 ˙ footnoteFor the mechanism of how liquid financial assets can works as a collateral, see Gottardi et al. (2017) and Parlatore (2019) The idea behind the new assumption is that holding reserves gives EMEs a better reputation or it can be used as a collateral as US treasury billls are used as collaterals in the repo market in the 57 IR t is the amount of reserves after government depletes its reserves. 58 We don’t specifiy the process off t , but the process off t should include some hazardous events. In this sense,f t can follow a first-order Markov process whose Markov chain includes disaster states described in Barro (2006). More generally, f t can be the summation of a “normal” AR(1) process and a compound Poisson process, which delivers a negative random shock with a certain probability in every period. 177 reality. f R 2(0;1) 59 and unlikef y t , it is constant because the collateral value of reserves composed of safe assets is highly invariant to global financial conditions. The social planner accumulates reserves in the same way with the baseline model. Again, we define w t b t (1+ r t )+ R t . This is net foreign asset position in terms of liquid assets. Then the problem of the planner is formulated as follows. V(w;y;f;q)= max R 0 u(c T ;y N )+bE V 0 w 0 ;y 0 ;f 0 ;q 0 s:t: c T =(1q)y T +Q(qq 1 )K T −b 0 + b(1+ r(b))+ R(1+ r)−R 0 b 0 = 8 > > < > > : solution of the credit constraint i fff c solution of the euler equation i ff >f c We find that it is convenient to formulate the valuation in the following way. V(w;y;f;q)= max R 0 u(c T ;y N )+bE u c 0 T ;y 0 N +V 00 b 00 ;R 00 ;y 00 ;f 00 ;q 00 Notice that we substituted b 00 ;R 00 for w 00 . Assuming the credit constraint does not bind, the first order condition is r rGb 0 db 0 dR 0 E u 0 T = dw 0 dR 0 Pr f 0 f 0c E f 0 f 0c ¶b 00 c ¶w 0 u 0 T +b ¶V 00 ¶b 00 c + Pr f 0 >f 0c E f 0 >f 0c ¶b 00 u ¶w 0 u 0 T +b ¶V 00 ¶b 00 u (3.43) 59 Theoretically, it is not impossible to have f R > 1 as it is in Shousha (2017). However, we exclude such a case because of legal structures on the reserves. As it is noted in Alfaro and Kanczuk (2009), which rejected the idea that reserves can be used as collateral, under the Foreign Sovereign Immunities Act of 1976 of the United States and similar laws in other countries, central bank assets, including international reserves, are usually protected against attachment. This means that reserves can be accepted as collateral only if EME governments are willing to pledge as such. Although it is highly unrealistic to think EMEs will default leaving so much reserves behind, we restrictf R to be smaller than 1. 178 where db 0 dR 0 = ¶b 0 ¶R 0 + ¶b 0 ¶R 00 ¶R 00 ¶R 0 and dw 0 dR 0 = ¶w 0 ¶R 0 + ¶w 0 ¶b 0 db 0 dR 0 = 1+ 1+ r 2G j db 0 dR 0 . It is easy to notice that equation (3.43) is analogous to equation (15), (16) in the baseline model. Once we replace V 00 with u 00 T , it is identical to equation (3.21), (3.22). Another difference is ¶b 0 ¶R 00 ¶R 00 ¶R 0 in db 0 dR 0 = ¶b 0 ¶R 0 + ¶b 0 ¶R 00 ¶R 00 ¶R 0 . This term emerges due to the time-inconsistency we will discuss below. Overall, the reserve accumulation in (3.43) is much analogous to the one in proposition 8. There exists a cut- off of direct investment above which the social planner always wants to accumulate reserves, and as more direct investment capital inflows, the planner passively accumulates reserves to maintain the macroeconomic balance and thereby making the economy more prudential to sudden stops in the future. Reserve depletion during sudden stops and the Time-inconsistency First, we characterize the reserve depletion by the social planner during sudden stop. Using equation (3.43), we can easily derive the first-order condition of reserve depletion. That is bGb 0 E u 0 T db 0 dR 0 + u T b(1+ r)E u 0 T + u T b 1+ r 0 E u 0 T db 0 dR 0 = dw 0 dR 0 Pr f 0 f 0c bE f 0 f 0c ¶b 00 c ¶w 0 u 0 T +b ¶V 00 ¶b 00 c + Pr f 0 >f 0c bE f 0 >f 0c ¶b 00 u ¶w 0 u 0 T +b ¶V 00 ¶b 00 u (3.44) The equation (3.44) is identical to the equation (3.43) except for [u T b(1+ r 0 )E[u 0 T ]] db 0 dR 0 . For our convenience, defineL(r) u T b(1+ r)E[u 0 T ] andL(r 0 ) u T b(1+ r 0 )E[u 0 T ]. Then both L(r) andL(r 0 ) decrease in R 0 . See ¶c 0 T ¶R 0 =1− ¶b ¶R 0 0 < 0 since 0< ¶b ¶R 0 0 f y ¶ p ¶R 0 y N +f R <f R < 1. Hence, holding reserves (less depletion) reduces the current consumptions — hence, higher wedges in the consumptions —, but leaves more resources for the future. With these comparative statics, we characterize the reserve depletions during sudden stop in the proposition 8. Proposition 8 Suppose the EME is under a sudden stop crisis. That is,f <f c : Then the reserve holding under the sudden stop R 0 has following comparative statics. 1) R 0 increases inf. That is, ¶R 0 ¶f > 0. 179 2) Both of H(f 0 ) and G(f 0 ) are the CDFs off 0 , and H(f 0 )< G(f 0 ). Then R 0 is larger under H(f 0 ) than G(f 0 ). Proof) See the Appendix C. The interpretation of lemma 3 is straightforward. If the ongoing crisis is more severe, more reserves are depleted (less reserve holding). If the crisis becomes more persistent, i.e.,f 0 becomes smaller in all the states, givenf, then the less reserves are depleted (more reserve holding). To explain more, during a sudden stop, the planner decides whether to deplete reserves to grant it to households or hold it to use as collateral. Since we assumef R < 1, reserve depletion gives more liquidity so as to raise current consumption of tradable goods, while the depletion leaves fewer resources for the future. Hence holding reserves is a sort of auxiliary policy, which is painful now, but desirable for the preparation for the possible crisis in the future. The use of reserve during sudden stops are described in Figure 14 below. The use of reserve in Figure corresponds to the empirical findings in Aizenman and Sun (2012) that EMEs during the Global Financial Crisis hesitated to deplete reserves, and similar patterns of using reserves during 2013 tapering tantrum and recent market turmoil in Brazil and Chile. 60 Figure 3.14: Reserve Depletion during Sudden Stops 60 The central banks in Barzil and Chile announced that they will lend reserves to the domestic banks and will recover the reserves by the time they announce. 180 Proposition 8 tells us that there is some discretion for the planner to use reserves during sudden stops. It implies that using reserves is a particular bail-out policy under discretion. A usual side effect from such ex-ante interventions with discretion is time-inconsistency. As one might expect, the bail-out in lemma 3 is not time-consistent. Furthermore, it turns out that reserve management policy, both reserve accumulation and depletion is in general time-inconsistent. We introduce the following proposition. Lemma 3 (Time-inconsistency) Denote a state in period t+k byw t+k . DefinefT t+k (w t+k )g 0k¥ as the reserve management policy function of the planner. Similarly, letfT t+k (w t+k )g 0k¥ be the reserve management policy function of the planner with a commitment power. Then, T t+k (w t+k )6= T t+k (w t+k ). Proof) See the Appendix C. Therefore the reserve management policy is time-inconsistent in both of accumulation and deple- tion. The reserve management in the future will impact the borrowing or saving by households today, but the planner without commitment power has no way to enforce the future planner to take account of such an effect. Lemma 3 above only tells us that the reserve management in general is not time-consistent. It does not suggest a “direction” of the inconsistency: is it underaccumulation or overaccumula- tion? or underdepletion or overdepletion? Unfortunately, we have no clear-cut answers to these questions. We have such ambiguous results because the direction of externality is not determin- istic. If the EME suffers from overborrowing or undersaving, promising more accumulations and less depletions during sudden stops is time-consistent so that the households borrow less or save more responding to higher marginal utilities in the future. On the contrary, if the EME is under oversaving by the households, then it should be time-consistent to promise less accumulations and more depletions since it brings down the saving today through lower marginal utilities in the future. Since whether the EME is in a state of oversaving or undersaving varies along with time and state, 181 it is not trivial to find the directions of the time-inconsistency. Although it would be an interesting theoretical point, we do not analyze further as it is not our main focus in this paper. 3.3.3 Endogenous Direct Investments and Capital Price Finally, we look at the extension where the direct investment and capital price are endogenous. Same as the last two extensions, we describe our new environment and then show why our key results do not change rather than solve the model fully. In the reality, direct investment in an EME depends on various factors such as locations, natural resources, or macroeconomic stabili- ties. Obviously, still we cannot reflect all the realistic features. We simply assume that the direct investment is an increasing function of the profitability of the investment. First, define p T = å 2 t=0 M t 0 A t Q 0 where M t 0 is the discount factor of the investor. p T is the gross return rate to the direct investment in the tradable goods sector capital. Then we can assume qQ 0 K T = F T (p T ) (3.45) Hence, how much direct investors purchase the tradable goods capitals depends on the profitability. Of course, we assume F 0 > 0. We also endogenize the capital price. To model the capital price, we can posit that the domestic capital market is perfectly competitive. Then the capital price must be Q 0 = 2 å t=0 b t u T t u T 0 A t = A 0 + A 1 1+ r 1 +E A 2 (1+ r 1 )(1+ r 2 ) 182 We can easily see the reserve accumulation will impact the capital price through the changes in the interest rates; the borrowing rates and returns to the saving are the linear functions of the borrowing and saving, which are changed by reserve accumulation. Then from previous sections, once b 1 > 0 we expect r 1 increases in reserve accumulation, while r 2 changes in the opposite ways to the reserve accumulation, depending on whether the credit constraint binds or not. However, with plausible parameter values, we can expect the effect through r 1 dominates. Hence, we reasonably conjectures ¶Q 0 ¶R 1 < 0 Then of course,q increases in reserve accumulation R 1 because the lower capital price boosts the profitability of the direct investment. It provides two opposite different implications of reserve accumulations. First, the reserve accumulation by the social planner would have a sort of self-multiplication effect. As we saw in the last section, reserve accumulation increases in the amounts of direct investment. Here the reserve accumulation attracts more direct investments which in turn calls for more reserve accumulation. Second, in the other way, reserve accumulation discounts the capital price, which is unfavorable for the EME. However, such negative impacts, which might be small, could be offset by possible positive effects, which we abstract in our model. For example, any knowledge spillover effects can be probably more than offsetting the negative impacts. Overall, although it is hard to make a strong assumption since the environment in our model is not rich enough, the optimal reserve accumulation could rise once we let the direct investment and the capital price be endogenous. Another strict restriction imposed in our model is that we exclude direct investments in the nontradable goods sector. Now we allow the investments in the nontradable goods capital. Let’s denote the share of foreign direct investors in the domestic tradable goods and nontradable gods capital markets byq T andq N respectively. Direct investors interested in holding K N also decide 183 their investments based on the expected profitability. The crucial difference is the investors con- vert the returns of nontradable goods to tradable goods. 61 Hence the expected profitability of the nontradable goods capital investments is p N =E 0 " å 2 t=0 M t 0 p t A t p 0 Q 0 # Notice that the expected profit increases in the expected appreciationE h p t>0 p 0 i . Also, from previous results, we can notice d(E[p t p 0 ]) dR 1 > 0 f or t= 1;2 Intuitively, the reserve accumulation has effects of increasing net foreign assets in the future, which raise the price of the nontradable goods prices in the future, but lowers the price in the present. Therefore, the direct investment in nontradable goods sector increases in the reserve accumulation. This is the mechanism examined in handful papers that studied the relationship between foreign direct investment and reserve accumulation. Matsumoto (2019) argued that EMEs accumulate re- serves to attract more direct investments. In contrast, we provide the causality in the opposite direc- tion: direct investments cause reserve accumulations. However, we also have a similar mechanism with Matsumoto (2018). Reserve accumulation causes the currency depreciation, while promising appreciations in the future, and therefore it is more profitable to invest in the EME. But, we do not believe that EMEs accumulate reserves to attract direct investments as we discussed in the litera- ture review section. 62 More plausible scenario is that EMEs facing lots of direct investment capital inflows make corresponding outflows in the form of reserve and the reserve accumulation again attracts more direct investments, which calls for even more reserve accumulation. 61 Such assumption is common in the local currency sovereign debt literature. 62 Another difficulty in the explanation that reserve is accumulated to attract direct investments through currency depreciation is a possibility that EME policy authorities may depreciate their currency in the future to dilute “real liabilities” measured in foreign currencies. If currency valuations are truly important in making direct investment decisions, such a time-inconsistent incentive to depreciate currency in the future must significantly impugn the direct investors’ interest so as to make the reserve accumulation less useful policy to attract direct investments. Such an ex-ante desire to depreciate currency is widely studied in the local currency sovereign debt literature; see Du et al. (2016), Engel and Park (2018), and Perez and Ottonello (2019). These papers concluded that desires to dilute local currency significantly limit the gain from local currency sovereign debts. 184 We close this section summarizing our findings in the remark below. Remark 3 Once we endogenize direct investment and the capital price, and allow nontradable goods sector capital investment as above, we have following properties. 1) Reserve accumulation makes it more profitable for foreign direct investors to purchase do- mestic capitals since it discounts the current capital price through lower domestic interest rates or lower real exchange rates. 2) Therefore, reserve accumulation may attract more direct investments, which in turn incen- tives EMEs to accumulate even more reserves. As a result, we have a sort of loop mechanism, by which magnifies both of direct investment inflows and reserve outflows. 3.4 Concluding Remarks In this paper, we provide a novel theory of reserve accumulation of EME. Our view of the reserves accumulation of EMEs is that it is the capital outflows by public sectors to supplement insufficient capital outflows by private sectors. When an EME receives excessively large amounts of capital inflows in the form of direct investment or equity portfolio investment, the EME needs to invest abroad to maintain the macroeconomic balance. If private sectors in the EME have a problem making overseas investments enough due to some reasons such as reliance on foreign financial intermediations for the overseas investment, policy authorities in the EME need to invest abroad instead of the inefficient private sectors. Therefore, policy authorities accumulate reserves to pump out excessive foreign currency liquidity from the certain types of capital inflows. Our theory is motivated by empirical regularities we found. We found that reserve accumu- lation of EMEs is positively and strongly correlated with 1) capital inflows in the types of direct investment and equity portfolio investment, 2) current account surpluses and 3) capital outflows by private sectors. Furthermore, the extra capital inflows we defined as the summation of current account surplus and net direct investment and equity portfolio investment inflows show highly pos- itive correlations with reserve accumulation. To describe the empirical facts we found, both reserve 185 outflows and private sector capital outflows increase when EMEs receive more direct investment or equity portfolio investment capital inflows, or have higher current account surpluses. Motivated by the facts, we construct a tractable three period model. Our model is built on the framework of Fisherian deflation model developed in the capital control literature. Into the framework, we incorporated imperfect capital mobility for both debt capital inflows and debt cap- ital outflows, and direct investment capital inflows. The imperfect capital mobility allows us to overcome the Ricardian equivalence critique on reserve accumulation. However, reserve accumu- lation turns out to be an inefficient policy tool in environments where private agents borrow abroad; the Ricardian equivalence is broken due to higher borrowing rates and therefore it is very costly. However, when EMEs receive large amounts of direct investment inflows so that the decision of the private agents alters from borrowing abroad to saving abroad, the efficiency of reserve accu- mulation is dramatically raised. Because more overseas investments by the private sectors create more negative externality from it, less investment by the private agents responding to reserve ac- cumulation is beneficial to the economy. Moreover, the reserve accumulation is not substitutable by capital controls — subsidies or taxes on overseas investment when saving abroad — once the intertemporal maximization of the private agents is to save, whereas it is it perfectly substitutable if the decision is to borrow. 63 This result is driven in our simple model, but our key insights survive or even become stronger although we generalize our environments in several different ways. Our model also provides important implications about ongoing debates of the currency manip- ulation. In our model, when EMEs receive direct investment capital flows beyond a certain level, the EMEs passively accumulate reserves by the amounts of the direct investment flows above the level. In terms of the currency valuation, the intervention to accumulate reserves works in a way of preventing the currencies from appreciating although it is not an object itself. Such patterns correspond to the empirical findings in Levy-Yeyati et al. (2013); EMEs intervene in the foreign exchange market to prevent currency appreciation, not to depreciate the currency. The intervention to discourage the currency appreciation is much different from descriptions in a common criticism 63 Again, we note that this result is assuming the capital control is perfectly efficient, which is not in the reality. Hence, the capital control should not be a perfect substitute for reserve accumulation in the reality. 186 on the reserve accumulation that EMEs accumulate reserves to depreciate their currencies to boost their exports. Moreover, our model predicts that EMEs receiving excessively large amounts of capital inflows can accumulate seemingly excessive reserves. Therefore, our model implies that the amount of reserve holding is not a good litmus for the test of currency manipulation. In this paper, we used a tractable three period model and abstracted from important features in the reality to clarify our novel key insights. One obvious way of developing the idea in this paper is to construct a medium scale DSGE model embedded with our key ideas. In richer environments with production technology or more sophisticated modeling of direct investment, we can try to quantify the optimal reserve accumulation and compare the outcomes from the DSGE model to the actual levels of reserve holding in EMEs. Another direction in which we can develop the ideas in this paper is to search more for private overseas investment in EMEs. To the best of our knowledge, our paper is the first in modeling frictions on overseas investment in EMEs and we adopted a way of giving us a nice and tractable result. Although studies of external assets of EMEs are very scant in the literature, external assets of EMEs except for international reserves are not negligible any more and we need to understand it more properly. Deeper understanding of EMEs’ overseas investment will certainly give important implications about reserve accumulation and of course, it can be important for the understanding of international monetary system as well. 64 Lastly, it will be also interesting to extend our model to incorporate the mercantilist view. In our model, large capital inflows in the form of direct investment or equity portfolio investment generate domestic booms in consumptions, but no changes in production since the model economy is de-facto an endowment economy. However, in a richer environment with different sectors with production technologies, domestic currency appreciations caused by the capital inflows must impact the productions in different sectors. It is not easy to predict what would be an outcome, but our prior is the domestic currency appreciation will result in a reallocation of domestic resources: It extends nontradable goods sectors, while causing contractions of tradable goods sectors. This prediction corresponds to the empirical findings in Benigno et al. (2016). In such environments, the planner accumulate 64 Recently, Horn et al. (2019) empirically study China’s overseas lending. 187 reserves to prevent domestic currency appreciations so as to restore the efficient resource allocation among different sectors. We believe all the issues above give us hard, but interesting questions unanswered in this paper. We leave these issues to future research. 188 References [1] Acharya, V . V . and Krishnamurthy, A. 2018. “Capital Flow Management with Multiple In- struments.” NBER Working Paper, No. 24443. [2] Aizenman, J. 2011. “Hoarding International Reserves versus a Pigovian Tax-cum-Subsidy Scheme.” Journal of Economic Dynamics and Control, 35 (9): 1502–1513. [3] Aizenman, J. and Lee, J. 2007. “International Reserves: Precautionary Versus Mercantilist Views, Theory and Evidence.” Open Economies Review, 18 (2): 191–214. [4] Aizenman, J. and Pasricha, G. K. 2007. “Why do emerging markets liberalize capital out- flow controls? Fiscal versus net capital flow concerns.” Journal of International Money and Finance 39: 28–64. [5] Aizenman, J. and Sun, Yi. 2012. “The Financial Crisis and Sizable International Reserves Depletion: From Fear of Floating to the Fear of Losing International Reserves?” International Review of Economics & Finance, 24: 250–269. [6] Alfaro, L., and Kanczuk, F. 2009. “Optimal Reserve Management and Sovereign Debt.” Jour- nal of International Economics, 77 (1): 23–36. [7] ——. 2013. “Debt Redemption and Reserve Accumulation.” NBER Working Paper, No. 19098. [8] Alfaro, Laura, Sebnem Kalemli-Ozcan, Vadym V olosovych. 2014 “Sovereigns, Upstream Capital Flows, and Global Imbalances.” Journal of the European Economic Association, 12 (5): 1240–1284. [9] Alvarez, F., Atkeson A., and Kehoe, P. J. 2009. “Time-Varying Risk, Interest Rates, and Exchange Rates in General Equilibrium,” Review of Economic Studies, 76 (3): 851–878. [10] Antras, P. and S. R. Yeaple. 2013. “Multinational Firms and the Structure of International Trade,” In Handbook of International Economics, 4. [11] Arce, F., Bengui, J, and Bianchi, J. 2019. “A Macroprudential Theory of Foreign Reserve Accumulation.” Federal Reserve Bank of Minneapolis Working Paper, No. 761. [12] Barro, Robert J. 2006. “Rare disasters and asset markets in the twentieth century.” The Quar- ter Journal of Economics, 121 (3): 823–866. 189 [13] Bayoumi, Tamim, Gagnon, Joseph and Saborowski, Christian. 2015. “Official financial flows, capital mobility, and global imbalances.” Journal of International Money and Finance, 52: 146–174. [14] Benigno, G., Chen, H., Otrok, C., Rebucci, A., and Young, E. R. 2016. “Optimal Capital Controls and Real Exchange Rate Policies: A Pecuniary Externality Perspective,” Journal of Monetary Economics, 84: 147–165. [15] Benigno, G. and Fornaro, L. 2014. “Reserve accumulation, growth and financial crises.” Manuscript, London School of Economics and Political Science. [16] B´ en´ etrix, A., Philip R. Lane, and Jay C. Shambaugh. 2015. “International currency exposures, valuation effects and the global financial crisis.” Journal of International Economics, 96: S98–S109. [17] Bianchi, J. 2011. “Overborrowing and Systemic Externalities in the Business Cycle.” Ameri- can Economic Review, 101 (7): 3400–3426. [18] Bianchi, J., Hatchondo, J. C. and Martinez, L. 2018. “International Reserves and Rollover Risk.” American Economic Review, 108 (9): 2629-2670. [19] Bianchi, J. and Enrique G. Mendoza, 2018.“Optimal Time-Consistent Macroprudential Pol- icy,” Journal of Political Economy, 126 (2): 588-634. [20] Bocola, L. and Lorenzoni, G. 2019. “Financial Crisis, Dollarization, and Lending of Last Resort in Open Economie.” Manuscript, Northwestern University. [21] Broner, F., T. Didier, A. Erce, and S. L. Schmukler (2013) “Gross capital flows : Dynamics and crises,” Journal of Monetary Economics, 60: 113-133. [22] Bruno, V . and Shin, H. S. 2015. “Cross-Border Banking and Global Liquidity.” Review of Economic Studies, 82 (2): 535–564. [23] Bussiere, M., Cheng, G., Chinn, M., and Lisack, N. 2015. “For a Few Dollars More: Reserves and Growth in Time of Crises.” Journal of International Money and Finance, 52: 127–145. [24] Caballero, R. J. and Krishnamurthy, A. 2003. “Excessive Dollar Debt: Financial Develop- ment and Underinsurance.” Journal of Finance, 58 (2): 867–894. [25] Calvo, G., and C. Reinhart, 2002. “Fear of Floating.” Quarterly Journal of Economics, 117 (2): 379–408. [26] Cavallino, P. 2019. “Capital Flows and Foreign Exchange Intervention.” American Economic Journal: Macroeconomics, 11 (2): 127–170. [27] Chang, R. 2018. “Foreign Exchange Intervention Redux.” NBER Working Paper, No 24463. [28] Choi, W. J. and Taylor, A. M. 2017. “Precaution versus Mercantilism: Reserve Accumulation, Capital Controls, and real exchange rate.” NBER Working Paper, No. 23341. 190 [29] Davis, J. Scott., Ippei Fujiwara, Kevin X.D. Huang and Jiao Wang. 2019. “Foreign Exchange Reserves as a Tool for Capital Account Management.”, Globalization Institute Working Pa- per, No. 351. [30] Dominguez Kathryn M.E., Y . Hashimoto, and T. Ito. 2012. “International Reserves and the Global Financial Crisis.” Journal of International Economics, 88 (2): 388–406. [31] Dominguez, Kathryn M.E. 2010. “International Reserves and Underdeveloped Capital Mar- kets.” In NBER International Seminar on Macroeconomics 2009, edited by Lucrezia Reichlin and Kenneth West, 193–221. University of Chicago Press. [32] Dominguez, Kathryn M.E. 2019. “Emerging Market Exchange Rate Policies: Stabilizing or Manipulation?” Manuscript, University of Michigan. [33] Du, Wenxin and Pflueger, Carolin E and Schreger, Jesse. 2016. ”Sovereign Debt Portfolios, Bond Risks, and the Credibility of Monetary Policy.” NBER Working Paper, No. 22592. [34] Durdu, C. B., Mendoza, E. G., and Terrones, M. E. 2009. “Precautionary Demand for Foreign Assets in Sudden Stop Economies: An Assessment of the New Mercantilism.” Journal of Development Economics, 89 (2): 194–209. [35] Fanelli, S. and Straub, L. 2019. “A Theory of Foreign Exchange Interventions.” Manuscript, Massachusetts Institute of Technology. [36] Farhi, E. and Werning, I. 2014. “Dilemma not Trilemma? Capital Controls and Exchange Rates with V olatile Capital Flows.” IMF Economic Review, 62: 569–605. [37] Fernandez, Andres, Michael Klein, Alessandro Rebucci, Martin Schindler, and Martin Uribe, 2016. “Capital Control Measures: A New Dataset,” IMF Economic Review, 64: 548–574. [38] Forbes, Kristin and Warnock, Francis, 2012. ”Capital flow waves: Surges, stops, flight, and retrenchment”, Journal of International Economics, 88 (2): 235–251. [39] Gabaix, X. and Maggiori, M. 2015. “International Liquidity and Exchange Rate Dynamics.” Quarterly Journal of Economics, 130 (3): 1369–1420. [40] Gottardi, P., Maurin, V ., Monnet, C., 2017. “A theory of repurchase agreements, collateral re-use, and repo intermediation.” EUI working paper ECO. [41] Tarek A Hassan, Thomas M Mertens, and Tony Zhang. 2019. “Currency Manipulation,” Manuscript, Boston University. [42] Hernandez, J. 2017. “How International Reserves reduce the Probability of Debt Crises.” Manuscript, University of Pennsylvania. [43] Horn, Sebastian, Reinhart, Carmen and Trebesch, Christoph. 2019. “China’s Overseas Lend- ing.” NBER Working Paper, No. 26050. [44] Hur, S. and Kondo I. O. 2016. “A Theory of Rollover Risk, Sudden Stops, and Foreign Reserves.” Journal of International Economics, 103: 44–63. 191 [45] Jeanne, O. 2015. “The Macroprudential Role of International Reserves.” American Economic Review 106 (5): 570–573. [46] Jeanne, O. and Korinek, A. 2010. “Managing Credit Booms and Busts: A Pigouvian Taxation Approach.” NBER Working Paper, No. 16377. [47] Jeanne, O. and Ranciere, R. 2011. “The Optimal Level of International Reserves for Emerg- ing Market Countries: a New Formula and Some Applications.” Economic Journal, 121 (555): 905–930. [48] Jeanne, O. and D. Sandri. 2018. ”Global Financial cycle and liquidity manage- ment”.Manuscript, Johns Hopkins University. [49] Korinek, A. 2018. “Regulating Capital Flows to Emerging Markets: An Externality View.” Journal of International Economics, 111: 61–80. [50] Korinek, A. and Sandri, D. 2016. “Capital Controls or Macroprudential Regulation?” Journal of International Economics, 99 (S1): 27–42. [51] Korinek, A. and Serven, L. 2016. “Undervaluation through Foreign Reserve Accumulation: Static Losses, Dynamic Gains.” Journal of International Money and Finance, 64: 104–136. [52] Lane, Philip and Milesi-Ferretti, Gian Maria. 2007. “The external wealth of nations mark II: Revised and extended estimates of foreign assets and liabilities, 1970-2004.” Journal of International Economics, 73 (2): 223–250. [53] Levy-Yeyati, E. 2008. “The Cost of Reserves.” Economic Letters, 100 (1): 39–42. [54] Levy-Yeyati, E., F. Sturzenegger, and P. A. Gluzmann. 2013. “Fear of appreciation”, Journal of Development Economics, 101 (1): 233–247. [55] Matsumoto, H. 2019, “Foreign Reserve Accumulation, Foreign Direct Investment, and Eco- nomic Growth”, IMES Discussion paper. [56] Mendoza, E. G. and Bianchi, J. 2018. “Optimal Time-Consistent Macroprudential Policy.” Journal of Political Economy, 126 (2): 588–634. [57] Obstfeld, M., Shambaugh J. C., and Taylor, A. M. 2010. “Financial stability, the trilemma, and international reserves.” American Economic Journal: Macroeconomics, 2 (2): 57–94. [58] Ostry, Jonathan, Ghosh, Atish, Habermeier, Karl, Laeven, Luc, Chamon, Marcos, Qureshi, Mahvash and Kokenyne, Annamaria. 2011. ”Managing Capital Inflows; What Tools to Use?”, IMF Staff Discussion Notes, No 11/06. [59] Ottonello, P. and Perez, D. J. 2019. “The Currency Composition of Sovereign Debt.” Ameri- can Economic Journal: Macroeconomics, 11 (3): 174–208. [60] Parlatore, Cecilia. 2019. “Collateralizing liquidity.” Journal of Financial Economics, 131 (2): 299–322. 192 [61] Prasad, Eswar S. 2011. “Role Reversal in Global Finance.” NBER Working Paper, No. 17497. [62] Rodrik, D. 2006. “The Social Cost of Foreign Exchange Reserves.” International Economic Journal 20 (3): 253–266. [63] Shousha, S. F. 2017. “International Reserves, Credit Constraints, and Systemic Sudden Stops.” Board of Governors of the Federal Reserve System International Finance Discussion Papers, No. 1205. [64] Wang, M. 2019. “Foreign Direct Investment and Foreign Reserve Accumulation.”, Manuscript, Columbia University. [65] Winant P., Ostry, J., Ghosh, A. and Basu, S. 2017. “Managing Capital Outflows: The Role of Foreign Exchange Intervention.” Manuscript, International Monetary Fund. 193 Chapter 4 Appendices A Appendix to Chapter 1 A.1 Local currency debt and equity liability data • Argentina: Argentine peso denominated bond portfolio investment is from local currency sovereign debt data from Arslanalp and Tsuda (2014). Some local currency debts are ob- served in QEDS (Quarterly External Debt Statistics) provided by World Bank 1 , but in the IIP of Argentina from IMF, external debts of deposit-taking corporations are negligible. There- fore, I assume that the small amount of local currency debts belong to FDI liabilities in the form of debt instrument. No information about the currency denomination of equity port- folio investments in IIP. I simply assume that all equities in the category are local currency denominated. • Brazil: Central bank of Brazil (BCB) provides IIP data that divides portfolio investments into securities issued in onshore (domestic) and offshore (foreign) markets. Officials at BCB confirmed that all equities and bonds issued in Brazil are almost all Brazil real denominated by regulation. Other local currency debts in the category of other investments in the classi- fication of IIP are identified from the data provided by officials at BCB (directly provided conditional on not sharing with others). 1 The QEDS is co-managed with BIS and IMF 194 • Bulgaria: Complete information of local currency debts and bonds of Bulgaria is provided from the central bank of Bulgaria (BNB, directly provided conditional on not sharing with others). All equity portfolio investments in IIP are assumed to be local currency denomi- nated. • Chile: Central bank of Chile (Banco Central de Chile) provides the data on debt securities issued in Chile, but held by nonresidents. Currently, I have no information on the currency denomination of the bonds, but the government bonds in the data well match Arslanalp and Tsuda (2014). In addition, deposits in Chile, held by nonresidents, are available from the ex- ternal debt data provided by the bank. The summation of the debt securities and the deposits are little short of the total Chile peso denominated external debts, which is also provided by Banco Cental de Chile; the remaining debts are presumably the peso denominated FDI debts. All equity portfolio investments in IIP are assumed to be local currency denominated. • Colombia: QEDS provides total Colombian peso denominated external debts, but the re- ported numbers in the data set are much smaller than Columbian peso denominated sovereign bonds in Arslanalp and Tsuda (2014). Like several other EMEs, it is very probable that the numbers in QEDS are under-reported. Hence, I used the data in Arslanalp and Tsuda (2014), and assumed that there is no other Colombian peso denominated external debts as it turns out external debts in the form of deposits are negligible in Colombia. All equity portfolio investments in IIP are assumed to be local currency denominated. • Czech Republic: Czech koruna denominated external debt unusually increased since 2014. The increasing pattern corresponds to a sudden increase in nonresidents koruna deposits in Czech Republic. 2 Hence, I reasonably assumed that all external debt in the form of deposits are Czech koruna denominated external debts; the external debts of different instruments are from the cental bank of Czech Republic (Czech National Bank, CZB). Czech koruna denominated sovereign bonds are computed as follows. The Ministry of Finance of the 2 It is 20-30% of GDP and definitely unusual. About the driving force behin such unusual phenomenon, currently I have no clue. 195 Czech Republic provides government debt data that classify the governments into domestic currency versus foreign currency, and domestic creditors and foreign creditors. I assumed that all foreign currency denominated sovereign external debts belong to foreign creditors. Then the Czech koruna denominated sovereign external debts are total external government debts credited by foreign investors minus the total foreign currency denominated government debts. The summation of the korona denominated deposits held by nonresidents and the korona denominated sovereign bonds reasonably matches the total local currency external debts of Czech Republic in QEDS. All equity portfolio investments in IIP are assumed to be local currency denominated. • Hungary: Currency compositions section in IIP is filled out in Hungarian data. Hence, I can easily identify local and foreign currency debts in different categories. All the equity portfolio investments in IIP are assumed to be local currency denominated. • India: Ministry of Finance in India annually publishes a report of external debts including ratios of India rupee denominated debts in both of total and sovereign debts. Since FDI debts in India are almost negligible, I assume that all rupee currency debts in the report are either of investments in rupee denominated bonds or rupee denominated deposits. The same report provides the information on Indian rupee deposits held by nonresidents; NRI deposits except for FCNR. The total rupee denominated external debts are substantially more than the summations of the rupee denominated sovereign bonds and rupee deposits held by non-residents. The remaining rupee denominated external debts are foreign portfolio investments, through FII account, in rupee corporate bonds or special bonds. 3 All equity portfolio investments in IIP are assumed to be local currency denominated. • Indonesia: Central bank of Indonesia (Bank Indonesia, BI) annually publishes the report of external debts of Indonesia, which documents total Indonesian rupiah denominated external 3 Other parts of external debts in Indian rupee are Rupee Debt: outstanding state credits (both defense and civilian) extended to India by the erstwhile Union of Soviet Socialist Republic (USSR). Hence, Rupee Debt is a legacy from the past and currently the remaining Rupee Debt is negligible. 196 debts and sovereign external debts. The summation of the rupiah sovereign external debts and deposits held by nonresidents almost exactly matches the total rupiah external debt. All equity portfolio investments in IIP are assumed to be local currency denominated. • Korea: Central bank of Korea (Bank of Korea, BOK) provides annual data on currency compositions of external assets and liabilities for all the categories in IIP. • Malaysia: Central bank of Malaysia (Bank Negara Malaysia, BNM) provided me with monthly data on external debts in Malaysia showing Malaysia ringgit denominated debt securities held by nonresidents; directly contact officials at BNM. A series of reports from BNM documents the same information at semiannual frequency and the report shows de- posits in Malaysia, held by nonresidents. I assumed all the deposits are denominated in ringgit and the summation of the deposits and debt securities reasonably matches the total ringgit external debts of Malaysia, documented in the periodical reports of BNM. 4 All equity portfolio investments in IIP are assumed to be local currency denominated. • Mexico: Gross External Debt Position (GEDP) data from central bank of Mexico (Banco de Mexico, Banxico) shows all Mexico peso denominated debt securities held by nonresidents. QEDS also provides Mexico peso denominated external debts; the documented numbers are smaller than GEDP data. As a Baxico senior official confirmed to me that GEDP data are correct, I ignore QEDS data and assumed that all Mexico peso-denominated external debts of Mexico are the peso debt securities in GEDP data. The official also confirmed that all equity portfolio investments in IIP are Mexico peso denominated. • Peru: Peru sol denominated sovereign debt data is from Arslanalp and Tsuda (2014). I cannot find any information of nonresidents deposit in Peru. Becaus of the deficient information, I use the number reported in Du and Schreger (2016): the paper documents that as of 2012, 7% of external debts of corporate sectors in Peru is sol denominated, and using the reported 4 For 2012 and 2013, the computed ringgit external debts are little more than the separately reported total ringit external debts. 197 ratio, I assume that 7% of external debts other than Peruvian government external debts are sol denominated. All equity portfolio investments in IIP are assumed to be local currency denominated. • Philippines: Central bank of the Philippines (Bangko Sentral ng Pilipinas, BSP) provided me with the data on Philippines peso denominated external debts in both government and private sector; directly contacted BSP. Actually, it turns out that peso denominated external debt in private sector is almost zero. The contacted official also confirmed that all the equity portfolio investments in IIP are Philippines peso denominated. • Poland: Central bank of Poland (National Bank of Poland, NBP) provided me with currency compositions of external assets for all the categories in IIP; directly contacted BNP. All equity portfolio investments in IIP are assumed to be local currency denominated. • Romania: Romania leu denominated sovereign external debts are from Arslanalp and Tsuda (2014). 5 Deposits held by nonresidents are assumed to be leu denominated. The total leu denomiated external debts are computed as the summation of the leu denomiated government external debts and the deposits held by nonresidents. • Russia: Central bank of Russia (Bank of Russia, CBR) provided a few different data sets showing currency denominations in external assets and liabilities. 6 Only missing informa- tion in the data is the amount of Russia ruble deposits, held by nonresidents. I assumed that ruble denominated external debts of the banking sector in Russia are the ruble denominated deposits as it is in many of the sample EMEs. • South Africa: South Africa central bank (South African Reserve Bank) publishes a report of external debts every quarter, and the reports document the amount of South Africa rand denominated debt securities held by nonresidents and total rand denominated external debts. 5 Leu denominated government external debts are higher than the total leu denominated external debts reported in QEDS. 6 One can download the same data from the CBR website. 198 Then I added deposits in South Africa, which I can identify from the same periodical. 7 All equity portfolio investments in IIP are assumed to be local currency denominated, and it is confirmed by an official at South Africa Reserve Bank (direct contact). • Thailand: Central bank of Thailand (Bank of Thailand, BOT) directly provided me with the data of Thailand bhat denominated government external debts, different forms of external debts in bhat and the total Thailand external debt in bhat. Other than the government bonds in bhat, BOT also issues its own bonds in bhat, some of which are purchased by foreign investors. The information is available at BOT website. All equity portfolio investments in IIP are assumed to be local currency denominated, and it is confirmed by an official at BOT (direct contact). • Turkey: Turkey lira denominated sovereign external debts are from Arslanalp and Tsuda (2014). The data of lira denominated deposits held by nonresidents is provided by the cen- tral bank of Turkey (T¨ urkiye Cumhuriyet Merkez Bankasi, TCMB). All equity portfolio investments in IIP are assumed to be local currency denominated. A.1.1 Sector Level Foreign Currency Assets and Liabilities • External assets and liabilities of each sector in foreign currency: IIP data from IMF break down external assets and liabilities into different sectors (financial corporate sector, govern- ment, central bank and others 8 ) and different types (equity portfolio, bond portfolio, and other debt instruments). From the external liabilities in different categories, I deduct corre- sponding local currency debt liabilities. 7 One can download same data from South African Reserve Bank online download system. The required codes for rand denominated external debt securities and deposits held by non-residents are KBP5512J and KBP5579J respec- tively. 8 The category of others includes both household sector and nonfinancial corporate sector. Since it is unlikely that households hold large amounts of external assets and liabilities without intermediations of banks, which is reported in banking sector liabilities and assets in IIP data, I assumed all the assets and liabilities in the category of “others” belong to nonfinancial corporate sector. 199 • Foreign currency deposits: Data set by Dr. Dalgic includes foreign currency deposits in the sample EMEs, all the EMEs in the sample except for Brazil. The data only shows the ratio of foreign currency deposits to the total deposits. Therefore, I used the deposit to GDP ratios of CEIC database. • Foreign currency loan: Same data set provided by Dr. Dalgic includes foreign currency loans between domestic creditors and domestic borrowers for 12 EMEs; Argentina, Bulgaria, Chile, Columbia, Czech Republic, Hungary, Indonesia, Mexico, Poland, Romania, Russia, and Turkey. For the last of 7 EMEs among the 19 EMEs (except for Brazil in the 20 EMEs), I simply assumed that the foreign currency loan is the same as the foreign currency deposit; for the 12 EME that I have information, the amount of foreign currency deposits are almost equivalent to foreign currency loans. As a result, I know the ratios of the foreign currency loans to the total credit to private nonfinancial sectors. The data of total credit from domestic financial sector to nonfinancial sectors is available from World Bank. Then I can easily compute the foreign currency loan to GDP ratios. • Foreign currency loan to households: The data set of Dr. Dalgic includes the foreign cur- rency loan to household credit ratios for the following EMEs; Argentina, Bulgaria, Colom- bia, Czech Republic, Hungary, Indonesia, Mexico, Poland, Romania, and Russia. For the remaining EMEs, I assumed that all the domestic foreign currency loans are made to non- financial corporates since, among the EMEs that I have information, foreign currency loans to households are negligible except for Poland and Romania. Multiplication of the foreign currency loan to household credit and household debt to GDP ratios give me the household foreign currency loan to GDP ratios. I used household debt data from the CEIC database. A.1.2 Additional Tables 200 Table A.1: Sources for the Local Currency Equities and Debts Country Sources Argentina - No available national source - Arslanalp and Tsuda (2014) Brazil - Direct contact (Banco Central do Brasil) - International Investment Position from Banco Central do Brasil Bulgaria - Direct contact (Bulgarian National Bank) Chile - Central Bank of Chile Statistical Database Columbia - No available national source - Arslanalp and Tsuda (2014) Czech Rep. - Czech National Bank Statistics - Ministry of Finance of the Czech Republic Statistics Hungary - No available national source - Currency composition table in the IMP IIP database India - The Ministry of Finance of the India Quarterly External Debt Report Indonesia - Bank Indonesia External Dept Report Korea - Bank of Korea Statistics Portal Ecos Malaysia - Direct contact (Bank Negara Malaysia) - Bank Negara Malaysia Quarterly Bulletin Mexico - Banco De Mexico Economic Information System Peru - No available national source - Arslanalp and Tsuda (2014) Philippines - Direct contact (Bangko Sentral ng Pilipinas) - Arslanalp and Tsuda (2014) Poland - Direct contact (Narodowy Bank Polski) - Currency composition table in the IMP IIP database Romania - No available national source - Arslanalp and Tsuda (2014) Russia - Bank of Russia Statistics South Africa - South African Reserve Bank Quarterly Bulletin Thailand - Direct contact (Bank of Thailand) - Bank of Thailand Statistics Turkey - Central Bank of the Republic of Turkey Statistical Data (EVDS) 201 Table A.2: External Liabilities of EMEs 2004 2018 LCE LCD FCL LCE LCD FCL Argentina 0.01 (0.02) 0.01 (0.02) 0.58 (0.97) 0.02 (0.05) 0.00 (0.00) [0.00] 0.43 (0.95) Brazil 0.05 (0.12) 0.00 (0.00) 0.37 (0.88) 0.12 (0.30) 0.06 (0.17) [0.06] 0.20 (0.53) Bulgaria 0.02 (0.04) 0.01 (0.02) 0.50 (0.94) 0.01 (0.02) 0.02 (0.04) [0.00] 0.42 (0.94) Chile 0.05 (0.11) 0.00 (0.00) 0.40 (0.89) 0.09 (0.16) 0.04 (0.07) [0.03] 0.40 (0.77) Colombia 0.02 (0.06) 0.01 (0.03) 0.31 (0.91) 0.02 (0.04) 0.07 (0.15) [0.07] 0.39 (0.81) Czech – – – 0.03 (0.05) 0.38 (0.55) [0.09] 0.28 (0.40) Hungary 0.11 (0.14) 0.30 (0.38) 0.39 (0.49) 0.11 (0.16) 0.17 (0.25) [0.14] 0.39 (0.59) India 0.09 (0.29) 0.04 (0.13) 0.18 (0.58) 0.05 (0.21) 0.07 (0.29) [0.03] 0.12 (0.50) Indonesia 0.08 (0.14) 0.06 (0.11) 0.42 (0.75) 0.10 (0.23) 0.07 (0.16) [0.06] 0.26 (0.61) Korea 0.17 (0.43) 0.02 (0.05) 0.21 (0.53) 0.25 (0.48) 0.08 (0.15) [0.06] 0.20 (0.38) Malaysia 0.33 (0.46) 0.00 (0.00) 0.39 (0.54) 0.18 (0.26) 0.18 (0.26) [0.13] 0.34 (0.49) Mexico 0.11 (0.33) 0.01 (0.03) 0.21 (0.64) 0.11 (0.23) 0.09 (0.19) [0.09] 0.29 (0.58) Peru 0.06 (0.11) 0.00 (0.00) 0.47 (0.89) 0.09 (0.20) 0.07 (0.15) [0.05] 0.29 (0.65) Philippines 0.05 (0.07) 0.00 (0.00) 0.62 (0.93) 0.18 (0.41) 0.01 (0.02) [0.01] 0.24 (0.57) Poland 0.05 (0.10) 0.20 (0.42) 0.23 (0.48) 0.07 (0.11) 0.17 (0.29) [0.13] 0.34 (0.60) Romania 0.01 (0.03) 0.05 (0.15) 0.28 (0.82) 0.01 (0.04) 0.04 (0.13) [0.03] 0.28 (0.83) Russia 0.21 (0.45) 0.00 (0.00) 0.26 (0.55) 0.10 (0.33) 0.05 (0.17) [0.04] 0.15 (0.50) South Africa 0.21 (0.55) 0.00 (0.00) 0.17 (0.45) 0.45 (0.52) 0.16 (0.19) [0.14] 0.26 (0.29) Thailand 0.16 (0.35) 0.02 (0.04) 0.28 (0.61) 0.22 (0.42) 0.10 (0.20) [0.06] 0.19 (0.37) Turkey 0.04 (0.09) 0.00 (0.00) 0.39 (0.91) 0.04 (0.07) 0.05 (0.07) [0.03] 0.56 (0.86) Average 0.10 (0.20) 0.04 (0.07) 0.35 (0.72) 0.11 (0.21) 0.09 (0.18) [0.06] 0.31 (0.61) Note: 1) All the numbers are the Liabilities to GDP ratios. 2) LCD: Local Currency Debts, LCE: Local Currency Equities. and FCL: Foreign Currency Liabilities 3) The numbers in ( ) are the ratios of each type of external liabilities to total external liabilities excluding FDIs. 4) The numbers in [ ] are local currency bond portfolio investments. 5) The data on Czech Republic in 2004 are not available. 6) Most of the FC (Foreign Currency) liabilities are FC debts. Table A.3: Correlations with Local Currency External Liabilities Stock Mkt. Cap. Public Debt Mkt Cap. Avg. Inflation 2) Government Effectiveness 3) GDP per capita 4) Trade Openness Financial Openness 5) LC Equity 0.81 (0.69) - -0.19 (0.03) 0.32 (0.10) 0.05 (0.00) 0.24 (0.05) -0.03 (0.00) LC Bond - 0.61 (0.37) -0.36 (0.02) 0.44 (0.17) 0.30 (0.09) 0.47 (0.22) 0.22 (0.03) Note: 1) The numbers in the parentheses are the R-squared of the univariate regression. 2) Average inflation is the average of the annual inflation for the last 10 years. 3) Government effectiveness index is one of the six categories in the World Bank Governance Indices. 4) GDP per capital is measure by US dollar in 2017 (World Economy Outlook 2020 October, IMF) 5) Financial openness measures are the indices of equity and bond market openness indices in Fernandez et al. (2016). 202 B Appendix to Chapter 2 B.1 Other Data I denote sources of other data used in the cross-country panel regressions • International Investment Position: International Monetary Fund • Exchange Rates and Stock Indices: Bloomberg • Other Controls – Trade Openness: World Bank – Financial Openness: Chinn and Ito website (http://web.pdx.edu/˜ito/Chinn-Ito website.htm) – Oil Price: IMF Commodity Data Portal Crude Oil Price Index – Commodity price: IMF Commodity Data Portal Non-Fuel Commodity Price Index – Short term interest rates: 3 Month Treasury Bill Rates from CEIC database (Brazil, Columbia, Czech Republic, Hungary, India, Mexico, Philippines, Russia, South Africa, and Thailand), 3 Month Interbank Interest Rates from CEID database (Indonesia, Peru, Poland, Romania, and Turkey), and 3 Month Interbank Interest Rates from IMF IFS (Argentina, Bulgaria and Chile) – Real Effective Exchange Rates: BIS Effective Exchange Rate Indices – Inflation: IMF IFS 9 – Industrial Production: CEIC database – M2 Monetary Aggregate: CEIC database 9 Monthly inflation in Argentina since 2015 is not available anywhere. Hence I extrapolated using nominal and real effective exchange rates of BIS effective exchange rate indices. 203 B.2 Portfolio of Global Investors Global investors are the international financial intermediaries who purchase local currency denom- inated equities and bonds in the small open economy. Like other component in the model, I model the global investors in a simple way, but also aim at capturing key features in the reality. Since this paper studies impacts of risk appetite shock to global investors, the global investors in the model need to be risk-averse. While there are different ways, I model the global investors as international financial intermediaries under VaR constraint, following Miranda-Agrippino and Rey (2019). Global investor in the model at time t has her own capital W G t and can raise outside financing in foreign currency in the form of one period debt to invest in different assets indexed by j2 f1;2;::::;Ng. Let p t and R t+1 are the vectors of the global investor portfolio and the excess return of the risky assets over the safe asset respectively. The optimization problem of the global investor is formulated as follows. max X t E t h p 0 t R t+1 1 N R f t+1 i sub ject to VaR t W G t where VaR t =a[Var[p 0 t R t+1 ]] 1 2 . The solution to the problem is p t = W G t al t [Var(R t+1 )] 1 E t R t+1 1 N R f t+1 (B.1) wherel t = h E t R t+1 1 N R f t+1 0 [Var(R t+1 )] 1 E t R t+1 1 N R f t+1 i 1=210 andVar(R t+1 ) de- notes the variance. Hence, the solution is identical to the optimal portfolio of a mean-variance investor. Also, notice that any shock to the capital of the global investor W G t or expected volatility of the world risky assetsVar(R t+1 ) leads to changes in the risky asset holdings of the global investor. To study the optimal portfolio of the global investor more specifically, let’s formulate the prob- lem as a consideration of an investment in a “marginal” asset: the investor had already formed a market portfolio composed of N-1 different assets, hence all the available risky assets except for i. 10 Hencel t is the sharpe ratio. 204 And further, the marginal asset i follows R i t v N R t i ;s 2 i +q i s 2 m i and Cov R i t ;R m i t =q i s 2 m i . Hence,q i is the “market beta” for asset i: The share of asset i, denoted by x i t , is given by x i t = R i R f = R m i R f q i s 2 i =s 2 m i + R i R f =(R m i R f )q i R i R f =(R m i R f )+ 1 (B.2) It is easy to show that x i decreases inq i if p i R i <(1 p i )R m i . We have two different assets in the model small open economy, the capital and the government bond. However, I assume that the share of the two assets in the total portfolio is small enough, so that I can take the result in equation (55) to both the capital and the bond. Not let W G t = W G e v t . Hence, I interpret risk-on/off shocks as shocks to the capital of the global investors 11 . V t corresponds to VIX and therefore it is a measure of the risk appetite of the investors. In addition, I assume that V t follows a mean-reverting process similarly with VIX. Thus v t =rv t1 +n t (B.3) wheren t s N 0;s 2 n andr v 2(0;1). Henceforth, I calln t > 0 “risk-off” shock andn t < 0 “risk-on” shock. I need to simplify the specification in equation (55) to make it suitable for quantitative analysis. I can reasonably assume R i ' R m i , i.e., R i R f R m i R f ' 1. Then taking a first-order approximation around R i ' R m i gives me the approximation of p i t ,e p i t e x i t = 1 s 2 i =s 2 m i + 1 2q i + 1 1q i s 2 i =s 2 m i + 1 2q i ! R i R m i R m i R f ! ' 1 s 2 i =s 2 m i + 1 2q i + 1 s m i 1 1q i s 2 i =s 2 m i + 1 2q i ! R i R m i (B.4) 11 This interpretation is in line with Miranda-Agrippino and Rey (2019) and Bruno and Shin (2015a). 205 where s m i is a constant close to R m i R f . s m i denotes spread of the global portfolio over the return to the safe asset. This another approximation is to reduce the number of the parameters I need to estimate for the calibration. To make it even more tractable, I assume that the parameters regarding the risk propertiess 2 i , s 2 m and q i are invariant in short run. We can think of investors who update their belief sporadi- cally. 12 Then I finally get e x i t 'c i 0 +c i 1 R i R m i (B.5) Then, let’s denote the money invested in the asset i by p i t . It is p i t = W G c i 0 e v t 1+ c i 0 c i 0 R i R m i Once I replace W G c i 0 with 1 G i and fully express the terms, the demand from the global investors for the equity and government bonds in the small open economy are given by 13 p k t = Q t k f t e 1 t = 1 G k e v t " 1+ c k 0 c k 0 E t e t e t+1 R k t+1 R m t+1 (v t ) # p b t = q t b f t e 1 t = 1 G b e v t " 1+ c b 0 c b 0 E t e t e t+1 R b t+1 R m t+1 (v t ) # B.3 Pricing to Markets and Exchange Rate Channel In this subsection, I introduce an extension of the simple model before I build a more general model to be used for quantitative exercises. The purpose of the extension is to illustrate how resilient nonfinancial corporates in EMEs can be to local currency depreciation so that the seemingly large amounts of net foreign currency debts of nonfinancial corporates do not show a significance in the cross-country regressions. For tractability and simplicity, in this subsection, I treat domestic banks 12 I implicitly assume that the investors update the belief of expected return more frequently. We can think the information to predict the expected return is more available or cheaper. 13 The seminal paper Gabaix and Maggiori (2015), and following papers derive similar forms from agency frictions between global financial intermediaries and investors. 206 as a conglomeration of financial and non-financial corporates. This is a way to illustrate desired mechanism, while keeping consistency in modeling techniques. While maintaining simplicity even in the extended model, I give one change to the simple model. I adopt monopolistic competition to the model. Following the standard in the literature, final goods are produced from a variety of differentiated goods y i;t ;i2[0;1] under perfect compe- tition according to CES technology as below. Y t = Z 1 0 y h1 h i;t di h h1 whereh > 1. Each differentiated intermediate good is by the standard Cobb-Douglas technology. y i;t = A t (k i;t ) a (l i;t ) 1a where subcript i denotes inputs used by producer i. Following the standard in the literature, I assume that the intermediate goods producers are under monopolistic competition and thus face a downward sloping demand curve as follows. y i;t = p i;t P t h Y t where p i;t is the nominal price of goods i and P t is the aggregate price index as follows. P t = Z 1 0 p 1h i;t di 1 1h Notice I do not introduce nominal rigidity; in each period, the intermediate goods producers can set their prices optimally. The producers can separately set their prices in foreign markets, as argued in Local Currency Pricing (LCP) hypothesis. 14 However, while the producers can optimally change their prices in local markets, producers, who export to foreign markets, take foreign market prices as given. This is a way to keep the model simple, while capturing observed empirical 14 For the different pricing in different markets of exporters, see the seminal paper Betts and Devereux (2000) 207 features. As it is well known, exchange rate pass-through into export prices is usually low in reality. 15 Also, for many exporting goods, the exporting prices are determined under strategic considerations of the exporters, from which I abstract here or the prices are determined in large international markets, like some commodities in the reality. Then the two different prices in domestic and foreign markets are determined as follows. p i;t = h h 1 mc t and p i;t = p t (e t p t > mc t ) Please notice that the price in the foreign markets is assumed to be higher than the marginal costs, and therefore the local currency depreciation, higher exchange rates will gift higher mark-ups to the exporters if the marginal costs in local currency are fixed. Since the price is higher than the marginal costs, the monopolistic producer can have some profits as follows. p i;t = 1 h Y d t |{z} Domestic Pro f it +Y t (p t ) h (e t mc t ) | {z } Export Pro f it where Y d t = c d t + I t + G and e t is the real exchange rate. mc t = 1 A t (az t ) a ((1a)w t ) 1a and z t and w t are the real rental cost of capital and the real wage in terms of “domestic price of the final goods.” In other words, the marginal cost of the producers is denominated in local currency. Now I show the “economic” profits of the corporates whose revenues are partially denominated in foreign currency move opposite to risk-appetite of the global investors. That is, risk-off (on) shocks increase (decrease) the profits of the corporates. I highlight this in the following lemma. Lemma 4 Risk-off (on) shocks increase (decrease) the corporate profitsp t . That is, dp t dn t > 0 Therefore, the profits of corporates increase when risk-appetite of global investors unexpectedly falls, despite the negative impacts on the capital market. This is a particular case because I abstract from some features in the reality, like nominal rigidity, which generates aggregate demand exter- nality. The implication from the lemma should be understood as such that profits of the exporting 15 See the excellent survey Burstein and Gopinath (2014) 208 corporates are impacted less than others or the profits are relatively stable from the risk-appetite shocks. Of course, in reality, profits of export oriented firms in EMEs can even increase although the economy falls into a recession. How are the corporates in the model benefiting from the risk-off shocks? Recall the costs of the corporates are denominated in local currency in the sense that real wages and rental costs are measured by marginal products in domestic markets. On the contrary, parts of the revenues are denominated in foreign currency. Therefore, local currency depreciation raise markups for the corporates. While local currency depreciation benefits the corporates on the profit side, the depreciation should raise the real debt burden of foreign currency debts if the corporates have foreign currency debts. Thus, the positive effects and negative effects offset each other, and which effect is dominat- ing depends on amounts of the foreign currency debts and the magnitude of the positive impacts on the profit, which again depend on different conditions such as share of exports in the outputs of the corporates, i.e., trade openness of the economy for the country representative firm. To see it more clearly, let’s look at the impacts of local currency depreciation on the net worth. 16 Recall that the exchange rate channel works through the impact of the exchange rate on the net worth of the domestic banks, which include the nonfinancial corporates here. ¶N t ¶e t de t dn t = 1 ¶Q t ¶N t k d t1 1 " ¶p t ¶e t + ¶Q t ¶ p k t e t ¶ p k t e t ¶e t R t d t1 # (B.6) In equation (B.6), the conventional balance sheet effects are captured by R t d t1 ; negative effects of local currency depreciation and the impacts on the profits are captured by ¶p t ¶e t ; positive effects of the depreciation. Because of the different effects offseting each other, the sign of ¶N t ¶e t de t dn t is inconclusive. In my regression, I have no proper measure of the impacts of exchange rate movements on the corporate profits, ¶p t ¶e t . If ¶p t ¶e t is positively correlated with the foreign currency debt d t1 as it is 16¶N t ¶e t ¶p t ¶e t = Y t (p t ) 1h P t and ¶Q t ¶N t = jf t q (jK t1 1) 2 +4j(N t f t +p k t e t) 209 likely in reality, then consistently with the empirical results, more net foreign currency debts in nonfinancial corporate sectors do not necessarily mean higher fragility to local currency deprecia- tion. To give it another way, let’s imagine that one looks at different EMEs with different levels of foreign currency debts in non-financial corporate sectors, and estimates correlations between the foreign currency debts and fragility measures such as changes in stock indices. If ¶p t ¶e t is not properly controlled due to some unobservable features like different pricing in exports, then the correlation corr ¶N t ¶e t de t dn t ;d t1 is hard to be meaningfully high. Then the following question is whether the positive impacts on the profits are positively cor- related with the foreign currency debts. Theoretically, one can build a model where desires to stabilize cash flows lead corporates to borrow more in foreign currency when they export more or leverage constraint from tail risk, like Vale at Risk constraint, incentivizes exporting firms to issue more foreign currency debts since exporting firms have less risk from foreign currency debts as their profits increase in local currency depreciation. Empirically, finding relevant evidence is challenging and it is beyond the scope of this paper. Recently, Dalgic (2020) documents that in Turkey foreign currency debts are centered on large exporters. The central message in this paper is the decline of the risk from foreign currency debts in EMS, but the rise of the new risk from equity and local currency portfolio investment capital flows. To focus on the new channel uncovered in this paper, I do not pioneer more about the assessment of foreign currency debt risks of nonfinancial corporates in EMEs. Instead, I highlight the findings in the following remark. In different types of models where profits of exporters increase in local currency depreciation, the exporters whose profits increase more in local currency depreciation will borrow more in foreign currency. Then more foreign currency debts do not necessarily lead to higher fragility: during a risk-off event, the positive effects on the profits largely offset the negative impacts on the foreign currency debts. Discussion of the implication Broadly speaking, the implication here is profits of the exporters are likely to increase in local currency depreciation. To the best of my knowledge, such effects 210 have not been extensively studied. However, all the underlying assumptions are in line with re- cent progress in the international macroeconomics literature. As revealed in influential papers of Dominant Currency Pricing (DCP), for example, Gopinath and Stein (2020) and Gopinath et al. (2019), most tradable goods are denominated in dominant currency, in fact USD. On the contrary, the “domestic” costs of corporates in EMEs are local currency denominated and rigid in short run. The obvious example is the wages in EMEs and wage rigidity in EMEs has been discussed in many papers such as Schmitt-Grohe and Uribe (2016). Then, it is clear that local currency depreciation itself boosts the profitability of exporters in EMEs as their revenues are denominated in foreign currency like USD, whereas much of their costs are denominated in local currency. One assumption that can alter the conclusion above is positive covariance between the risk appetite shocks and foreign demands for exports from EMEs. In other words, if the trade shocks substantially move together with the risk appetite shocks, then exporters will be hit harder. Cer- tainly, the two different shocks are positively correlated with each other to some extent, but many specific cases of risk-off shocks hardly accompany trade shocks. For instance, risk-off shocks driven by US Fed monetary policy normalization do not necessarily cause negative trade shocks as Fed would roll back the expansionary monetary policy, conditioning on Fed judges US economy is resilient enough. However, a global crisis like 2008 Global Financial Crisis or recent COVID-19 Crisis accompanies both large trade shocks and risk appetite shocks. 17 To accommodate such tail risks, I need more informative data and need to conduct a more sophisticated theoretical analysis. These are beyond the scope of this paper. One straightforward prediction from the model is a positive correlation between trade openness and net foreign currency debts of nonfinancial corporate sectors. In my 20 sample EMEs, the observed correlation is 0.31. In reality where pricing in exports and price elasticities of exporting goods are different among EMEs, sensitivities of corporate profits to exchange rates depend on 17 The extreme crises must dampen the profitability of the exporters in EMEs, but the role of exchange rates in this context is unclear. Besides the negative demand shocks, the higher exchange rates still help the exporters with lessening the negative impacts. 211 many factors other than trade openness. Identification of the factors is also beyond the scope of this paper. Last, I emphasize that the conclusion here does not imply net foreign currency debts of non- financial corporates in EMEs are efficient from the viewpoint of financial stability. Rather, the statement should be understood as positive correlations between countercyclical components in nonfinancial corporates profits and foreign currency debts in the sectors: thus, more foreign cur- rency debts do not necessarily lead to higher fragility in the data. The foreign currency debts in reality may be determined by the risk-hedging desires of the corporates, but the foreign currency debts in the first-best equilibrium should be determined taking account of pecuniary externalities and aggregate demand externalities. 18 Since the focus in this paper is on the channels through which risk appetite shocks are transmitted, I do not further analyze the efficiency of foreign cur- rency debts. 19 B.4 Who Can Borrow More in Equity and LC Bond? An important question related to the research agenda in this paper is “Which EMEs can borrow more in equities and LC debts?” In other words, “How can we explain different structures of external liabilities of EMS, which were reported in section 2 in this paper?” Answer to these questions is also important for the validity of the cross-country regression in section 2, which motivated our model; explaining different amounts of equity external liabilities and LC debts can help me with dealing with the concern about the endogeneity. 18 For a more serious analysis, I need more detailed information about different features of exports in the EMEs, such as different pricing or price elasticities of the exports. 19 My presumption is the net foreign currency debt levels observed in the data are still higher than the first-best equilibrium levels. This is because the risk-on/off shocks impact EMEs not only through foreign currency debts, but also through the capital market channel, which I uncover and highlight in this paper. To stabilize the economy from the global financial cycles, the net worth of the corporates need to be procyclical to exchange rates; higher exchange rates (local currency depreciation) increase the cash flows of the corporates, as higher (lower) exchange rates follow risk-off (on) shocks. To explain more, the net worth of the corporates are stabilized from exchange rates with some foreign currency debts, but without the foreign currency debts the local currency depreciation can strengthen the net worth so as to cover the negative impacts of risk-off shocks through the capital price. 212 Figure B.1: Stock Market Capitalization, Trade Openness, and LC external liabilities Note: all measures are the average of yearly data from 2012 to 2017 I examined a few economic fundamentals that are possibly correlated with the amounts of ex- ternal equity liabilities and LC debts. While most of the fundamentals such as economic growth rate, inflation or government debt to ratio do not show significantly high correlations, two funda- mentals turn out to be particularly relevant. First, amounts of both equity and LC debt are highly correlated with domestic capital markets: that is, EMEs with larger stock markets tend to receive more equity portfolio investments from abroad and similarly, EMEs with larger bond markets tend to receive more (LC denominated) bond portfolio investments. For the LC debts, trade openness also significantly matter. The correlation between equity external liabilities 20 to GDP ratio and stock market capitalization to GDP ratio is 0.91. For LC debts, the correlation between domes- tic bond market outstanding to GDP ratio and LC debts to GDP ratio is 0.61, and the correlation between trade openness and LC debts to GDP ratio is even higher; the correlation is 0.69. To explain the correlation, I consider an optimal portfolio construction of global investors. In the appendix, I analyzed how global investors under VaR constraints build her portfolio. Imag- ine a global investor who has a portfolio composed of m 1 number of assets. For the global 20 As data permits, I only included local currency denominated equities, which are issued in each of the EMEs. 213 investor who is seeking an investment opportunity in equity market in country i, the amount of the investments is determined by p i q i ; : = R i R f R m i R f q i b q i p i ;Q i k d i s 2 i =s 2 m i + R i R f R m i R f q i b q i p i ;Q i k d i R i R f R m i R f + 1 (B.7) where R i is the expected return to the equity in country i,s 2 i ands 2 m i are the standard deviations of asset i and the pre-determined portfolio. q i is the covariance of the equity return and the return to the pre-determined portfolio. From the findings in section 2 and 3, I know that the covariance between equity market return in country i and the risk appetite shock to global investors q i depends on the share of the investors in the equity market in country i, b q i = p i p i +Q t k d i . 21 The size of the capital market can be proxied by Q i k d i . Then it is obvious larger Q i k d i must be associated with higher p i : holding the investment by global investors p i fixed, larger Q i k d i leads to a lower share of global investor b q i so as to lower the covariance with global risky asset pricesq i , and then more global investors are attracted to the market because of the low exposure to the global systemic risk. More formally, equation (B.7) forms a fixed problem in which global investors’ share in the equity market in country i is determined in equilibrium. Global (foreign) investors make a decision based on the exposure to their own risk and the exposure is determined by the decision of the investors. Therefore the following proposition follows. Proposition 9 In the optimal portfolio decision in (B.7), 1) Given R i ands 2 i , p i increases in Q i k d i . 2) Moreover, for different country i and j, ifs 2 i =s 2 j and R i = R j ; then b q i = b q j in equilibrium. The first statement summarizes the discussion above. To see what the second statement means, imagine there are two different EMEs and the average return R i and idiosyncratic volatilitys i 2 are the same in the two EMEs. 22 If the share of global investors is higher in one country, then the 21 Here I am abstracting from currency and also ignoring the possible comovements from the trade channel. 22 Precisely, this is slightly misleading since the foreign investor participation would alter the average return and idiosyncratic volatility. However, these two factors should be less sensitive to foreign investors’ share in the market than the covariance. Hence, the main insights would not be changed. 214 equities in the country are more sensitive to the risk appetite shocks to the global investors so that the equities are less attractive to the global investors since the returns are more correlated with the investors’ own risk profile. Then the global investors move capitals to the other country and therefore the global investor shares in the two EMEs are must be the same in equilibrium. A similar story can be applied to the relationship between trade openness and LC bond portfolio investment. Higher trade openness, given LC portfolio bond investment, make the currency less correlated to the investors’ own risk profile, and thereby attracting more global investors into the LC bond markets in the country. B.5 Omitted Algebras and Proofs Proof of Proposition 1 To prove the statements in the proposition, I find it useful to found the lemma below. Lemma 5 The equilibrium of the small open economy is represented by the equations below, which shows the capital market clearing condition, the foreign exchange market clearing condition, and the the law of motion for the risk appetite of global investors respectively. Q t = f 1 (e t ;v t ) 0= f 2 (Q t ;e t ;v t ) Proof) First, remember that I assume c 1 i = 0 so that the capital inflows from global investors, p k t and p b t , are determined regardless of the expectation. The resources constraint of this economy is as follows. AK a t1 L 1a = c d t + I t + j 2 I t K t1 2 K t1 + G+ Ex t 215 The optimality condition of the capital producer, 1+jI t = Q t , pins down the investment given the capital price, Q t . From the equation (2.19), we know Q t is a function of the states, v t ande t . Real exchange ratee t determines the exports, Ex t . Since the output in this economy is determined from the previous period, K t1 , capital price Q t and real exchange ratee t determine the domestic goods consumption c d t . The real exchange rate e t is determined by the foreign exchange market clearing condition (2.21). Note that the imported goods consumption is determined given domestic goods consump- tion and the real exchange rate by the equation below. c m t e t = c d t w 1w (B.8) Thus, c m t is a function of Q t ,e t and v t . In equation (2.21),e t is determined by c m t , and the invest- ments by global investors, p k t and p b t , which are solely determined by v t . This tells me that Q t and v t uniquely determinee t . This completes the proof. Now we prove the proposition. Plugging in N t =s (z t + Q t )k d t1 R t d t1 +x(z t + Q t )k d t1 into equation (2.19) yields Q t = j j 1 +f(s+x) k d t1 K t1 1 + s j 2 j 1 +f(s+x) k d t1 K t1 1 2 + 4j p k t e t +f((s+x)z t k d t1 sR t d t1) K t1 2 (B.9) dQ t dn t = e t 1+ de t dn t e t s f(s+x) k d t1 p k t e t +j 1 K t1 p k t e t K t1 p k t e t 2 + 4j 1 K t1 p k t e t 1+f (s+x)z t k d t1 sR t d t1 p k t e t (B.10) To prove dQ t dn t > 0, I need to show de t dn t e t < 1. 216 Plugging into the equation (B.8) to the foreign exchange market clearing condition yields 1 1w Y t e g t w 1w Y t I t j 2 I t K t1 2 K t1 G ! = R k t1 k f t1 + R b t1 b f t1 e t p k t + p b t By the implicit function theorem, I have de t dn t e t = p k t 1+ dQ t d(p k t e t) w 1w j 1 + I t k f t1 + p b t Y t 1w ge g1 t + p k t 1+ dQ t d(p k t e t) w 1w (j 1 + I t ) k f t1 + p b t (B.11) It is obvious that de t dn t e t < 1: It proves the first statement in the proposition. Next I show ¶ 2 Q t ¶n t ¶ b q t j e t < 0. To derive the desired result, I need to show ¶Q t ¶n t j e t decreases in p k t e t . For the purpose, I find it is convenient to denote K t1 p k t e t by x t . Then the term in the denominator turns out to be a following quadratic equation. H(x t )= 0 @ f(s+x) k d t1 K t1 + j 1 + 1 ! 2 + 4j 1 f (s+x)z t k d t1 sR t d t1 K t1 ! 1 A 2 x 2 t +4j 1 x t Since x t is inverseIy related with p k t e t , I want to show H 0 > 0 for x t > 0. It is equivalent to x t > 4j 1 2 f(s+x) k d t1 K t1 +(j 1 + 1) 2 + 4j 1 f (s+x)z t k d t1 sR t d t1 K t1 ! Since x t > 0;the sufficient condition for the inequality is f(s+x) k d t1 K t1 + j 1 + 1 ! 2 >4j 1 f (s+x)z t k d t1 sR t d t1 K t1 ! As I assumed in the proposition. 217 Lastly, I prove the third statement. It is trivial. Notice k d t = N t Q t f and furthermore N t Q t =(s+x)k d t1 + (s+x)z t k d t1 sR t d t1 Q t It is straightforward that N t Q t increases in Q t if (s+x)z t k d t1 sR t d t1 < 0. This completes the proof. Corollary 3 if K t1 t K t , k d t1 t k d t ;and Q t t 1, then I can approximate ¶Q t ¶n t j e t as follows. ¶Q t ¶n t j e t t e t 1+ dS k t dn t S k t r f(s+x) 1q t q t + j 1 q t 1 q t 2 + 4 j 1 q t 1+f 1q t q t (s+x)z t sR t f1 f It is easy to see that, given the value of dS k t dn t S k t , approximated ¶Q t ¶n t j e t does increase inq t if f(s+x) 1q t q t + j 1 + 1 2 >4j 1 f 1q t q t (s+x)z t sR t f1 f . Corollary 3 pro- vides a comparative statics matching the empirical results from the cross-country panel regressions. Proof of Proposition 2 See ifc 0 k =c 0 b = 0, then dQ t dn t = e t 1+ dS k t dn t S k t + de t dn t e t s f(s+x) k d t1 p k t e t +j 1 K t1 p k t e t K t1 p k t e t 2 + 4j 1 K t1 p k t e t 1+f (s+x)z t k d t1 sR t d t1 p k t e t (B.12) de t dn t e t = p k t 1+ dQ t d(p k t e t) w 1w j 1 + I t k f t1 1 dS k t dn t S k t + p b t 1 dS b t dn t S b t Y t 1w ge g1 t + p k t + p b t + p k t dQ t d(p k t e t) w 1w (j 1 + I t ) k f t1 1 dS k t dn t S k t (B.13) It is easy to see that under the assumptions in the proposition, 0< de t dn t e t < 1. Then by the assumption, de t dn t e t > 0 and moreover, by the assumption that ge g1 t 1w > dS b t dn t S b t h b t , de t dn t e t < 1. 218 Backing to dQ t dn t , see dQ t dn t = e t 1+ dS k t dn t S k t + de t dn t e t s f(s+x) k d t1 p k t e t +j 1 K t1 p k t e t K t1 p k t e t 2 + 4j 1 K t1 p k t e t 1+f (s+x)z t k d t1 sR t d t1 p k t e t As dS k t dn t S k t ! 0,1+ dS k t dn t S k t !1, but de t dn t e t is strictly smaller than 1. It proves that dQ t dn t < 0 for dS k t dn t S k t small enough. Lastly, I show ¶h t ¶h b t > 0> ¶h t ¶h k t or ¶h t ¶h b t > ¶h t ¶h k t > 0 where h h b t ;h k t = de t dn t e t , andh b t = p b t Y t andh k t = p k t Y t . I transform equation (58) to de t dn t e t = h k t 1+ dQ t d(p k t e t) w 1w j 1 k f t1 1 dS k t dn t S k t +h b t 1 dS b t dn t S b t ge g1 t 1w +h b t +h k t 1+ dQ t d(p k t e t) w 1w j 1 k f t1 1 dS k t dn t S k t (B.14) Since 1 dS b t dn t S b t > 1 and 0< 1 dS b t dn t S b t < 1, but 0< de t dn t e t < 1, it is straightforward that ¶h t ¶h b t > 0, and ¶h t ¶h b t > ¶h t ¶h k t . Furthermore, since de t dn t e t < 1 and the “coefficient” ofh k t in the numerator is smaller than the denominator. That is, 1+ dQ t d p k t e t w 1w j 1 k f t1 ! 1 dS k t dn t S k t < 1+ dQ t d p k t e t w 1w j 1 k f t1 1 dS k t dn t S k t Therefore, it can be either of ¶h t ¶h k t > 0 or ¶h t ¶h k t < 0. 219 Proof of Corollary 1 dI t dn t < 0 is trivial. To show d(NX t ) dn t > 0, let’s recall the net export ise g1 t Y t c m t . The exports obviously increase in n t as I showed de t dn t > 0 under the conditions. Then I only need to show dc m t dn t < 0. See(1w)c m t e t =wc d t . From the resource constraint, c m t e t = w 1w (Y t I t G EX t ) By the assumption dI t dn t + ¶EX t ¶e t de t dn t > 0, c m t e t decreases in n t . Since de t dn t > 0 by the proposition 1, dc m t dn t < 0. Proof of Lemma 4 The result comes from the allocation of the output. Notice Y t = C t + I t + G+ EX t I factorize the aggregate output into domestic demands and foreign demands. That is, Y d t = C t + I t + G= Z 1 0 y d i;t h1 h di ! h h1 EX t = Z 1 0 y i;t h1 h di h h1 y d i;t and y i;t are the intermediate inputs for domestic demands and foreign demands (exports) re- spectively. By the assumption, the price of export goods in the foreign market is fixed. Then, the demand from the foreign market EX t is invariant to the risk-on/off shock n t . Then, accordingly, the do- mestic demand is invariant as well. The value of the total output in terms of the domestic price is Z 1 0 y d i;t h1 h di ! h h1 + p t e t p t Z 1 0 y i;t h1 h di h h1 = w t L+ z t K t1 +p t (B.15) 220 See R 1 0 y d i;t h1 h di ! h h1 and R 1 0 y i;t h1 h di ! h h1 are invariant to the risk-on/off shocks. Fur- thermore, p t and p t are invariant as well because p t = h h 1 mc t Notice the marginal cost mc t is invariant as well because there is no technological shock. In the same way, w t and z t are invariant as well. In the equation (60), the LHS increases inn t since risk-off (on) shock raises (lowers)e t . To hold the equality between the LHS and the RHS in the equation (60), the profitp t has to change accordingly. This gives me the desired result. To see it another way, the real profit measured by the domestic price is p t = 1 h Y d t |{z} Domestic Pro f it +Y t (p t ) h p t e t p t 1h h | {z } Export Pro f it Since Y d t is given, it is easy to seep t increases inn t . Proof of Proposition 9 Notice equation (B.7) defines a fixed problem. First, I show there exists a unique solution. It is trivial. RHS decreases in p i as I assume q i increases in p i , given Q i k d i . Since the RHS is positive for p i = 0, there exists a unique value of p i that solves the equation (52). Then it is easy to show the first statement in the proposition. Larger Q i k d i . leads to a lower b q i and thus a lowerq i . To restore the equality, p i has to increase. The second statement is also easy to show. Suppose Not. That is, b q i 6= b q j , but R i = R j and s i =s j . Here let’s say b q i > b q j . Then the global investor will move capitals to j country until q i =q j and equivalently b q i = b q j . B.6 Estimation of the Parameters First, I describe how I estimated the parameters c 0 k c 1 k and c 0 b c 1 b . Transform equation (2.46) and (2.47) as follows 221 p k t = Q t k f t e 1 t = 1 e G k e v t " 1+ c 1 k c 0 k E t e t e t+1 R k t+1 R m t+1 (v t ) # p b t = b f t e 1 t = 1 e G b e v t " 1+ c 1 b c 0 b E t e t e t+1 R b t+1 R m t+1 (v t ) # Serious estimation of c 1 k c 0 k and c 1 b c 0 b poses a challenge. As I emphasized, these parameters matter only to the extent that it matters for capital flows. Thus, I estimate the parameters from the capital flows data in Korea. The statistics portal of the financial supervisory service (FSS) provides monthly data of foreign investors’ holding of public equities and bonds issued in domestic financial markets in Korea. The data of the equity holdings and the bond holdings begin from January 2005 and January 2006 respectively. However, it is observed in the data of the bond that the foreign investors’ holding of the bonds in the Korean markets have been on a “stable trend” since 2010; before 2010, the foreign holdings had kept rising except for global financial crisis 23 . Therefore, for the bond, I only use the data after 2009. First, I take the log of the data of the public equity and bond holdings by foreign investors, and then remove the linear trend. Let’s denote the detrended portfolio investment by thee p j t , and let the growth ofe p j t be g e p j t Define the growth of the portfolio investment g p j t p j t p j t1 . Then I have ln g p j t =(v t v t1 )+ln 1+ c 1 j c 0 j E t e t e t+1 R k t+1 R m t+1 (v t ) ! ln 1+ c 1 j c 0 j E t e t e t+1 R k t+1 R m t+1 (v t ) ! (B.16) where e t e t+1 R k t+1 and e t e t+1 R b t+1 are the returns to the Korean stock market Index (KOSPI Index) and 3 year maturity government bond in the USD, and the return to the alternative investment, R m t+1 (v t ), is the quarterly yields on the BAA grade corporate bond in the US. Now I have two moment conditions to estimate the c 1 j c 0 j E t1 h g p j t g e p j t i = 0 (B.17) 23 One way to understand the period is to think the time as a transitional period like a transitional path from one steady state to another steady state. 222 Table B.1: GMM estimation results c 1 k c 0 k c 1 b c 0 b Estimated Values 3.645*** 9.363*** (0.124) (0.015) Observations 174 112 P-value of J-test 0.411 0.932 E t1 h (v t v t1 ) g p j t g e p j t i = 0 (B.18) The intuition of using the difference in VIX is that the difference between the theoretical moment and empirical moment should be uncorrelated with the growth of VIX once the model is correctly specified under the parameter values close to the true value. One difficulty in this GMM estimation is the limited number of the observation. To circumvent the problem, I compute the changes in the portfolio investment from a quarter ago in every month, and accordingly the returns. That is, for every month, I compute the changes in the portfolio investments and VIX in “three months”, and spreads in the USD between the portfolio investments in Korean markets and the alternative asset, also in three months. Then I estimate the parameters from the “quarterly” changes in every month. In such a way, I can increase the number of the sample to 174 and 112 for the equity and bond respectively. The results of the estimation are reported in table B.1 above. The estimation should be understood as a way to calibrate the model to the Korean economy in the context of this paper, not a very robust and precise estimation. Next, I illustrate the estimation ofc m andc . I can simply estimate the parameters by running the OLS regressions. ln(r j t )=a j + b b j V IX t + e j t (B.19) 223 Table B.2: Sensitivity of the Interest Rates to VIX (1) (2) (3) (4) V IX t 0.506*** 0.643*** (0.132) (0.090) n t 0.203*** 0.475*** (0.060) (0.045) ln(r j t1 ) 0.934*** 0.859*** (0.037) (0.034) Observations 83 83 83 82 R-square 0.154 0.892 0.388 0.903 r m and r are the BAA corporate bond yields in the US and the JP Morgan Emerging Market Bond Index 24 . Notice that the purpose for the regression is to estimate the realized sensitivity of the yields on the risky bonds to the risk appetite, measured by VIX. One of the issues in the regression above is the autocorrelation in the error terms, which are possibly correlated with VIX. The regressions must be plagued by the autocorrelation. To see how much it matters, I run another regression below. ln(r j t )=a j +r j ln(r j t1 )+ b b b j n t + e j t (B.20) wheren t = ln(VIX t )E t1 [ln(VIX t )]. 25 The estimation results are introduced below. Notice that the results of (B.19) and (B.20) are similar to each other once we consider estimated autocorrelation coefficientsr j . A tricky part is that in the calibration, I implicitly assumed that the global risk-appetite process is a part of VIX; in other words, VIX includes the risk appetite and some noises. That means if I directly use the parameter value, I will underestimate the impact of risk-appetite shocks on the interest rates. Thus, I adjust the coefficient values so that one standard deviation of risk-on/off shock in the model changes the interest rates by the magnitudes same with the regression results. The standard deviation of v t in the calibrated model is 0.171, while the 24 I used the spread on BAA corporate bonds and similarly do not convert the EMBI index to the yields on the sovereign bonds of EMEs. In fact, in the sample period, the real return to US government bond is close to zero. 25 Same as the bank level regression,n t is estimated from AR (1) model. Extension of the AR (1) model to ARMA (1) does not significantly change the results. 224 standard deviation of VIX in the sample period is 0.353. The converted parameter values for the BAA corporate bond and EMBI index are 1.046 and 1.372 respectively. B.7 Contagion to Credit Market As I stated, I abstract from rich features in financial markets in reality to focus on the key insight and make the analysis more tractable. As the channel in the model works through capital market, I assumed that all firms issue equity type securities, which are purchased by the representative do- mestic banks and global investors. However, firms in reality finance through different instruments and the most common instrument is bank loan. Furthermore, there are different types of financial intermediaries. In almost all countries, there are some commercial banks, which take deposits and supply credits to nonfinancial corporates in the form of bank loan, and on the other hand, some market-oriented financial intermediaries, like investment banks in the US or insurance companies, finance different ways and invest in different financial securities such as equities and bonds. In this section, I introduce another empirical exercise in which I test the capital market channel in realistic environments. In the regressions of investment-type financial intermediaries in Korea, I tested how the impacts of global financial shocks on the capital markets affect financial interme- diations of the financial intermediaries. Here, I test how the impacts on the capital markets are propagated into credit markets. Like many other countries, commercial banks in Korea mainly finance through deposits, but they also rely on market funding. The commercial banks issue short- term bank debentures and sell certificate of deposits (CDs) and the financing of these forms account for 13.4% of the total liability of the banks in my sample period. Much of the bank debentures and CDs are purchased by the ”investment bank” type financial intermediaries. According to the un- official statistics, approximately 60-70% of the securities are purchased by the investment banks. Henceforth, let’s call the liabilities of bank debentures and CDs ”noncore” liabilities. 26 26 The noncore liability of commercial banks in Korea and related fragility of the banks during the global financial crisis are well documented and discussed in Shin and Shin (2011). 225 Figure B.2: Contagion to Credit Market Then a natural prediction is global financial shocks are propagated into commercial banks, credit markets, through the noncore liability. To be more specific, a risk-off shock results in falls in asset prices in the capital market in Korea, which induce the investment banks to deleverage. The contraction of the investment banks decreases the investment banks’ demands for bank debentures and CDs issued by commercial banks. The commercial banks will have trouble funding via bank debentures and CDs and as a result, the banks will be forced to reduce their credit supplies to nonfinancial corporates and households. 27 To make it short, the impacts on the capital market are contagious to the credit market through the noncore liabilities of commercial banks. On the other hand, the commercial banks have some mild exposures to Korean won deprecia- tions. Net foreign currency debts of the banks are, on average, 1% of the total assets in the sample period. Therefore, I can test how the two different channels, the capital market channel and the 27 The flow of funds described here is quite different from the US and perhapd other emerign and advanced economies. Hence, I do not claim the propagation channel and mechanism here are general. It is Korea specific rather general. The goal of the empirical study is to test the capital market channel mechanism in realistic environ- ments although it is little specific. 226 exchange rate channel, work in financial intermediations of commercial banks in Korea. Figure 11 below the two transmission channels in the regressions, which I introduce below. The regression equation is as below. Dln(A i;t )=a i +a t +b 0 q i;t1 n t +b 1 f c i;t1 n t +G 0 X i;t1 +e i;t (B.21) q i;t1 is the ratio of the noncore liabilities to the total liabilities and f c i;t1 is the ratio of net foreign currency debt to total liability ratio. n t is unexpected changes in VIX, ln(VIX t )E t1 [ln(VIX t )], which is the same as the capital market channel regressions. If commercial banks with higher reliance on the noncore liability will be affected by the risk- on/off shocks more than others,b 0 will be significantly negative. Similarly,b 1 will be significantly negative if risk-on/off shocks affect asset growth of the commercial banks through net foreign currency debts. While I control different characteristics of the commercial banks, it is crucial to control the group heterogeneity of the banks. As usual in many countries, the number of commer- cial banks in Korea is limited: there are 16 banks, including some foreign owned banks and the banks specialized in certain provinces. Among the two special groups, the group of foreign owned banks might move in a different way than other banks. Hence, I control the group by putting the interaction term between the foreign group and the unexpected changes in VIX. There is also a special bank, Industrial Bank of Korea (IBK), which heavily relies on the special bank debenture issuance. The problem is IBK is in part owned by government bank, and therefore their opera- tions are heavily affected by government policy, although the behavior of IBK is similar to other commercial banks in many aspects. Considering the unique feature of IBK, I control the bank in a similar way to the foreign owned bank group. In addition, I control financial derivative holdings and sizes of the banks as the two variables turn out to be significant. In particular, controlling the derivative holdings is important since the banks hedge risks related to market fluctuations driven by global financial shocks; most impor- tantly, exchange rate risks. 227 Table B.3: Contagion to Credit Market (1) (2) (3) (4) (5) (6) (7) Dln A R i;t1 0.06 0.01 0.01 0.02 0.02 0.03 0.03 (0.04) (0.04) (0.04) (0.04) (0.04) (0.04) (0.04) q i;t1 n t -0.04* -0.25*** -0.17** -0.16** -0.15** f-0.15** -0.13 † (0.02) f(0.08) (0.07) (0.07) (0.07) (0.07) (0.08) FC i;t1 n t 0.16 0.13 0.13 0.15 (0.19) (0.21) (0.21) (0.20) D i;t1 A i;t1 n t 1.19*** 1.19*** 1.12*** 1.17*** 1.15 (0.17) (0.16) (0.18) (0.17) (0.18) Foreign ownedn t -0.05*** -0.04*** -0.04*** -0.04*** -0.04*** -0.04*** (0.01) (0.01) (0.01) (0.01) (0.01) (0.02) Localn t -0.01 -0.01 -0.01 (0.01) (0.01) (0.01) IBKn t 0.08*** 0.05*** 0.04* 0.03 † 0.03 0.03 (0.03) (0.02) (0.02) (0.02) (0.03) (0.03) q i;t1 -0.00 0.06 0.07* 0.07* 0.07* 0.07* 0.08* (0.04) (0.04) (0.04) (0.04) (0.04) (0.04) (0.04) FC i;t1 -0.05 -0.04 -0.01 -0.01 (0.06) (0.06) (0.06) (0.06) D i;t1 A i;t1 -0.09 -0.09 -0.08 -0.08 -0.09 -0.08 (0.08) (0.08) (0.08) (0.08) (0.08) (0.08) Size i;t1 -0.64*** -0.79*** -0.81*** -0.81*** -0.76*** -0.77*** (0.15) (0.15) (0.15) (0.15) (0.16) (0.16) A R i;t1 N i;t1 c i;t -0.01 -0.01 (0.09) (0.09) C i;t1 A i;t1 n t 1.41 (2.01) Country FE YES YES YES YES YES YES YES Time FE YES YES YES YES YES YES YES Observation # 727 727 727 727 727 727 727 R-squared 0.33 0.39 0.44 0.44 0.44 0.45 0.45 # of banks 16 16 16 16 16 16 16 Note: 1) *** p¡0.01, ** p¡0.05, * p¡0.1, * p¡0.1, † p¡0.15. 2) The dependent variable is the growth of asset, except for cashes and tangible assets. 3) D i;t1 A i;t1 is derivate assets to total asset ratio. C i;t1 A i;t1 is the cash holding to total asset ratio. c i;t is the estimate of changes in values of financial securities held by the banks, due to unexpected changes in VIX. A R i;t1 N i;t1 is the leverage ratio, the effective asset to net worth ratio, where the effective asset is the all the assets excluding cashes and tangible assets. 228 The results in table B.3 are mostly as predicted. The coefficient of the interaction term between the noncore liability and the risk-on/off shock is negative and significant in all different specifica- tions. In contrast, all the interaction terms of net foreign currency debt are insignificant. That probably reflects that the net foreign currency debts are very small parts of the total assets and thus the balance sheet effects cannot be sizable enough. As the net foreign currency debts of the banks are relatively small, the banks might easily hedge their risks in derivative markets. B.8 Regressions with Sector Level Currency Mismatches Same as the regressions without aggregate level currency mismatches, I first introduce the results of the exchange rate regressions. Again, for brevity, I introduce only the estimated coefficients of the key variables. The results for the other control variables are relegated to the appendix. I denote net foreign currency assets (foreign currency assets of debt instruments minus foreign currency debts) by NFC, and denote households, financial corporate sector, nonfinancial corporate sector, and government by HH, FC, NFC and G respectively. Hence, HH NFC indicates net foreign currency assets of households in the form of debt instrument 28 . Next, I introduce the results of stock indices regressions. Same as the exchange rate regressions with sector level net foreign currency assets, the overall results are much similar with the results of aggregate level currency mismatch: the foreign investor share is still negative with little more significance, and net foreign currency assets in different sectors are all insignificant. The interpretation of the insignificance of the sector-level currency mismatches is that foreign currency debts of each of the sectors in the EMEs are matched with foreign currency assets or foreign currency denominated revenues, and then accordingly local cur- rency depreciation do not seriously impact balance sheets of the different sectors in the EMEs. 28 Unlike the regressions of the aggregate level currency mismatches, I use net positions rather than putting foreign currency assets and liabilities separately. The net foreign currency assets allow me for more straightforward interpre- tations, as I focus on foreign currency asset and debt evaluation effects due to exchange rate changes. However, the results introduced below are highly consistent although I put foreign currency assets and debts separately. The results are omitted due to limited space. 229 Table B.4: Exchange Rate Regressions Sector level (1) (2) (3) (4) (5) (6) (7) Dln(V IX) t 0.040** 0.043*** 0.044*** 0.044*** 0.046*** 0.056** 0.046** [0.016] [0.013] [0.013] [0.013] [0.013] [0.022] [0.020] LCB GDP j t ” 0.189** 0.215*** 0.123* 0.152* 0.182** 0.193** [0.076] [0.086] [0.070] [0.083] [0.093] [0.088] LCE GDP j t ” -0.005 -0.028 -0.015 -0.026 -0.003 -0.053 [0.036] [0.034] [0.032] [0.073] [0.063] [0.053] HH NF GDP j t ” -0.242* -0.190 -0.196 -0.201 -0.100 [0.137] [0.144] [0.146] [0.140] [0.132] FC NF GDP j t ” -0.046 -0.054 -0.027 -0.060 -0.048 -0.038 -0.082 [0.045] [0.047] [0.041] [0.049] [0.060] [0.053] [0.074] NFC NF GDP j t ” -0.028 -0.024 -0.094 † -0.081 -0.094 -0.063 -0.039 [0.027] [0.025] [0.063] [0.060] [0.066] [0.051] [0.054] G NF GDP j t ” 0.079 0.061 0.081 0.038 0.092 [0.088] [0.080] [0.088] [0.076] [0.069] Reserve GDP j t1 ” -0.064** -0.068*** -0.011 -0.028 -0.029 -0.026 -0.029 [0.028] [0.027] [0.036] [0.038] [0.049] [0.043] [0.028] Country Fixed Effect Yes Yes Yes Yes Yes Yes Yes Time Fixed Effect No No No No No No Yes # of Obs. 1577 1577 1577 1577 1577 1577 1577 R-squared 0.052 0.046 0.027 0.022 0.033 0.090 0.232 Note: 1) *** p¡0.01, ** p¡0.05, * p¡0.1, * p¡0.1, † p¡0.15. 2) Exclude Brazil as there is no available data of foreign currency deposit in Brazil. 2) LCD: Local Currency Debt, LCB: Local Currency Bond Portfolio, LCE: Local Currency Equity, HH NFC: net foreign currency assets of household sector , FC NFC: net foreign currency assets of financial corporate sector, NFC NFC: net foreign currency assets of non-financial corporate sector and G NFC: net foreign currency assets of government sector. 3) Driscoll-Kraay standard errors. 4) Regression (7) adds more controls (commodity price index and related groups) to regression (5). 5) Time fixed effects are not two-way fixed effects, but time dummies (random effects), because one of the key explanatory variables, the VIX log difference is the time series variable. To explain a little more, let’s think of the financial corporate sectors that borrow abroad in for- eign currency and supply the foreign currency to domestic foreign exchange markets. A risk-off shock can cause local currency depreciation so as to dampen the balance sheets of the financial corporates, which in turn results in less foreign currency supplies due to the deleveraging of the fi- nancial intermediaries. However, such a scenario can be realized only when the financial corporate sectors have large enough net foreign currency debts and we saw in many EMEs that the financial corporate sectors are squared-off in foreign currency. Similarly, unexpected local currency depreci- ation probably increase real debt burdens of nonfinancial corporates in EMEs since the corporates have sizable net foreign currency debts as we confirmed in the data in the last subsection. However, 230 Table B.5: Stock Indices Regressions Sector level (1) (2) (3) (4) (5) (6) (7) Dln(V IX) t -0.078*** -0.078*** -0.085*** -0.077*** -0.077*** -0.057*** -0.031 [0.016] [0.016] [0.015] [0.017] [0.018] [0.018] [0.027] LCB GDP j t ” 0.104 0.103 0.200* 0.209** 0.243** 0.248** [0.119] [0.120] [0.105] [0.103] [0.104] [0.116] LCE Mkt: Cap: j t ” -0.091*** -0.089** -0.115*** -0.117*** -0.113*** -0.126** [0.043] [0.038] [0.039] [0.038] [0.037] [0.063] HH NF GDP j t ” 0.065 0.138 0.136 0.094 0.090 [0.125] [0.120] [0.121] [0.124] [0.091] FC NF GDP j t ” -0.011 0.100 -0.031 -0.024 0.004 0.049 [0.095] [0.080] [0.086] [0.086] [0.074] [0.164] NFC NF GDP j t ” -0.044 -0.026 -0.002 -0.006 -0.003 -0.004 [0.031] [0.061] [0.058] [0.062] [0.063] [0.053] G NF GDP j t ” -0.001 -0.006 -0.042 -0.056 [0.074] [0.063] [0.079] [0.089] Reserve GDP j t1 ” 0.051 † 0.051 0.018 0.029 0.026 0.029 -0.023 [0.035] [0.036] [0.019] [0.041] [0.044] [0.042] [0.066] Country Fixed Effect Yes Yes Yes Yes Yes Yes Yes Time Fixed Effect No No No No No No Yes # of Obs. 1577 1577 1577 1577 1577 1577 1577 R-squared 0.055 0.059 0.067 0.058 0.053 0.064 0.208 Note: 1) *** p¡0.01, ** p¡0.05, * p¡0.1, * p¡0.1, † p¡0.15. 2) LCD: Local Currency Debt, LCB: Local Currency Bond Portfolio, LCE: Local Currency Equity, HH NFC: net foreign currency assets of household sector , FC NFC: net foreign currency assets of financial corporate sector, NFC NFC: net foreign currency assets of non-financial corporate sector and G NFC: net foreign currency assets of government sector. 3) Exclude Brazil as there is no available data of foreign currency deposit in Brazil. 4) Driscoll-Kraay standard errors. 5) Regression (7) adds more controls (commodity price index and related groups) to regression (5). 6) Time fixed effects are not two-way fixed effects, but time dummies (random effects), because one of the key explanatory variables, the VIX log difference is the time series variable. currency depreciation can boost the profitability of some of the nonfinancial corporates who export to foreign markets as their revenues from the exports are foreign currency denominated while much of their costs like wages are local currency denominated. If the net foreign currency debts of the nonfinancial corporates are correlated with more foreign currency revenues, then foreign currency depreciation do not necessarily lead to deleveraging of the nonfinancial corporates. Besides the main result, a noteworthy difference from the regressions of aggregate currency mismatch data is that in stock indices regressions (4) - (7), the coefficients of local currency bond portfolio investments become positively significant; higher local currency bond portfolio to GDP ratios are associated with more resilient stocks markets to global financial shocks. 231 To interpret the results, recall the results in the exchange rate regressions: currencies in EMEs having more local currency debts in the form of bond portfolio investment tend to be more sensitive to risk-on/off shocks. That is, a risk-off shock depreciates currencies of EMEs and the depreciation are larger for an EME if the EME was receiving more local currency bond portfolio investments. There are two mechanisms by which the currency depreciation caused by LC bond portfolio in- vestment outflows help the stock markets with becoming resilient from risk-on/off shocks. First, the depreciation discounts the stock prices in foreign currency so as to attract more in- vestors. Second, as I already discussed the results of the exchange rate regressions, the deprecia- tion might cause positive effects on the profitability of exporters in the EME, given other impacts through foreign currency debts, because the revenues from exports are fixed in the foreign currency (hence higher in the local currency), but much of the costs, for example wages, are fixed in local currency. As one can easily see, the second reasoning also explains why I do not have significant results for net foreign currency assets of nonfinancial corporate sectors. If the positively significant coeffi- cient captures the positive effects of local currency depreciations on the profits of the nonfinancial corporates, then it suggests that increases in the profits following local currency depreciations can offset the negative valuation effects of foreign currency debts. B.9 Additional Empirical Results For the tables of the full results of the cross-country panel regressions and more results as the robustness check, see the online appendix 232 C Appendix to Chapter 3 C.1 List of Countries and Omitted Figures and Tables C.1.1 List of Countries used in Regressions Argentina, Bolivia, Brazil, Bulgaria, Chile, Colombia, Croatia, Czech Republic, Guatemala, India, Indonesia, Kazakhstan, Korea, Malaysia, Mexico, Mongolia, Pakistan, Paraguay, Peru, Philip- pines, Poland, Romania, Russia, South Africa, Thailand, Turkey, Uruguay, Vietnam (28 EMEs) C.1.2 Omitted Figures Figure A.1 plots the change in the reserve outflows-to-GDP and extra capital inflows-to-GDP for all the sample countries during the period 1998-2017. As we discussed in Figure 7, for most sample countries, the two series show strong co-movement. Figure A.2 plots per capital GDP and the ratio of reserve asset to total external assets during the period 2013-2017. This Figure shows a negative association between the two variables. That is, the country with higher GDP per capita tends to have less international reserves of the total external assets. The relation might be explained by the following reasons. First, developed coun- tries generally have more efficient financial systems, which suggests their institutional quality is superior enough to generate private capital outflows. Second, in the course of growth, an EME can reduce the reliance on public capital outflows. Korea is a good example in this regard. In 2003, the ratio of reserves to the total external assets in Korea was 60.1%. But, it has decreased to about 40% recently. On the flip side, it means the portion of private external assets has increased as we discussed in section 2. This divergence between the private and public outflows may be caused by structural changes in private sector’s overseas investment such as improved access to foreign financial assets due to technological advances. 233 Figure C.1: Reserve Outflows and Extra Capital Inflows Note: All values are scaled by GDP. Source : IMF BOP/IIP C.1.3 Omitted Tables Here, we list the regression results discussed in section 2.2. A difficulty in our regression analysis is that our sample periods include both of the periods of reserve accumulation and the periods of reserve depletion. We want to exclude the periods of sudden stops such as the crises in Latin America in 2002 or other periods alike because we want to see how reserve “accumulation” is associated with different types of capital inflows; during a sudden stop, we should see reserve depletion with falling debt capital inflows and relatively stable FDI inflows, which might cause positive correlations of reserve outflows with debt inflows, but negative correlations with FDI inflows. 234 Figure C.2: External Asset Structure and GDP per capita Note: Y axis (RS TOT) is defined as the ratio of reserve assets to the total external assets. X axis (GDP per capita) is dollar value. All values are averaged over 2013-2017. Source : IMF BOP/IIP To handle the issue, we opt to run regressions in the sample period of 2003-07. This is to avoid the periods of the crises in the sample period 2930 . The results of the regressions are provided in Table A.1 below, with the results of the whole sample periods, 1998-2017. As we expected, almost all the coefficients of capital inflows are positive and significant regardless of our choice of sample periods. First, the coefficients of the equity portfolio are larger than the others and moreover the largenesses are statistically significant. Second, once we trim down the sample period to 2003-07, the coefficients of direct investment inflows become larger while the coefficients of debt inflows become smaller or insignificant. 29 The currency crises in Malaysia and Indonesia in 1999, the sovereign debt crisis in Argentina in 2002 that prop- agated to other Latin America countries, and obviously the Global Financial Crisis in 2008 along with its subsequent turbulent periods such as tapering tantrum in 2013. 30 Furthermore, 2003-07 was the time under the mood of the great moderation except for the subprime mortgage default in 2007, which was yet to propagate to emerging markets. 235 Table C.1: (a) The whole sample period (1998-2017) Reserve outflows-to-GDP V ARIABLES (1) (2) (3) FDI inflows 0.27*** (0.090) EQ inflows 0.82*** (0.186) DT inflows 0.31*** (0.083) Current account 0.26*** 0.16** 0.25*** (0.047) (0.060) (0.055) Constant 0.01* 0.01*** 0.01*** (0.003) (0.001) (0.001) Observations 140 140 140 R-squared 0.204 0.192 0.264 Number of Country 28 28 28 Driscoll and Kraay (1998) standard errors *** p<0.01, ** p<0.05, * p<0.1 C.2 Omitted Algebras and Proofs Proof of Lemma 1 First, we show the the object function of the social planner is concave in b 1 . Define V u c T 0 ; y N 0 +E 0 bu c T 1 ; y N 1 +b 2 u c T 2 ; y N 2 . Then taking the derivative of V with respect to b 1 gives us dV db 1 = −u T 0 +b Z f f 1+ r −2G j b 1 1− db 2 dw 1 u T 1 dF f +b 2 Z f f 1+ r −2G j b 1 db 2 dw 1 (1+ r −2G b b 2 ) u T 2 dF f (C.1) where w 1 = b 1 (1+ r 1 ) because we have no reserve accumulation yet. 236 (b) The sample period of 2003-2007 Reserve outflows-to-GDP V ARIABLES (1) (2) (3) FDI inflows 0.60*** (0.061) EQ inflows 1.33*** (0.111) DT inflows 0.25*** (0.036) Current account 0.42*** 0.31*** 0.31*** (0.032) (0.030) (0.046) Constant 0.02*** 0.04*** 0.04*** (0.003) (0.001) (0.001) Observations 140 140 140 R-squared 0.247 0.306 0.209 Number of Country 28 28 28 Driscoll and Kraay (1998) standard errors *** p<0.01, ** p<0.05, * p<0.1 Note: Panel Regressions with country and year fixed effects. All dependent variables are scaled by GDP. Source: IMF BOP/IIP Since the social planner can choose b 2 for f2 f;f c where f c is the cut-off for the social planner, which is different from (3.18), the equation (C.1) changes to dV db 1 = −u T 0 +b Z f f 1+ r −2G j b 1 u T 1 dF f + (C.2) b 2 Z f c f 1+ r −2G j b 1 db 2 dw 1 u T 1 +(1+ r −2G b b 2 )u T 2 dF f It is easy to see b R f f 1+ r −2G j b 1 u T 1 dF f is decreasing in b 1 since the return to the saving is decreasing (increasing borrowing rates in the borrowing). Hence, the first and second terms in the equation (C.3) decreases in b 1 . In the third term, from (3.15) we can check db 2 dw 1 < 0 and j db 2 dw 1 j is decreasing in b 1 . f c obviously decreases in b 1 . Since b 2 j f<f c decreases in b 1 , we can immediately see u T 1 (1+ r −2G b b 2 )u T 2 decreases in b 1 for all f2 f;f c . Therefore, all the terms are decreasing in b 1 . It implies d 2 V db 2 1 < 0. 237 Now we show the claims in the lemma. First, we show b h 1 < b sp 1 if b h 1 < 0. To show b priv 1 < b sp 1 , we first derive the first order condition of b 1 for the social planner. b sp 1 is characterized as below −u T 0 +b Z f f (1+ r −2G b b 1 ) 1− db 2 dw 1 u T 1 dF f b 2 Z f f (1+ r −2G b b 2 ) db 2 dw 1 (1+ r −2G b b 1 ) u T 2 dF f =0 (C.3) where w 1 = b 1 (1+ r 1 )+ R 1 . This can be represented by −u T 0 +b Z f f (1+ r 1 )u T 1 −G b b 1 u T 1 dF f −(1+ r −2G b b 1 )b Z f f db 2 dw 1 u T 1 b(1+ r 2 −G b b 2 )u T 2 dF f = 0 (C.4) Suppose b 1 = b priv 1 and b 2 = b priv 2 . Then we have u T 0 =b R f f (1+ r 1 )u T 1 dF f and also u T 1 = b(1+ r 2 )u T 2 i f f >f c . This gives us −b Z f f G b b 1 u T 1 +b 2 (1+ r −2G b b 1 ) Z f f db 2 dw 1 G b b 2 u T 2 dF f −b(1+ r −2G b b 1 ) Z f c f db 2 dw 1 u T 1 −b(1+ r 2 )u T 2 dF f > 0 (C.5) If f2[f;f c ), then u T 1 −b(1+ r 2 )u T 2 > 0: Absolutely the remaining terms are all positive. It implies that b h 1 0, then there always existsg 0 2(0;¥] such that forG s 2(0;g 0 ) b h t ¡b sp t . Ifg 0 2(0;¥), then there existsg 1 2(g 0 ;¥) such that forG s 2(g 0 ;g 1 ) b h t ¿b sp t . Since b h 1 > 0;we replaceG b withG s in front of b 1 in equation (C.2). This gives us 238 dV db 1 = −b Z f f G s b 1 u T 1 +b 2 (1+ r −2G s b 1 ) Z f f db 2 dw 1 G b b 2 u T 2 dF f −b(1+ r −2G s b 1 ) Z f c f db 2 dw 1 u T 1 −b(1+ r 2 )u T 2 dF f (C.6) First, notices dV db 1 is continuous inG s as long as b 1 > 0. All the variables are continuous and the mappings in dV db 1 are continuous as well. Then we show lim G s #0 dV db 1 > 0 and lim G s "¥ dV db 1 > 0 These properties come from lim G s #0 G s b 1 = lim G s "¥ G s b 1 = 0: lim G s #0 G s b 1 = 0 is obvious. lim G s "¥ G s b 1 = 0 can be easily shown by a contradiction. If lim G s "¥ G s b 1 6= 0, then the gross return must be zero. Then the households can be better off by letting b 1 = 0. IfG s b 1 = 0, then the equation (C.6) will be dV db 1 =b 2 (1+ r ) Z f f db 2 dw 1 G b b 2 u T 2 dF f −b(1+ r ) Z f c f db 2 dw 1 u T 1 −b(1+ r 2 )u T 2 dF f > 0 (C.7) It proves there always existsg 0 2(0;¥] such that forG s 2(0;g 0 ) b h t ¡b sp t since dV db 1 is continuous in G s . Finally, we show ifg 0 2(0;¥), then there existsg 1 2(g 0 ;¥) such that forG s 2(g 0 ;g 1 ) b h t ¿b sp t . To show it, it is enough to show we may have dV db 1 < 0 for someG s by again the continuity of dV db 1 in G s . In equation (C.6), we can manipulate the distribution off and relative values of y T 1 and y T s so thatf c !f andb 2 is small enougn to haveb R f f G s b 1 u T 1 >b 2 (1+ r −2G s b 1 ) R f f db 2 dw 1 G b b 2 u T 2 dF f . This holds as long asG s b 1 > 0, and we showed dV db 1 < 0 for someG s . This completes the proof. 239 Proof of Proposition 4 In the same way we did in the proof of lemma 1, we let b 1 and b 2 as functions of R 1 . Taking a derivative with respect to R 1 gives us dV db 1 = u T 0 +b(1+ r)E u T 1 + db 1 dR 1 u T 0 1+ r −2G j b 1 b Z f f u T 1 dF f ! +b db 1 dR 1 1+ r −2G j b 1 +(1+ r) Z f f db 2 dw 1 u T 1 b(1+ r −2G b b 2 )u T 2 dF f + df c dR 1 bu c T 1 b 2 f c+ +b 2 u c T 2 b 2 f c+ bu c T 1 b 2 f c +b 2 u c T 2 b 2 f c We know bu c T 1 (b 2 (f c+ )) +b 2 u c T 2 (b 2 (f c+ )) =bu c T 1 (b 2 (f c )) +b 2 u c T 2 (b 2 (f c )) . Also we know u T 0 b(1+ r 1 )E u T 1 = 0 and u T 1 b(1+ r 2 )E u T 2 = 0 forf2 f c ;f . Taking all of these and letting dw 1 dR 1 = db 1 dR 1 1+ r −2G j b 1 +(1+ r) gives the equation (3.21), (3.22). Derivation of the Optimal Tax First, we show the borrowing of decentralized households under the optimal taxation is same as the direct choice of the social planner. The Euler equation of the households under the taxation will be u T 0 (b 1 ;;)=b(1+ r 1 )(1+t d ) Z f c f u T 1 (b 1 ;b 2;c ;;)dF f + Z f f c u T 1 (b 1 ;b 2;u ;;)dF f ! (C.8) The solution of the planning ist d such that the equation (C.8) is ex-post identical to the equation (C.6), which characterizes the borrowing determined by the government. Of course, it implies b priv 1 (t d )= b sp 1 : Ignoring the term r 1 t d , solving fort d such that the equation (C.8) is same as the equation (C.6) yields the characterization of the optimal taxation in the equation (C.8). Proof of Proposition 5 The first statement in the proposition follows from the definition of the optimal tax. 240 To prove the second and third statements, rewriting the first order conditions in the equations (3.24) and (3.25) are as follows u T 0 +bu T 1 1+ r 2G s b sp 1 +b Z f c f u T 1 +bu T 2 (1+ r 2 ) db 2 dw 1 1+ r 2G s b sp 1 = 0 (C.9) u T 0 +bu T 1 (1+ r)+b Z f c f u T 1 +bu T 2 (1+ r 2 ) db 2 dw 1 (1+ r)+h = 0 (C.10) whereh is the multiplier of the non-zero constraint of reserve accumulation. If b sp 1 < 0, then 1+ r 2G s b sp 1 > 1+ r. It impliesh > 0. Hence, if If b sp t+1 < 0 and b sp 1 = b t+1 then R = 0 If b sp 1 > 0, then we may have 1+ r 2G s b sp 1 = 1+ r. It implies b sp 1 = r r 2G s Then the optimal reserve accumulation is R 1 = w 1 r r 2G s where w 1 is the optimal net foreign liquid assets chosen by the social planner. The social planner problem of choosing b 1 and R 1 can be understood the two steps where the planner firstly chooses w 1 and then choose how to compose w 1 with b 1 and R 1 . As one might expect, for b 1 < r r 2G s b 1 is always more efficient. For b 1 > r r 2G s vice versa. Proof of Corollary 2 We formulate the problem to the problem the social planner determine the optimal w 1 and then compare b 1 and R 1 . For the planner who only use the optimal taxation b sp 1 1+ r G j b sp 1 = w 1 (C.11) 241 For the planner who only use the reserve accumulation b h 1 1+ r G j b h 1 + R 1 = w 1 (C.12) where j= b;s: Assume that w 1 < 0. Then obviously b 1 < 0 in the both of equations (C.11) and (C.12), and b h 1 r, it is always better to reduce R 1 and raise (reduce) b h 1 b h 1 . It implies V(t)> V(R) if b sp 1 < 0. Next now assume w 1 > 0. Let’s compute the return to the two different policies. For the optimal taxation policy in equation (C.11) r 1 = 1+ r G s b sp 1 Similarly, we can compute the return to the EME in equation (17). Denote the return byb r 1 b r 1 = R 1 =(1+ r) R 1 =(1+ r)+ b h 1 r+ b h 1 R 1 =(1+ r)+ b h 1 ; r G s b h 1 For the same w 1 , b h 1 r 1 . This completes the proof. Proof of Proposition 7 To prove the second statement we make a reasonable assumption Assumption A1 . Let the LHS of equation (3.29) be h L (b 1 (q;R(q));q;R(q)) and the the RHS be h R (b 1 (q;R(q));q;R(q)). Then we assume 1.j dh L dq j R =j ¶h L ¶q + ¶h L ¶b 1 ¶b 1 ¶q j>j dh R dq j R =j ¶h R ¶q + ¶h R ¶b 1 ¶b 1 ¶q j 2. ¶h L ¶b 1 ¶b 1 ¶R 1 + ¶h L ¶R 1 > 0 To see why the assumption A1 can easily hold, first see dh L dq , dh R dq < 0. Then b 1 changes the LHS directly, but affetct RHS through b 2 ;which makes the RHS less responsive to b 1 : The second in- equality easily holds for most of the parameter values including some extreme values. Now we prove the statements in the proposition. 242 To show the first statement, recall that we have dh L dq ; dh R dq < 0. Withq = ˜ q, h L (b 1 ;q)= 0, while h R > 0 since b 2 < 0 for allf. It implies that atq = ˜ q R 1 > 0. Since ¶b 1 ¶q > 0, R 1 > 0 for allq > ˜ q. Then we need to show there existsd > 0 such that R 1 > 0q2 ˜ qd; ˜ q . See both h L and h R are continuous in q and R 1 . For d small enough, h L b 1 ˜ qd;0 ; ˜ qd;0 is positive, but smaller than h R by the continuity of h L and h R . It implies R 1 > 0q2 ˜ qd; ˜ q . Second statement can easily follow from the assumption. Envoking the envelope condition yields ¶h L ¶q + ¶h L ¶b 1 ¶b 1 dq + ¶b 1 dR dR 1 dq + ¶h L ¶R 1 dR 1 dq = ¶h R ¶q + ¶h R ¶b 1 ¶b 1 dq + ¶b 1 ¶R dR 1 dq + ¶h R ¶R 1 dR 1 dq Then we must have dR 1 dq > 0 under the assumption A1. Lastly, we prove the third and forth statements in the proposition. The resource constraints in tradable goods are c T 0 (q;R 1 )+ b 1 (q;R 1 )+ R 1 =(1q)y T 0 + Q 0 qK c T 0 ˜ q;0 + b 1 ˜ q;0 = 1 ˜ q y T 0 + Q 0 ˜ qK Since(1q)y T 0 + Q 0 qK= y T 0 +(Q 0 A 0 )qK, we have R=(Q 0 A 0 ) q ˜ q K+ c T 0 ˜ q;0 c T 0 (q;R 1 )+ b 1 ˜ q;0 b 1 (q;R 1 ) We know, by definition, r −rG s b 1 (q;R 1 ) 1 db 1 (q;R 1 ) dR 1 > 0 and r −rG s b 1 ˜ q;0 1 db 1( ˜ q;0) dR 1 = 0. It implies b 1 ˜ q;0 > b 1 (q;R 1 ) To show the last statement, notice that ¶c T 0 ¶q > 0, ¶c T 0 ¶R 1 < 0 and c T 0 ˜ q;R 1 ˜ q < c T 0 (q;0). Thus we have c T 0 ˜ q;R 1 ˜ q < c T 0 ˜ q;0 . If ¶c T 0 ¶q + ¶c T 0 ¶R 1 dR 1 dq < 0, then it is obvious. Even if ¶c T 0 ¶q + ¶c T 0 ¶R 1 dR 1 dq > 0, there always existsd > 0 such that c T 0 ˜ q+d;R 1 ˜ q+d < c T 0 (q;0). This completes the proof. 243 Proof of Proposition 8 Recall the equation (3.43) bG b b 0 E u 0 T db 0 dR 0 + u T b(1+ r)E u 0 T + u T b 1+ r 0 E u 0 T db 0 dR 0 = dw 0 dR 0 Pr f 0 f 0c bE f 0 f 0c ¶b 00 c ¶w 0 u 0 T +b ¶V 00 ¶b 00 c + Pr f 0 >f 0c bE f 0 >f 0c ¶b 00 u ¶w 0 u 0 T +b ¶V 00 ¶b 00 u Let Pr(f 0 f 0c )bE f 0 f 0c h ¶b 00 c ¶w 0 u 0 T +b ¶V 00 ¶b 00 c i Q c and Pr(f 0 >f 0c )bE f 0 >f 0c h ¶b 00 u ¶w 0 u 0 T +b ¶V 00 ¶b 00 u i Q u . It is easy to see both ofQ c andQ u decreases in w 1 . To show the first statement, notice that for a lowerf,j db 0 dR 0 j becomes larger and hence smaller dw 0 dR 0 . Also u T b(1+ r)E[u 0 T ] and u T b(1+ r 0 )E[u 0 T ] become larger as well. WhetherG b b 0 E[u 0 T ] db 0 dR 0 increases by the assumption. Given w 1 ,Q c andQ u are invariant, and w 1 increases for smallerf, which makes the falls inQ c andQ u . Therefore, for smallerf <f c , the marginal costs are higher, but the marginal benefits are lower. Hence, the optimal plan is to accumulate less reserves (deplete more reserves). That is, ¶R 0 ¶f j f<f c> 0. Next, we prove the second statement. Suppose that there happens a left shift of the pdf off 0 such that ˜ Pr(f 0 < a)< Pr(f 0 < a) for all a2(0;1) where ˜ Prob is the cdf of the new distribution. Then obviously, the RHS increases in the new distribution. In the LHS, onlyE[u 0 T ] rises due to expected tighter credit constraint. Since we assumeb(1+ r)>b(1+ r 2G b b 0 ) db 0 dR 0 , the LHS falls. To restore the balance between the LHS and RHS, we need higher R 1 . This is as desired. Proof of Lemma 3 Suppose the credit constraint does not bind in the state w t . Then b h t+1 (w t ) is determined by the euler equation u T t b h t+1 =b(1+ r t+1 )E u T t+1 b h t+1 ;T t+1 : More formally, b h t+1 is a functinonal of the fuction T t+1 (w t+1 ). We know, in general V(t) b t+1 6= 0, and of coure it is not in the optimality conditions of the social planner. 244 C.3 Other Mircofoundations of Frictions on Capital Outflows We introduce results with different microfoundations of frictions on private capital outflows. We consider the case where overseas investments are intermediated by domestic financial experts, and the case the intermediation is done by the domestic financial experts but the experts conceal parts of their income from the intermediation fees. Now we assume that there is a continuum of domestic financial experts and they face heteroge- neous participation costs, same as the IFIs in section 3. For the households who are not financial experts, the decision is same as the model in section 3. The difference is now the rents 1 2 G s b 2 1 back to the households since the financial experts belong to the family of households. Without going through all the necessary steps, we introduce the condition for the optimal reserve accumulation. u T 0 b(1+ r)E u T 1 | {z } ConsumptionWedge at r = dw 1 dR 1 b 2 6 6 6 4 Z f c f d(−b 2 ) dw 1 u T 1 b(1+ r 2 )u T 2 dF f | {z } Marginal Value o f Borrowing −bG b E u T 2 b 2 db 2 dw 1 | {z } Lower r 2 3 7 7 7 5 (C.13) The main difference is the term −bG s E u T 1 b 1 db 1 dR 1 vanishes since the rents from the intermediation belong to the households. Also notice that the saving b h 1 by the decentralized households is always short of the optimum for the planner because the households do not take account of the negative externalities from the insufficient wealth in the next period; the terms in the RHS in the equation (C.13), but the term of negative externality of additional savingbG s E u T 1 b 1 disappears; the de- creasing return to the saving is mere a result from increasing marginal costs. Therefore, in the new environment where the intermediation is done by domestic financial experts who face increasing cost of the intermediation, the amount of reserve accumulation, assuming all the states are equal, is strictly smaller than when the intermediation is done by foreign parties. 245 Additionally let’s assume that the financial experts conceal parts of the incomes of the rents; i.e., they conceal parts of the profits of the intermediation 1 2 G s b 2 1 ; denote it asp. For the simplicity, we consider the case all the rents 1 2 G s b 2 1 are concealed and the experts invest the concealed incomes in the assets that bear no interest. The experts conceal the rents in the period 1 and consume the rents in the period 2 regardless of the states in the period 1. Then the concealment results in “inefficient resource allocations across time”; it reduces the tradable goods resources in the period 1 by 1 2 G s b 2 1 while it raises the resources in the period 2 by the same amount; it is inefficient because the economy in the period 1 needs to borrow against the outputs in the period 2. Again, we just introduce the optimality condition of reserve accumulation as below. u T 0 b(1+ r)E u T 1 | {z } ConsumptionWedge at r =b Z f c f dw 1 dR 1 d(−b 2 ) dw 1 u T 1 b(1+ r 2 )u T 2 dF f | {z } Marginal Value o f Borrowing −b 2 G b E u T 2 b 2 ¶b 2 ¶w 1 dw 1 dR 1 + ¶b 2 ¶p dp dR 1 | {z } Lower r 2 bG s b 1 db s 1 dR 1 E u T 1 bE u T 2 | {z } Loss due to concealment (C.14) where db 2 dR 1 = ¶b 2 ¶w 1 dw 1 dR 1 + ¶b 2 ¶p dp dR 1 . SeeE u T 1 bE u T 2 > 0 sinceE u T 1 b(1+ r 2 )E u T 2 0. The concealment makes two changes to the optimality condition in the baseline model;bG b E h u T 2 b 2 ¶b 2 ¶p dp dR 1 i andbG s b 1 db s 1 dR 1 E u T 1 bE u T 2 . First, the concealment generates inefficient income streams and it is captured bybG b E h u T 2 b 2 ¶b 2 ¶p dp dR 1 i . The economy needs more resources in the period 1 to lessen the costs of sudden stop and reduce the inefficient borrowing from IFIs, but the concealment worsen the problem; it adds some resources to the period 2 whereas it reduces equivalent resources in pe- riods 1 and thus the households need to borrow even more. As a result, the reserve accumulation reduces the costs of the concealment by replacing the private overseas investments. Second, since the concealed assets are invested inefficiently, the replacement of private overseas investments by the reserve accumulation is beneficial as it is shown in the term bG s b 1 db s 1 dR 1 E u T 1 bE u T 2 . 246 Consequently, introducing overseas concealment by domestic financial experts worsen the prob- lems of overseas investments by private sectors and accordingly raises the benefits of reserve ac- cumulation, which substitutes for the undesirable investments by households. The optimal reserve accumulation in the equation (C.14) must be higher than the equation (C.13) but it is hard to predict that the reserve accumulation is higher or lower than baseline model in section 3.3. Our conjec- ture is that it might be still lower than the the baseline model. However, introducing cases where households conceal their assets for themselves or other costs of the concealment undo the lower reserve accumulation. C.4 Reserve Accumulation of Saving Glut EMEs In this section in the appendix, we remove one assumption that we have kept throughout this paper. Now we assume that our model EME saves even in period 1. We can think of an EME experiencing rapid aging or saving excessively due to some reasons 31 . In the data, some EMEs, in particular East Asian EMEs, have large amounts of positive net foreign assets. We show how our model can explain reserve accumulation in such EMEs. To model EMEs that save in period 1, we assume y T 0 >(1q)y T 1 >(1q)(1s)y T 2 . Hence, we can posit b 1 ;b 2 > 0 and therefore there is no chance of sudden stop in period 1 32 . Interestingly, the planner in the EME might be incentivized to accumulate reserves even when b 2 > 0. The optimal reserve accumulation in period 1 is characterized by r −rG s b 2 1 db 2 dR 2 = 0 (C.15) The equation (C.15) does not necessarily hold, but it holds if R 2 > 0. On the contrary, if equation (C.15) does not hold, we have R 2 = 0. As one might expect, db 2 dR 2 is almost invariant to R 2 . Then, conditioning on R 2 > 0, the equation (A.20) almost pins down b 2 . The intuition behind this results 31 For example, low financial development in certain EMEs lead the EMEs to save excessively. See Caballero et al. (2008), Mendoza et al. (2009) and Maggiori (2017). 32 It might be little extreme and such results stem from our modeling of sudden stop 247 is rather straightforward. The return to private overseas investments (for the planner, not for the households) decrease in the investment amounts and it implies that the return to the investment beyond a certain level is even below the return to the reserve. Hence, the planner chooses reserve accumulation by a mean of national saving. Now we characterize the reserve accumulation in period 0, while accommodating the results in equation (C.15). The optimal reserve accumulation is characterized by the equation (C.16) below. r −rG s b 1 (q) 1 ¶b 1 (q) ¶R 1 + ¶b 1 (q) ¶R 2 dR 2 dR 1 E u T 1 = − dw 1 dR 1 bG s E u T 2 b 2 db 2 dw 1 (C.16) See the RHS is negative and as we saw in equation (C.15), b 2 is almost fixed once we have R 2 > 0. Once the value of RHS is highly invariant, then we need to have invariant LHS as well: the RHS pins down b 1 in the LHS; let’s denote it as b c 1 . It implies that for EMEs to need external assets more than b c 1 , the extra demands for external assets beyond b c 1 is absorbed by reserve accumula- tion. To understand it, think of an EME that has current account surpluses and receive lots of direct investments. The EME needs to accumulate external assets, but accumulating external assets by private sectors accompanies nonnegligible inefficiencies. As we have repeated throughout this pa- per, in such a case the planner accumulate reserve to replace the inefficient overseas investment by private sectors. Hence, passive capital inflows (direct investments or equity portfolio investments) to such EMEs generate seemingly excessive reserve accumulation. Although we cannot make a strong claim because our model is not suitable for quantitative analysis, the result in this section potentially explain why some East Asian EMEs with high net positive foreign asset positions, for example Malaysia, Thailand, and China, hold high levels of international reserves from 30 to 40 percent of GDP. 248 C.5 Capital Outflow Restrictions in EMEs Compared to capital inflow restrictions, the discussion on capital outflow restriction in EMEs is scarce. However, in reality, the restrictions on capital outflow in EMEs are stricter than capital in- flow. To show how capital outflow restrictions have changed, we exploit capital control measures constructed by Fernandez et al. (2016). Figure C.3 (a) presents the evolution of capital control restrictions on both inflows and outflows for our sample countries. The two series show strong co-movement over time. Remarkably, EMEs have maintained higher levels of capital outflow re- strictions than capital inflow. This suggests that residents in EMEs would have more difficulty investing abroad. Fernandez et al. (2016) also note that this pattern of strict outflow restrictions is more noticeable for EMEs than advanced economies. Figure C.3 (b) shows how capital outflow restrictions in EMEs evolve by major asset classes. The outflow controls on money market instru- ments, which include bonds with a maturity of one year or less, have maintained a relatively high level. As an example, we will illustrate the evolution of the regulations on capital flows in Korea. The capital outflow indicator shows that Korea imposed relatively strict restrictions on capital outflows by 2004. 33 While there were some relaxations on capital control in the 1980s, outward investment by residents in Korea was strictly controlled by the government. For instance, domestic residents were not permitted to purchase foreign securities. Starting in 1995, residents were allowed to freely purchase and hold foreign currency only up to USD 10 thousand per year per individual. For outward FDI, there were a list of investment areas that had to be reviewed by the government and the quantification conditions such as self-financing. After Korea’s accession to the OECD, from the late 1990s, regulations on capital transactions were gradually reduced. To deal with a continued trade surplus and support the competitiveness of Korean firms, the government encour- aged domestic firms and residents to invest abroad. For example, the investment restriction per 33 From 1998 to 2004, the overall capital outflow control index is 0.64 on average. In 2005 it decreased to 0.2. Then, the average index from 2006 to 2017 is 0.11. For an inflow index, it shows a similar trend to the outflow index. However, as discussed, the average inflow index is 0.51 before 2004, which shows that outflow restrictions were stricter then. 249 Figure C.3: Capital Outflow Restrictions (a) Inflows and Outflows Restrictions (b) Outflows Restrictions (by asset class) Note: A higher value means strict restrictions. All values are averaged across the sample countries. Source: Capital Control Measures: A New Dataset, Fernandez et al. (2016) project was abolished and the maximum investment amount by individual investors raised to USD 10 million in 2005. 250

Abstract (if available)

Abstract In this dissertation, I study the changes in external liabilities and assets of emerging market economies since the early-2000s and the following implications of the changes for the financial stability and the optimal policies of the economies. In chapter 1, I construct a dataset, which measures the external liability composition of emerging market economies in different instruments and currencies. The new dataset shows emerging market economies have much lower currency exposures than in the past. Also, the observed pattern in the dataset suggests that the ever-increasing local currency external borrowings of the emerging market economies since the early 2000s, original sin dissipation, is related to the capital market development in emerging market economies. Chapter 2 is a study of channels through which risk-appetite shocks to global investors, i.e., global financial shocks, are transmitted to emerging market economies. First, I empirically show that much of the transmission of global financial shocks to emerging market economies is reflected in equity and local currency bond portfolio investment capital flows. I then develop a small open economy model which, augmented with leverage constrained banks and foreign investors who purchase equities and bonds, can replicate these empirical findings qualitatively. Quantitative analysis of the model suggests that global financial shocks can account for 50% of the equity price volatility and 30% of the investment volatility in Korea, in which most of the external liabilities of the country are Korean won-denominated equities and debts. In short, all the analysis in chapter 2 implies that to a substantial extent, risk-appetite shocks to global investors are transmitted to emerging market economies via fickle portfolio capital flows to equity and local currency bond markets in the economies. In chapter 3, Dongwook Kim and I provide a novel theory of international reserve accumulation of emerging market economies. We view reserve accumulation as capital outflows by the public sector which supplements insufficient capital outflows by the private sector. In our model, when an emerging market economy receives large capital inflows in the form of direct or equity portfolio investment, the emerging market economy must invest abroad to maintain macroeconomic balance and prepare for a possible future sudden stop. If the private sector in the emerging market economy cannot invest externally sufficiently or invests inefficiently due to low financial expertise or poor institutional quality, supplemental international investments must be accomplished by the public sector as international reserve outflows.

Linked assets

University of Southern California Dissertations and Theses

Conceptually similar

PDF

Essays on monetary policy and international spillovers

PDF

Essays in macroeconomics and macro-finance

PDF

Essays in macroeconomics

PDF

Early-warning systems for crisis risk

PDF

Essay on monetary policy, macroprudential policy, and financial integration

PDF

Innovation: financial and economics considerations

PDF

Two essays on major macroeconomic shocks in the Japanese economy: demographic shocks and financial shocks

PDF

Growth and development in Africa: macroeconomic and microeconomic perspectives

PDF

The murky risk of trade protectionism in an interconnected and uncertain global economy: a state, industrial, and regional analysis

PDF

Essays on interest rate determination in open economies

PDF

Nationalisms in the era of global quality TV: how SVODs main/stream the local

PDF

Essays on macroeconomics and income distribution

PDF

Essays on business cycle volatility and global trade

PDF

Essays in corporate finance

Asset Metadata

Creator Han, Bada (author)

Core Title Three essays in international macroeconomics and finance

School College of Letters, Arts and Sciences

Degree Doctor of Philosophy

Degree Program Economics

Degree Conferral Date 2021-05

Publication Date 05/10/2021

Defense Date 03/14/2021

Publisher University of Southern California (original), University of Southern California. Libraries (digital)

Tag capital flows,currency exposure,global financial cycle,international reserve,local currency debt,OAI-PMH Harvest,original sin,sudden stop

Format theses (aat)

Language English

Contributor Electronically uploaded by the author (provenance)

Advisor Aizenman, Joshua (committee chair), Betts, Caroline Marie (committee member), Li, Wenhao (committee member), Ranciere, Romain (committee member)

Creator Email badahan@usc.edu

Permanent Link (DOI) https://doi.org/10.25549/usctheses-oUC112720141

Unique identifier UC112720141

Identifier etd-HanBada-9577.pdf (filename)

Legacy Identifier etd-HanBada-9577

Document Type Dissertation

Format theses (aat)

Rights Han, Bada

Internet Media Type application/pdf

Type texts

Source 20210512-wayne-usctheses-batch-837-shoaf (batch), University of Southern California (contributing entity), University of Southern California Dissertations and Theses (collection)

Access Conditions The author retains rights to his/her dissertation, thesis or other graduate work according to U.S. copyright law. Electronic access is being provided by the USC Libraries in agreement with the author, as the original true and official version of the work, but does not grant the reader permission to use the work if the desired use is covered by copyright. It is the author, as rights holder, who must provide use permission if such use is covered by copyright. The original signature page accompanying the original submission of the work to the USC Libraries is retained by the USC Libraries and a copy of it may be obtained by authorized requesters contacting the repository e-mail address given.

Repository Name University of Southern California Digital Library

Repository Location USC Digital Library, University of Southern California, University Park Campus MC 2810, 3434 South Grand Avenue, 2nd Floor, Los Angeles, California 90089-2810, USA

Repository Email cisadmin@lib.usc.edu