University of Southern California Dissertations and Theses

Implied equity duration: international evidence

(USC Thesis Other)

Implied equity duration: international evidence

PDF

Download a page range

Download transcript

Copy asset link

Request this asset

Transcript (if available)

Content

IMPLIED EQUITY DURATION: INTERNATIONAL EVIDENCE

by

Jesse Gardner

A Dissertation Presented to the
FACULTY OF THE USC GRADUATE SCHOOL
UNIVERSITY OF SOUTHERN CALIFORNIA
In Partial Fulfillment of the
Requirements for the Degree
DOCTOR OF PHILOSOPHY
(BUSINESS ADMINISTRATION)

August 2023

Copyright 2023 Jesse Gardner
ii

Dedication

To my wife, Yi; my son, Jesse; and my parents, Jesse and Carol. I could not have completed this
without you.

iii

Acknowledgements

I thank the members of my dissertation committee: Patricia Dechow (chair), Richard Sloan, and
Lukas Schmid for many helpful discussions and suggestions for this dissertation. I also thank
Katherine Bruere, Anthony Bucaro, Melissa Carlisle, Jonathan Craske, Marc Cussatt, Mark
DeFond, Ryan Erhard, Timothy Fogarty, Mary Margaret Frank, Justin Hopkins, Murali
Jagannathan, Kathryn Kisska-Schulze, Jonathan Kugel, Clive Lennox, Shelley Li, Luann Lynch,
Babak Mammadov, Nicholas Moellman, James Naughton, Maria Ogneva, Vivek Pandey, Gary
Previts, Jonathan Ross, Steven Schwartz, Matthew Shaffer, Paul Simko, Mark Soliman, Lorien
Stice-Lawrence, KR Subramanyam, Danko Tarabar, Courtney Yazzie, Joon Yoon, and workshop
participants at Binghamton University, Case Western Reserve University, Christopher Newport
University, Clemson University, University of Southern California, University of Virginia, and
Winthrop University their for constructive feedback. I thank my Ph.D. colleagues and the USC
Leventhal School of Accounting staff for constant support and encouragement. I gratefully
acknowledge generous financial support from the USC Marshall School of Business.

iv

Table of Contents

Dedication ...................................................................................................................................... ii

Acknowledgements ....................................................................................................................... iii

List of Tables ................................................................................................................................ vi

List of Figures ............................................................................................................................. viii

Abstract ......................................................................................................................................... ix

Chapter 1: Introduction ...................................................................................................................1

Chapter 2: Background ..................................................................................................................10
2.1 Measuring Equity Duration..........................................................................................10
2.2 Alternative Measures of Equity Duration ....................................................................13
2.3 Country-Level Governance ..........................................................................................16

Chapter 3: Countries Examined and Prediction Development ......................................................17
3.1 Countries ......................................................................................................................17
3.2 Predictions....................................................................................................................18

Chapter 4: Sample and Descriptive Statistics ................................................................................24
4.1 Data ..............................................................................................................................24
4.2 Descriptive Statistics ....................................................................................................25

Chapter 5: Empirical Results .........................................................................................................27
5.1 Visual Preview of Results ............................................................................................27
5.2 Governance (WGI) and Country-Year Duration .........................................................28
5.3 Future Focused Tax Policy and Equity Duration .........................................................30
5.3.1 Tax Loss Carryforwards and Country-Year Duration ..................................30
5.3.2 Innovation Boxes and Country-Year Duration .............................................33

v

Chapter 6: Implications for Asset Management ............................................................................35
6.1 Short Duration Premium ..............................................................................................35
6.1.1 Short Duration Premium within Developed and Emerging Markets ............36
6.1.2 Short Duration Premium within Countries ...................................................37
6.2 Determinants of the Short Duration Premium .............................................................38
6.3 Short Duration Premium Across Countries .................................................................39

Chapter 7: Robustness Tests of Alternative Discount Rates and Forecasting
Assumptions ...................................................................................................................................42

Chapter 8: Additional Analysis ......................................................................................................46
8.1 Components of the World Governance Indicators .......................................................46
8.2 Economic Policy Uncertainty Index ............................................................................48
8.3 Hofstede's National Cultural Dimensions ....................................................................50
8.4 GDP Per Capita ............................................................................................................55

Chapter 9: Conclusion....................................................................................................................56

References ......................................................................................................................................60

Appendix A ....................................................................................................................................65
Variable Definitions ...........................................................................................................65

vi

List of Tables

Table 1. List of 49 Countries in Sample by Development Status and Geographic Region ...........79

Table 2. Equity Duration and Forecasting Assumptions Necessary to Calculate Duration
(Countries Sorted by Duration) ......................................................................................................80

Table 3. Summary Statistics for Country-Year Sample.................................................................82

Table 4. Sample Average World Governance Indicators (Countries Sorted by Composite WGI
Measure) ........................................................................................................................................83

Table 5. Country-Level Governance and Country-Level Duration ...............................................85

Table 6. Statutory Tax Loss Carryback Periods by Country-Year (Countries Sorted by Sample
Average) .........................................................................................................................................86

Table 7. Tax Loss Carryforward Policy and Country-Level Duration ..........................................88

Table 8. Availability of Innovation Box Tax Policies by Country-Year (Countries Sorted by
Total Years Available) ...................................................................................................................89

Table 9. Innovation Box Tax Policy and Country-Level Duration ...............................................91

Table 10. International Duration Portfolio Characteristics ............................................................92

Table 11. Within Country Duration Portfolio Characteristics .......................................................95

Table 12. Determinants of the Short Duration Premium at the Country-Level .............................99

Table 13. Equity Duration Portfolios of Country Indices ............................................................101

Table 14. Alternative Discount Rates and Forecasting Assumptions ..........................................102

vii

Table 15. Country-Level WGI Components and Country-Level Duration .................................103

Table 16. Country-Level Economy Policy Uncertainty and Country-Level Duration ................106

Table 17. Hofstede's Country-Level Cultural Dimensions and Country-Level Duration ...........107

Table 18. Country-Level GDP per Capita and Country-Level Duration .....................................110

viii

List of Figures

Figure 1. Equity Duration Around the World ................................................................................71

Figure 2. Time Series of Country-Year Duration for Developed, Emerging, and ACWI .............72

Figure 3. World Governance Indicators Around the World ..........................................................73

Figure 4. Tax Loss Carryforwards Around the World ...................................................................74

Figure 5. Short Duration Premium by Country (Equal Weighted Portfolios) ...............................75

Figure 6. Short Duration Premium by Country (Value Weighted Portfolios) ...............................76

Figure 7. Country-Year Duration Portfolios During a Short Term Cash Flow Shock ..................77

Figure 8. Country-Year Duration Portfolios During Global Interest Rate Hikes ..........................78

ix

Abstract

I hypothesize that short and long duration firms are more prevalent in certain economies. Short
duration firms return cash flows to investors more quickly than do long duration firms. Therefore,
I predict the predominance of long duration firms where the risk to investors' capital is lower. I
study a sample of 23 developed and 26 emerging countries and find that, in equilibrium, economies
have more long duration firms where: financial markets are more developed, inflation is lower, the
government is stable and promotes private sector development, agents have confidence in contract
enforcement and property rights, and corruption is controlled. I also predict and find that countries
with tax policies that are more future-focused (e.g., longer tax loss carryforward periods) have
longer duration firms on average. I next examine whether the duration premium (i.e., the negative
relation between returns and equity duration) is observable within and across countries. The results
indicate that the duration premium is observable in most countries and that a country's duration
premium is correlated with the range of duration in that country. Overall, these results highlight
that duration is a relevant risk factor both within and across countries and that a country's
institutional setting and level of development can impact corporate duration.
1

Chapter 1

Introduction

Equity duration is a measure used to capture the average maturity (in years) of cash flows
to shareholders associated with a given stock. These cash flows take the form of dividends, stock
repurchases, and stock issuances. A short duration equity is one where most of its market value is
based on the forecasted expectation of a return of short-term cash flows to shareholders. This
measure has gained popularity in recent years with academics and practitioners.
1
However, almost
all work in this area has focused on U.S. markets.
2
I contribute to the literature by examining the
role of implied equity duration in a global setting. I study both the cross-sectional country
characteristics that are associated with equity duration, and I test for the existence and magnitude
of the short duration premium within and across countries.

1
Recent academic examples include Weber (2018), Ozdagli (2018), Dechow, Erhard, Sloan, and Soliman (2021),
Goncalves (2021), Chen (2022), Montagna and Bianchi (2022), and Gormsen and Lazarus (2023). From a practitioner
perspective, approximately 40 articles have been posted to seekingalpha.com in 2022 that mention equity duration,
while only approximately 20 and 5 such articles were posted in 2021 and 2020 respectively. There are minimal
mentions of the measure prior to 2020.
2
Gormsen and Lazarus (2022) and Kleintop (2022) are some exceptions.
2

First, I predict and find the predominance of short duration firms in countries where the
risk to investors' capital is higher. Using a sample of 23 developed and 26 emerging countries, I
study country-level market and governance characteristics that are associated with long equity
duration firms headquartered in a given country. Managing a long duration company requires the
ability to plan for projects many years in advance. Long duration equities tend to be growth stocks,
and growth often comes from research and development which requires cash outflows in the short-
term, with the hope of earning larger cash inflows in the long-term. This delayed gratification
requires stability within the company’s internal and external operating environments. A manager
will not want to start a 10-year project if they worry their industry could become restricted in the
future or if they fear that their government could seize their assets. Additionally, investors need to
feel protected and confident to support these long-term projects. Investors will require legal
protections for their investments, and they will want a strong reporting and regulatory
environment. I find that countries with strong governance, as defined by the Worldwide
Governance Indicators (WGI) project, have longer equity durations firms on average. The
measures of governance tested individually and in aggregate relate to the stability of the
government, the support from the government for private sector development, the confidence
agents have in contract enforcement and property rights, and the extent to which corruption is
controlled.
From a financial market perspective, I find that, in equilibrium, more financially
developed markets have proportionally more long duration firms. This market development
classification comes from Morgan Stanley Capital International (MSCI) which considers multiple
market accessibility criteria such as the openness to foreign ownership, the ease of capital inflows
and outflows, and the availability of investment instruments. Inflation is also an important country
3

characteristic when considering equity duration. While short duration firms have increased cash
flow risks, long duration firms are more sensitive to expected return shocks (Dechow, Erhard,
Sloan, and Soliman 2021). As interest rates and inflation tend to move in the same direction, I
expect to find countries with higher levels of inflation to have shorter duration firms on average.
Another important country-level characteristic that I predict and find to be associated with
equity duration is tax policy, specifically the length of tax loss carryforward periods and the
availability of innovation box tax incentives. Governments can encourage innovation, investment,
and a focus on long-term cash-flows by allowing for tax losses to be carried forward to be used to
offset future taxable income. These tax loss carryforward rules allow for firms to partly share the
downside risk of their investments with the government (Langenmayr and Lester 2018; Ljungqvist
et al. 2017). In 48 out of 49 countries in my sample, taxable losses are allowed to be carried forward
as of 2020. However, the length of the allowable carryforward period varies from country to
country. On the short end, some countries only allow for a three-year carryforward period. On the
long end, many countries now allow for taxable losses to be carried forward indefinitely. I find
that long duration firms are more likely to be headquartered in countries with generous tax loss
carryforward rules. The companies that benefit most from tax loss carryforwards are firms that are
not yet profitable. Firms that have not yet made a profit are likely to be long duration firms because
it is doubtful that they will return cash flows to their shareholders in the near term when losses
persist.
Another important tax policy is innovation box tax incentives which reduce tax rates on
innovation-related income. There are various types of innovation box tax incentives, but their main
purpose is to split a firm’s annual taxable income into two boxes to which two different corporate
tax rates are applied. The first box contains taxable income related to innovations (e.g., income
4

generated from patents or trademarks) which are taxed at a reduced tax rate (e.g., 9% in the
Netherlands). The second box contains all other taxable income for the firm-year and is taxed at
the standard corporate tax rate for the country (e.g., 25% in the Netherlands) (Silva-Gámez et al.
2022). Chen et al. (2019) have found their implementation to be associated with a reduction of
outbound tax-motivated income shifting and an increase in fixed asset investments and
employment in a sample of European countries. Innovation boxes are output focused tax policies
as they reward the creation of an intangible asset in contrast to research and development (R&D)
tax credits which reward the inputs of innovation. If a company creates a valuable intangible asset
that can generate cash flows many years into the future, the innovation box tax policy will ensure
that those future cash flows continue to be taxed at a reduced rate. With an R&D tax credit system,
there is an incentive to participate in R&D activities, but the same tax benefit is received if those
R&D activities lead to an innovation that provides a competitive advantage for one year or for ten
years. Therefore, innovation box tax policies are predicted to be related to longer duration
investments as they continue to provide a benefit to the firm far into the future if the investment
continues to pay off. No prediction is made for the relationship between duration and R&D tax
credits as they generally only provide a one-time benefit.
After establishing the determinants of cross country variation in duration, I next shift focus
to an asset manager’s perspective and consider the short duration premium within and across
countries. Prior research has found that equity duration captures a strong common factor in stock
returns and has a downward-sloping yield curve (Dechow, Sloan, and Soliman 2004). This result
has been called the short duration premium and is represented by a decreasing alpha in portfolios
sorted on, and increasing in, equity duration. Recent work has found the short duration premium
to have a significant risk-adjusted performance based on 3-factor, 5-factor, and q-factor models
5

(Goncalves 2021). Another common result in the prior literature is that portfolios increasing in
equity duration are also increasing in CAPM beta and total volatility which is attributable to their
increased discount rate risk. I therefore investigate whether these variables, and the short duration
premium, vary consistently across all countries or whether the country’s average duration also
plays a role.
As a first step, I build quintile portfolios, sorted on equity duration, for all firms in
developed (23 countries) and emerging (26 countries) markets. I find negative and significant
CAPM alpha spreads for both the developed and emerging portfolios. However, there are notable
differences between these groups. In the developed markets, CAPM beta is increasing in duration,
and idiosyncratic volatility is decreasing in duration. I find the opposite result in emerging markets,
that is, they have CAPM betas that are decreasing in duration, and idiosyncratic volatility is
increasing in duration. This could be because firms in emerging countries have shorter average
durations, therefore as a firm’s duration becomes longer, the firm has less in common with their
country’s overall market and any increase in risk is firm specific. The reverse would be true in
developed countries, as a firm’s duration becomes longer, they have more in common with their
country’s overall market and therefore share more of the same market risks.
The developed and emerging portfolios are then testing again in a three-factor model
following Fama and French (1993).
3
The same results hold for the equal weighted developed
portfolios, but the negative alpha spread becomes insignificant in the value weighted developed
market portfolios. For the emerging markets portfolios, the negative alpha spread in the equal
weighted portfolios weakens but remains significant depending on the inclusion of microcap
stocks. However, the negative alpha spread in the emerging market value weighted portfolios

3
I thank Ken French for making these factors available on his website.
6

become insignificant. To investigate where the alpha differential loses its significance in the
developed markets, I use the four developed markets regional breakdowns created by Fama and
French (2012): Europe, Japan, Asia Pacific excluding Japan, and North America. I find the 3-factor
alpha spread to be negative and significant in all four equal weighted portfolio groupings, and
negative and significant in Asia Pacific excluding Japan and North America for the value weighted
portfolios.
I continue to refine my results to a more granular level by testing whether the short duration
premium holds within individual countries. I compute implied equity duration for all available
firm-years within the MSCI All Country World Index (ACWI) between 2000 and 2020. While the
country constituents of this index have changed over time, I focus on the 23 developed countries
and 26 emerging countries that made up the index as of the end of 2019. For this test, I create
quintile portfolios sorted on firm-level duration for each of the 49 countries. I find a negative and
significant CAPM alpha spread in 44 out of the 49 countries tested when the portfolios are equal
weighted. When the portfolios are value weighted, the CAPM alpha spread is negative and
significant in 33 out of the 49 countries. The CAPM alpha spread is not positive and significant in
any of the specifications. These results show that the short duration premium exists not only in the
U.S., but in many other developed and emerging countries. Additionally, the short duration
premium can be found not only in portfolios of equities from a single country, but also in portfolios
of equities formed on a regional or market development level.
An important detail for investors to consider with equity duration in an international setting
is which country characteristics are related to larger short duration premia. To answer this question,
I regress the CAPM alpha spread computed from each country’s quintile portfolios on several
country characteristics including those related to country-level duration (i.e., the standard
7

deviation of this measure), the inputs of duration (i.e., the autoregressive coefficients of return on
equity and book equity growth, and the terminal values for return on equity and book equity
growth), and country characteristics (i.e., market development status, industry structure, and the
percentage of market value held by the largest firms). I find that countries with the highest standard
deviation in their country-level measure of duration have the largest CAPM alpha differential
between their fifth and first portfolios. This result holds across the inclusion of all control variables,
and with the alphas generated from value weighted portfolios and from equal weighted portfolios.
This country-level standard deviation is greatest in countries with higher historic costs of capital,
lower historic GDP growth rates, and lower percentages of firms in long duration industries. This
result complements the result found in Goncalves (2021) where the short duration premium in the
U.S. is strongest in years in which the standard deviation of equity duration is high.
Next, I create a country-year measure of equity duration by value-weighting the duration
of each stock within each country-year based on their market capitalization. I create quintile
portfolios sorted on country-level duration (each portfolio contains 9 or 10 countries) and find that
the CAPM alpha differential between portfolio 5 and 1 is negative and significant in an equal
weighted specification when the first quarter of 2020 is removed from the sample. This potential
exclusion is based on prior research which found this quarter to be one in which the short duration
premium is reversed (Dechow et al. 2021). The CAPM beta for these portfolios is increasing in
duration when the MSCI ACWI index is used as the market return. This result demonstrates that
equity duration can be a useful risk factor for investors not only at the firm-level, but also at the
country-level.
In summary, I find that equity duration varies around the world and is influenced or
correlated with economic development and is longer in countries with more established and robust
8

political systems. This is consistent with investors being more willing to take the risk of investing
long-term when they believe the institutional arrangements of the country support long-term
investments.
My dissertation makes three contributions. First, I extend the academic literature related to
equity duration. I am the first to create a country level measure of equity duration based on the
average of the equity durations of the firms headquartered in the country. I predict and find that
countries with stronger governance and with more future focused tax policies are better suited
operating environments for long duration investments. Second, I provide insights to asset
managers by proving that on average the short duration premium holds outside of the U.S., both
within other countries with firm level returns and across countries with country index returns.
Equity duration is a relevant risk factor to investors, and services such as MSCI should consider
including a country level measure of duration in their country index reports.
Finally, my findings are relevant to global organizations such as the United Nations (UN)
and the World Bank. As part of the UN’s Sustainable Development Goals report they identified
several factors that are hampering the global economic recovery in 2022 including rising inflation
and policy uncertainty.
4
My results show that with either of these factors present, equity durations
are shorted on average. When these factors are present, recommendations related to strengthening
governance or modifying tax policy may be able to help. My results have implications for the UN
Goals because they highlight that a societies’ willingness to invest long-term, economic
development, and standard of living are all interrelated.
The remainder of this dissertation is organized as follows. Chapter 2 presents a review of
prior literature on implied equity duration and describes the measure, as well as recent advances

4
https://www.un.org/sustainabledevelopment/wp-content/uploads/2022/07/Goal-8-infographic.pdf
9

in the measure. Chapter 3 develops the main predictions. Chapter 4 describes the sample selection
process and identifies the data sources used. Chapter 5 presents the research design and primary
empirical results. Chapter 6 reports findings relevant to asset managers. Chapter 7 reports the
robustness of the results in Chapter 5 to alternative measures of duration. Chapter 8 provides
additional analysis with alternative country characteristic measures. Chapter 9 concludes with a
summary.

10

Chapter 2

Background

2.1 Measuring Equity Duration

The implied equity duration measure is based on the Macaulay (1938) measure for bond
duration, which has primarily been used to measure bond price sensitivity to changes in the yield
to maturity. However, unlike fixed income securities, equity securities do not have their future
cash flows to shareholders set in advance. Because of this, forecasting assumptions are required to
calculate the future cash flows of a firm (hence the “implied”). Equity securities do not have a set
maturity date either, so assumptions need to be made about cash flows beyond the finite forecasting
period. The general formula for implied equity duration, initially developed by Dechow et al.
(2004), was designed to handle these issues. The formula below splits the duration measure into
two parts: the finite forecast horizon and the terminal value:

𝐷𝑢𝑟𝑎𝑡𝑖𝑜𝑛 =
∑ 𝑡 ∗ 𝐶𝐹
𝑡 /(1 + 𝑅 )
𝑡 𝑇 𝑡 =1
𝑀𝐸
0
+ (𝑇 +
1 + 𝑅 (𝑅 − 𝑔 )
) ∗
𝑀𝐸
0
− ∑ 𝐶𝐹
𝑡 /(1 + 𝑅 )
𝑡 𝑇 𝑡 =1
𝑀𝐸
0
(1)

11

Where ME0 denotes the current market capitalization of equity (stock price at the start of the year
multiplied by shares outstanding), CFt denotes the net cash distributions to equity holders by the
company in year t. R denotes the discount rate. g is the steady state growth rate for the firm after
the finite forecasting period. Finally, T is the number of years in the finite forecast horizon, and it
is set to 15 years as in Weber (2018) and Dechow et al. (2021). Cash flows beyond the finite
forecast horizon are assumed to be paid out to investors as a level perpetuity as in Dechow et al.
(2004), Weber (2018), and Dechow et al. (2021) (i.e., g = 0). Cash flows in this measure are
calculated based on the assumption of clean surplus accounting. That is:

𝐶𝐹
𝑡 = 𝐸𝑎𝑟𝑛𝑖𝑛𝑔𝑠 𝑡 − ∆𝐵𝐸
𝑡 = 𝐵𝐸
𝑡 −1
∗ (𝑅𝑂𝐸 𝑡 − 𝐺 𝑡 ) (2)

Following Dechow et al. (2004), I forecast ROEt (i.e., Return on Equity; Earningst/BE t − 1)
by assuming that ROE follows a first-order autoregressive (AR(1)) process with a set long-run
mean. I forecast Gt (i.e., future growth in book equity) based on a first-order autoregressive process
of past sales growth, with a set long-run mean, as Nissim and Penman (2001) have shown that this
is a better predictor of future growth in book equity than past book equity growth. I calculate the
AR(1) coefficients for ROE and G separately for each of the 49 countries in my sample for the 27-
year period between 1995 and 2021. This calculation results in a persistence coefficient of 0.53
and 0.18 for ROE and G respectively within the United States. These results are in line with the
0.57, 0.41, and 0.39 ROE persistence coefficients and 0.24, 0.24, and 0.21 G persistence
coefficients in Dechow et al. (2004), Weber (2018), and Dechow et al. (2021) respectively. The
other 48 countries in the sample have a mean ROE persistence coefficient of 0.51 and a mean G
persistence coefficient of 0.14.
12

The long-run mean of ROE is set to approximate the median cost of equity (i.e., R) for each
country for the period between 2000 and 2020 so that the ROE will converge to R. The cost of
equity is set equal to the Total Equity Risk Premium plus the Riskfree Rate as defined by
Damodaran (2020a) and Damodaran (2020b).
5
This measure results in a long-run mean ROE of
6% for the United States. While a 12% rate was used in Dechow et al. (2004), Weber (2018), and
Dechow et al. (2021), this 6% rate better represents the macroeconomic conditions during the
sample period. This is the same rate used in the current period example in Dechow et al. (2021).
The other 48 countries in the sample have a mean terminal ROE of 9%.
I set the long-run mean of G equal to the median GDP growth from 2000 to 2020 for each
country. The historic GDP annual percentage growth rates for each country were obtained from
the World Bank’s website.
6
The United States’ median GDP growth rate between 2000 and 2020
was 2%. This is less than the 6% terminal growth rate used in Dechow et al. (2004), Weber (2018),
and Dechow et al. (2021). It is closer to the 0% terminal growth rate used in the current period
example in Dechow et al. (2021). The other 48 countries in the sample have a mean terminal G of
3%. This process creates cross-sectional variation in ROE and G, but both parameters remain
temporally constant as in prior research.
7

5
I thank Aswath Damodaran for making this dataset publicly available. Current and historic dataset can be found on
his website: https://pages.stern.nyu.edu/~adamodar/. The risk-free rate is based on the “Differential Inflation Riskfree
Rate” computation as 10-year sovereign bond yields are not available for many of the countries in the sample.
6
https://data.worldbank.org/indicator/NY.GDP.MKTP.KD.ZG
7
I also winsorize the bottom and top five countries for long-run means of ROE and G and the terminal values of ROE
and G so that no country is too extreme in their forecasting assumptions.
13

2.2 Alternative Measures of Equity Duration

Two recent papers have attempted to improve the forecasting and discounting process in
computing the implied equity duration measure. Chen and Li (2022) assume that the terminal value
is a growth perpetuity, instead of a steady state perpetuity, with a growth rate equal to the long-
run growth rate (i.e., G = g). The equity duration measure represents a weighted average of the
time to receive payments, and the growth perpetuity assumption increases the amounts received in
later years. Therefore, this assumption results in a longer duration measure on average. The mean
duration in Chen and Li (2022) is 28.6 for their United States based sample period between 1975
and 2017. In Dechow et al. (2021), the mean duration is 17.4 for their United States based sample
period between 1964 and 2019. Another innovation of Chen and Li (2022) is the use of a vector
autoregression (VAR) to simultaneously forecast ROE and G instead of forecasting them with two
independent AR(1) processes. The VAR system allows for the inclusion of more relevant
accounting variables, and it could potentially offer better forecasting. The authors use 6 variables
to estimate the future ROE and G. Additionally, instead of calculating the VAR coefficients for
the full sample, the authors recompute the coefficients each year with data available at that time.
Finally, they allow the growth horizon, T, to vary by firm-year.
Another recent paper to create a modified implied equity duration measure is Goncalves
(2021). He also uses a VAR system, although his system is based on 12 state variables. Goncalves
(2021) makes several other contributions such as extending the finite forecast period to 1,000 years
and estimating a discount rate for each firm-year. In related work, Goncalves and Leonard (2023)
use this forecasting methodology to study explanations for the decline in the value premium. Chen
and Li (2022) note that while estimating a discount rate for each firm-year is theoretically
appealing, in practice their tests showed that it was easy to yield extreme estimates due to
14

forecasting errors.
8
As such, Goncalves (2021) caps his measure of duration at 500 years because
there are some firms-year observations with extremely long durations. When capping duration at
500 years, he did it in a way that preserves the rankings. These changes result in a longer mean
equity duration of 59.5 years for his United States based sample period between 1973 and 2018.
Chen (2022) creates a novel measure of equity duration that does not involve forecasting
assumptions but instead infers an equity’s duration by observing the equity’s price movements
around unexpected movements in the federal funds rates. Changes to the federal funds rate, and
U.S. monetary policy in general, will not have the same effect on international firms as domestic
firms, therefore this measure is not suitable for my setting. I opt to use the first order univariate
autoregressive processes as in Dechow et al. (2004), Weber (2018), and Dechow et al. (2021) as
the requirement of 6 variables from Chen and Li (2022) or 12 variables from Goncalves (2021)
would reduce the number of firms in my sample.
9
Comparability is the reason that I chose to
maintain the level perpetuity assumption (i.e., g = 0). Finally, I acknowledge that an exogenously
specified discount rate for the full sample is not theoretically ideal. My method of computing a
separate discount rate for each country represents a middle ground between using a single rate,
and Goncalves (2021)’s estimation of a distinct rate for each firm-year. In Chapter 5, I discuss the
results of two alternative specifications: one where I compute the implied discount rate for each
firm-year and another where I set the discount rate to 0% for all firm-years so that the difference
in equity duration comes solely from the differences in the timing of the forecasted cash flows. I
then compare those results to the primary analysis.

8
Chen and Li (2022) opted for the “theoretically unappealing” method of setting the discount rate to 12% for the full
sample as in Dechow et al. (2004), Weber (2018), and Dechow et al. (2021) as it performed better empirically and
still provided a reasonable estimate of duration.
9
Some variables like O-score in Chen and Li (2022) and payout yield in Goncalves (2021) require several
subcomponent variables to calculate. This could reduce the sample size as international databases often include
coverage of less accounting variables compared to their domestic counterparts (Dai 2017).
15

Finally, Gormsen and Lazarus (2022) study duration in an international setting. Their first
measure of duration is not built on forecasting and discounting cash flows to shareholders in the
Macaulay framework, but is instead based upon fitted values of a regression of analyst growth
forecasts on five common cross-sectional return predictors: CAPM beta, book to market,
profitability, investment, and payout. They find a positive alpha (they test short duration minus
long duration portfolios) in 20 out of 23 developed countries tested which appears to be significant
in 16 of those countries. However, the characteristics underlying their duration factor are based on
U.S. data and not international data. They create a second novel measure of equity duration using
single-stock dividend futures (i.e., dividend strips). This measure is restricted to firms that issue
dividends. Ye et al. (2019) demonstrate that less than half (46%) of non-financial non-utility firms
issue cash dividends over the period between 2000 and 2013 for a sample of 22 developed and
emerging countries. This dividend strip measure of duration is further limited to equities with such
futures available. I distinguish my work from Gormsen and Lazarus (2022) by computing equity
duration by forecasting and discounting cash flows to shareholders in the Macaulay framework, at
the country-level in addition to the firm-level and by reporting the country-level duration in years.
Additionally, I include emerging markets in my sample, which were not previously tested, and
find that their results differ from developed markets. I further contribute to the literature by
studying the cross-sectional differences in countries that are related to differences in country-level
duration and by testing the country specific characteristics that are related to the magnitude of the
short duration premium in duration sorted portfolios.

16

2.3 Country-Level Governance

With regards to the country-level measures of governance, Ellahie and Kaplan (2021) show
that firms in countries with weak institutions pay dividends earlier in their life cycles and are more
likely to commit themselves to a set dividend policy based on a percentage of earnings compared
to firms in countries with strong institutions. Building on these results, Kapons, Kelly, Stoumbos,
and Zambrana (2022) show that investors value dividends more in periods and environments where
trust is lower. These findings are related to my hypothesized relationship between country-level
equity duration and country-level governance. However, my results extend these papers as a firm’s
equity duration is not just decreased by dividend payments to shareholder, but is also impacted by
a firm’s stock issuances and repurchases which may not be directly related to their dividend policy.
Additionally, not all firm issue dividends, and the equity duration measure allows an analysis of
those firms that a direct payments test would exclude.

17

Chapter 3

Countries Examined and Prediction Development

3.1 Countries

I analyze 49 countries across various geographic regions and levels of development. These
countries are selected based on the availability of data that is used for my predictions. For a country
to be retained in the sample, I require it to be included in the MSCI ACWI as of the end of 2019.
10

This list overlaps well with the developed and emerging classifications used on Ken French’s
website for his factor model data. Table 1 provides a list of the 49 countries sorted alphabetically
after being grouped by both development status and geographic region as defined by MSCI as of
the end of 2019. There are 23 developed and 26 emerging countries in the sample. Collectively,
these countries represent over 90% of global GDP and six continents are represented.

10
https://www.msci.com/our-solutions/indexes/market-classification
18

3.2 Predictions

I expect the equity duration of a given firm to vary based on the country in which they are
headquartered. Countries vary on many parameters (e.g., corruption, rule of law, regulatory
quality, political stability, etc.), and I expect that several of these parameters could encourage or
discourage managers from planning for growth and making investments that are guaranteed to
result in negative cash-flows in the short-term, with the expectation that they will pay off with
positive cash-flows in the long-term. Kaufmann, Kraay, and Mastruzzi (2010) (KKM) document
the World Bank’s Worldwide Governance Indicators (WGI) project which defines country-level
governance as the traditions and institutions by which authority in a country is exercised. This
includes the capacity of the government to effectively formulate and implement sound policies and
the respect of citizens and the state for the institutions that govern economic and social interactions
among them. The WGI project has created two measures for each of these areas of governance and
has calculated values for these measures for over 200 countries and territories since 1996. The four
measures I use are Government Effectiveness (GE), Regulatory Quality (RQ), Rule of Law (RL),
and Control of Corruption (CC). Each measure combines corporate and individual statistics for
each country, as well as expert survey responses.
Government Effectiveness represents the quality of public services and the quality of policy
creation and implementation (KKM). It also captures a government’s ability to commit to a plan
and the extent to which their policy is created free from political pressure. I predict that countries
with high levels of government effectiveness will be well suited for long duration investments.
Regulatory Quality represents the government’s ability to permit and promote private sector
development (KKM). My sample is of publicly traded companies and does not capture all
government entities and activities, so the promotion of private sector development will help create
19

an environment suited for the observation of long duration firms. If private sector development
was not encouraged by a government, I would not expect private sector firms to commit to long
term investments. Rule of Law is the extent to which agents have confidence in and follow their
local legal framework (KKM). This includes managers’ confidence in contracts being enforced,
property rights being upheld, and courts ruling on cases consistently. A country with a strong rule
of law is predicted to be opportune for long duration investments as managers will be confident in
fair and consistent business practices today and in the future. Control of Corruption measures the
extent to which the state is captured by private interests and the extent to which public officials
use their power for their own gain (KKM). Control of corruption (i.e., limited corruption within a
country) is considered to be the positive outcome and I predict that an environment with limited
corruption will be a fairer environment which will allow managers to be comfortable committing
to long term plans. However, some prior research suggests that corruption and inefficiencies do
not necessarily need to go together. It could be possible that corruption acts as a lubricant and
keeps local economies running (Leff 1964). For a company that participates in corruption, they
may feel protected and more willing to take chances because they know that they have friends
within the government that will look out for them.
In addition to these four WGI measures, there is also Political Stability and Absence of
Violence/Terrorism (PV) and Voice and Accountability (VA). PV is the perceived likelihood that
the government will be substantially weakened or overthrown, it also includes the perceived threat
of politically motivated terrorism and violence (KKM). Instability of any type creates uncertainty
and may reduce managers’ willingness to commit to long duration investments. Additionally, the
threat of terror attacks on infrastructure will also undermine confidence in building up operations
in a given country. However, I exclude this measure from my composite to be consistent with
20

Ellahie and Kaplan (2021). VA represents an individual citizen’s impact on selecting their
government and political officials (e.g., voting procedures) (KKM). This measure also includes
the freedom of speech that citizens and the media have. I do not include this measure in my primary
analysis as it is not clear if a more voting rights will result in more or less stability. Slim majorities
in political parties could result in drastic swings in policy whereas one party rule may result in
stability.
The WGI measures have been used in the prior accounting literature to measure country-
level governance, sometimes referred to as institutional quality (e.g., Ellahie and Kaplan 2021).
Similar measures have been used to demonstrate the importance of rule of law, government
effectiveness, and regulator quality in developing financial markets within a country (e.g., La
Porta, Lopez-de-Silanes, Shleifer, and Vishny 1997; La Porta, Lopez-de-Silanes, Shleifer, and
Vishny 1998). I predict that countries with more mature capital markets and greater investor
protections are more suitable environments for long duration firms. For a firm to have a longer
duration, its market capitalization must be weighted more towards future cash flows to
shareholders compared to a higher weighting on near-term cash flows in short duration firms. For
this weighting to exist, I would expect that managers need to be comfortable planning for long-
term operations and investments. This requires stability within their headquartered country from a
legal perspective, and a perception of fairness of enforcement across firms within the country.
Because of this, I predict that countries with strong governance are hosts to headquartered firms
with equity durations that are longer on average compared to the equity durations of firms
headquartered in countries with weak governance.

21

Hypothesis 1: Firms will on average have longer implied equity durations in countries
with strong governance.

Corporate taxation rules and regulations vary significantly from country to country.
Differences appear in the corporate tax rate, tax loss carryforward rules, and the treatment of
intangible assets and their resulting cash flows. Prior research shows that increased taxation can
hinder innovation within a region. Mukherjee, Singh, and Zaldokas (2017) find that increasing
corporate tax rates at the U.S. state-level reduces future innovation (e.g., number of patents
produced, R&D investments made, and the introduction of new product lines). They also find that
higher corporate tax rates reduce innovation and risk-taking incentives. Li, Ma, and Shevlin (2021)
find that when U.S. states adopt addback statutes to reduce the tax benefit from creating intangible
assets, firms in those states decrease the number of patents they produce. Innovation is related to
equity duration because innovation relies on research and development investments which require
cash outflows in the short-term, but expect to produce a product or process that will generate cash
inflows in the long term. Another way for governments to encourage innovation, investment, and
a focus on long-term cash-flows is to allow for tax losses to be carried forward and used to offset
future taxable income. 48 of the 49 countries in my sample allow taxable losses to be carried
forward as of 2020.
11
However, the length of the allowable carryforward period varies from
country to country. On the low end, the Philippines only allow for a three-year carryforward period.
On the high end, many countries (e.g., the United States) now allow for taxable losses to be carried

11
The United Arab Emirates does not have a corporate tax system and therefore net operating losses do not exist.
22

forward indefinitely.
12
I predict that a country’s mean firm’s equity duration is increasing in the
length of that country’s tax loss carryforward period.
Additionally, some countries have created innovation box tax incentives that reduce the tax
rate on income generated by innovative technologies designed within their country. The
implementation varies by country as some require the creation of a patent, trademark, copyright,
etc. to qualify and there are also different treatments for intellectual property that is acquired. Chen
et al. (2019) find that, on average, the creation of innovation box tax incentives is associated with
a reduction of outbound tax-motivated income shifting and an increase in fixed asset investments
and employment in a sample of European countries. If a company creates a valuable intangible
asset that can generate cash flows many years into the future, the innovation box tax policy will
ensure that those future cash flows are taxed at a reduced rate. Therefore, I predict that country-
level equity duration is longer on average in countries in which innovation box tax incentives are
available.

Hypothesis 2: Firms will on average have longer implied equity durations in countries
with tax policies that favor long term planning and innovation.

To the extent that duration varies internationally within countries, and different countries
have different average firm equity durations, I expect the short duration premium to exist at both
the firm-year level within countries and at the country-year level across countries globally. Testing
the short duration premium internationally requires several assumptions. First, I need to have an
accurate forecasting procedure for each country. For example, if an AR(1) process is only

12
There are some limitations to the indefinite tax loss carryforward in certain countries. In the United States for
example, the indefinite tax loss carryforward can only be used to offset 80% of the adjusted taxable income each year.
23

appropriate for developed countries, then using this model in emerging countries may prevent me
from finding results. Similarly, the assumption to use sales growth as a proxy for book equity
growth may be appropriate in the U.S., but it may not be the best method internationally.
Ultimately, the existence of the short duration premium internationally is an empirical question.
Given that equity duration varies by firm-year, and that the average duration of the firms
within a country varies by country-year, I expect the magnitude of the short duration premium to
vary across countries. If equity duration represents a risk factor, then I expect the risk exposure
across duration portfolios to increase as the variability of equity duration increases within a
country. Therefore, I predict that countries with larger standard deviations in their country-level
measure of equity duration will have greater differences between the CAPM alphas computed for
the long duration portfolios compared to those for the short duration portfolios.

24

Chapter 4

Sample and Descriptive Statistics

4.1 Data

I obtain annual accounting data and monthly stock return data from FactSet Fundamentals
for my sample of domestic and international firms. FactSet’s global database is built on a copy of
Worldscope. In comparison to other international databases, FactSet has a good balance between
company size and coverage, and a reasonable selection of accounting items and geographic
exposure (Dai 2017). For robustness, I compute the country-level measure of equity duration using
data from Compustat Global and find the country-level measure from the two different sources to
have a correlation greater than 80%. While my sample may miss some smaller countries, the 49
countries in my sample combine to represent over 90% of global GDP. I obtain rules related to
each country-year’s tax loss carryforward periods from EY’s Worldwide Corporate Tax Guides
for the years between 2004 and 2020 which has the current policies for each country as of January
1 for the given year.
13
I use MSCI indices for each of the 49 countries in my sample to get a

13
https://www.ey.com/en_id/tax_law_guides/worldwide-corporate-tax-guide-20201 I assume that the tax loss
carryforward rules between 2000 and 2003 are consistent with 2004 as the EY guides are not available for those years
unless otherwise stated.
25

complete measure of each country's total equity market monthly returns which I obtain from
FactSet Fundamentals. The innovation box incentive availability by country-year are taken from
Silva-Gámez et al. (2022) and Chen et al. (2019). Industry data are based on SIC codes and come
from FactSet Fundamentals as well. Inflation data by country-year was downloaded from The
World Bank’s website. All other data sources were previously described. The primary sample
period is 2000 to 2020.

4.2 Descriptive Statistics

Table 2 provides the value weighted mean, equal weighted mean, median, and standard
deviation of each country's implied equity duration measure. These statistics are based on the full
panel of data for each country from 2000 to 2020. Panel A also contains the forecasting
assumptions that were used to calculate the equity duration for each country. The discount rate and
terminal values for ROE and G are median values for the sample period, while the autoregressive
coefficients are based on all available data for each country between 1995 and 2021. Panel B
provides the number of firm-years for each country in the sample and the composite measure of
the four WGI measures per country averaged over the sample period. The WGI measures are
reported in percentile rank terms ranging from 0 (worst) to 100 (best). The four measures are
interrelated and are strongly positively correlated. Panel B provides the length of the tax loss
carryforward period allowed for each country as of the end of 2020, and an indicator for whether
the country has an innovation box tax policy in place by 2020. Finally, this panel also includes the
percentage of firm-years in each country that are in shortest duration industries (Categories 1 and
3 of the Fama and French 12 industries) and the percentage of firm-years that are in longest
26

duration industries (Categories 6 and 10 of the Fama and French 12 industries).
14
The notation as
a long or short duration industry is based on a comparison of the average duration in each industry
for the sample used in this paper.
To minimize the influence of outliers, I winsorize all continuous variables at the 1
st
and
99
th
percentiles. Equity duration is computed for each firm-year using accounting information
lagged by three months (i.e., I assume a three-month lag on the public availability of financial
information after a fiscal year-end) and market values based on the final day of the previous fiscal
year. The country level measure of equity duration (Country-Year Duration) is based on the value
weighted mean for each country in the portfolio analysis tests so that larger companies are
represented for their greater influence on their country’s economy. Portfolios are formed in June
each year. For equal-weighted portfolios, I follow Hou, Xue, and Zhang (2019) and exclude
microcaps (defined as firms below the 20
th
percentile of market equity for the country’s sample-
year) to make sure results are not due to these firms. Equal weighted results are generally stronger
in all tests when these firms are retained.

14
#1: Consumer NonDurables -- Food, Tobacco, Textiles, Apparel, Leather, Toys. #3: Manufacturing -- Machinery,
Trucks, Planes, Off Furn, Paper, Com Printing. #6: Business Equipment -- Computers, Software, and Electronic
Equipment. #10: Healthcare, Medical Equipment, and Drugs.
27

Chapter 5

Empirical Results

5.1 Visual Preview of Results

As a first step, I demonstrate that Country-Year Duration varies across countries. I create
a visualization of this variation in Figure 1 by filling in each of the 49 countries in my sample on
a map of the world with a color gradient ranging from red (short duration) to green (long duration).
The areas with the highest duration on average are found in North America, western Europe,
Australia, and eastern Asia. Table 2 provides similar information numerically and visually as it is
sorted by the value weighted equity duration for each country. Singapore has the longest average
Country-Year Duration (29.58 years) and Pakistan has the shortest (13.48 years).
Figure 2 shows a time trend of the value weighted Country-Year Duration for groupings
of 23 developed countries, 26 emerging countries, and the combined 49 country world group which
are classified as the ACWI. Throughout the sample period, developed countries have longer
average equity durations than emerging countries. The equity duration computed for the full ACWI
remains fairly stable over the sample period and ranges from 24 to 26 years. I would not expect
significant changes to the equity duration of the ACWI over time unless there are radical changes
28

to the global economy. The biggest single year movement came in 2008 during the global recession
brought on by the financial crisis. During this time the ACWI equity duration decreased by
approximately 1.5 years.
Table 3 provides summary statistics at the country-year observation level for the dependent
variable, primary independent variables, control variables, and sample sizes used in the regressions
in Tables 5, 7, and 9. The Governance (WGI) measure is missing observations as data was not
available for 2001. Tax Loss Carryforward and Innovation Box are both missing observations as
the UAE does not have a corporate tax system.

5.2 Governance (WGI) and Country-Year Duration

I predict that countries with weak governance are hosts to headquartered firms with equity
durations that are shorter on average compared to the equity durations of firms headquartered in
countries with strong governance. Countries with more mature capital markets and greater investor
protections are predicted to be more suitable for long duration firms. For a firm to have a longer
duration, it must have a higher weighted average of time to distribute cash to shareholders (i.e., a
market capitalization that is weighted more towards future cash flows to shareholders) compared
to a lower weighted average of time for shareholders to receive payouts in short duration firms.
For this weighting to exist, managers need to be comfortable planning for long-term strategies and
investments. This requires stability within their country and a strong rule of law. This weighting
towards future cash flows also requires a belief from investors that the company will be able to
continue to operate as planned into the future.
29

Figure 3 demonstrates the variation in the average Governance (WGI) for each country
during the sample period. The countries range from as low as 25 to as high as 98 for this composite
measure. Table 5 reports results from the following country-year level regressions:

𝐶𝑜𝑢𝑛𝑡𝑟𝑦 − 𝑌𝑒𝑎𝑟 𝐷𝑢𝑟𝑎𝑡𝑖𝑜𝑛 𝑖 ,𝑡 = 𝛽 1
𝐺𝑜𝑣𝑒𝑟𝑛𝑎𝑛𝑐𝑒 (𝑊𝐺𝐼 )
𝑖 ,𝑡 + 𝜃𝐶𝑜𝑛𝑡𝑟𝑜𝑙𝑠 𝑖 ,𝑡 + 𝜀 𝑖 ,𝑡 (3)

Where Country-Year Duration is the equal weighted mean equity duration of all firms within a
country-year. Governance (WGI) is a composite measure of made up of the average of four WGI
measures for each country-year.
15
The sample period is 2000 to 2020 and both Country-Year
Duration and Governance (WGI) vary by year within countries.
Country-Year Duration is positively associated with Governance (WGI) in the univariate
regression reported in Column 1. In Column 2 I add in country level control variables and note
that Country-Year Duration is longer in countries with more developed financial markets
(coefficient of 1.748 on Developed) and is shorter in country-years with higher levels of relative
inflation (coefficient of -0.134 on Inflation). This specification also controls for the percentage of
firms in a country-year that are in the short duration industries (Short Dur. Industry %) and long
duration industries (Long Dur. Industry %). The last control represents the percentage of market
value for the country-year that is held by the top 10% of firms in that country-year (Largest Firm
%). With these controls in place, the coefficient on Governance (WGI) is 0.115 and is statistically
significant. This coefficient indicates that as Governance (WGI) increases by a standard deviation
(i.e., 22 points on the composite measure) Country-Year Duration increases by 42% of a standard
deviation which is equal to 2.5 years of equation duration, on average, with all else held equal.

15
The composite measure takes an average of the four main individual measures. Therefore, its scale remains 0 to
100.
30

Columns 3 through 5 test alternative specifications of Column 2 in an attempt to correct
for cross-sectional and time-series dependence (Gow et al. 2010). Column 3 replicates Column 2,
with t-statistics based on Newey-West corrected Fama-MacBeth standard errors (Fama and
MacBeth 1973; and Newey and West 1987). Column 4 replicates Column 2, with t-statistics based
on cluster-robust standard errors that are clustered by country. Column 5 replicates Column 2,
except now the panel has been collapsed down to a single observation per country with each
variable representing its sample average for the respective country.
In summary, the results indicate that the Governance (WGI) remains positive and
significant in all specifications. This suggests that the strength of a country’s governance is not
subsumed by these other metrics but is incrementally significant in explaining the cross sectional
differences in Country-Year Duration. This is consistent with my prediction that firms will on
average have longer implied equity durations in countries with stronger governance.

5.3 Future Focused Tax Policy and Equity Duration

Innovation is predicted to be positively related to equity duration because innovation can
result in long-term cash flows. One way for governments to encourage innovation is through tax
policy. I predict that firms will on average have longer implied equity durations in countries with
tax policies that favor long term planning and innovation.

5.3.1 Tax Loss Carryforwards and Country-Year Duration

One tax policy that is common around the world is the allowance for tax losses to be carried
forward and used to offset future taxable income. I predict that a country’s mean firm’s equity
31

duration (Country-Year Duration) increases in the length of that country’s tax loss carryforward
period (Tax Loss Carryforward). Figure 4 shows the variation in the tax loss carryforward policy
(Tax Loss Carryforward) by country. This figure provides a graphical representation of the number
of years that taxable losses can be carried forward to offset future taxable income. The number of
years allowed has changed for many countries during the sample period, so this figure is an average
for each country. Many countries such as the U.S. now allow for indefinite carryforwards with
minimal restrictions (these country-years are coded as 30 years). However, there are still many
countries that limit tax loss carryforwards to 10 years or less. The figure also shows that 48 out of
49 countries in my sample allow taxable losses to be carried forward as of 2020. However, the
length of the allowable carryforward period varies by country. Table 7 reports results from the
following country-year level regressions:

𝐶𝑜𝑢𝑛𝑡𝑟𝑦 − 𝑌𝑒𝑎𝑟 𝐷𝑢𝑟𝑎𝑡𝑖𝑜𝑛 𝑖 ,𝑡 = 𝛽 1
𝑇𝑎𝑥 𝐿𝑜𝑠𝑠 𝐶𝑎𝑟𝑟𝑦𝑓𝑜𝑤𝑎𝑟𝑑 𝑖 ,𝑡 + 𝜃𝐶𝑜𝑛𝑡𝑟𝑜𝑙𝑠 𝑖 ,𝑡 + 𝜀 𝑖 ,𝑡 (4)

The results in Table 7 indicate that long duration firms are more likely to be headquartered
in countries with generous tax loss carryforward rules. In this table and in the summary statistics
in Table 6, Tax Loss Carryforward measures the number of years a taxable loss can be carried
forward to offset future taxable income. Many countries in the sample allow for an indefinite
carryforward period. For these countries, there is often a limit on how much taxable income can
be offset in the current year with losses from prior years. For example, in the United States, taxable
losses from prior years can be used to offset up to 80% of current year adjusted taxable income. I
code these indefinite carryforward policies as equal to 30 years if the carryforward can be used to
offset 50% or more of the current year taxable income, and as 25 years if the indefinite
32

carryforward can be used to offset less than 50% of the current year taxable income. The country-
level policies are based on the tax laws in place at the start of each year. I manually collect the net
operating loss carryforward limit from EY’s Worldwide Corporate Tax Guide for each year. While
the net operating loss carryforward policy does not change often within a country, 18 countries do
experience one or more changes in their allowable loss carry forward limit during the sample
period. In every one of these 18 countries, the carryforward period was extended (net of any
reductions).
In Table 7 Column 2, the coefficient on Tax Loss Carryforward is positive and significant.
The coefficient of 0.061 indicates that as Tax Loss Carryforward increases by a standard deviation
(i.e., 11 years) Country-Year Duration increases by 12% of a standard deviation which is equal to
0.7 years of equation duration, on average, with all else held equal. Columns 3 through 5 perform
the same alternative specification tests as in Table 5. The coefficient on Tax Loss Carryforward
remains positive in all specifications but is only statistically significant in Column 4.
In summary, the results indicate that the Tax Loss Carryforward remains positive and
significant in most specifications. This suggests that the length of a country’s allowable
carryforward period for tax losses is not subsumed by these other metrics but is incrementally
significant in explaining the cross sectional differences in Country-Year Duration. This is
consistent with my prediction that firms will on average have longer implied equity durations in
countries with longer tax loss carryforward periods.

33

5.3.2 Innovation Boxes and Country-Year Duration

In addition to Tax Loss Carryforward in Table 7, I also predict there to be a positive
relationship between the availability of innovation box tax policies (Innovation Box) and Country-
Year Duration. Innovation Box is a dummy variable is equal to 1 if the country-year has innovation
box tax incentives in place. 16 countries in my sample have these incentives in place by the end of
the sample period, and only two (France and Ireland) had the policy in place during the entire
sample period. Table 9 reports results from the following country-year level regressions:

𝐶𝑜𝑢𝑛𝑡𝑟𝑦 − 𝑌𝑒𝑎𝑟 𝐷𝑢𝑟𝑎𝑡𝑖𝑜𝑛 𝑖 ,𝑡 = 𝛽 1
𝐼𝑛𝑛𝑜𝑣𝑎𝑡𝑖𝑜𝑛 𝐵𝑜𝑥 𝑖 ,𝑡 + 𝜃𝐶𝑜𝑛𝑡𝑟𝑜𝑙𝑠 𝑖 ,𝑡 + 𝜀 𝑖 ,𝑡 (5)

The positive and significant coefficient on Innovation Box in Column 1 indicates there is a
positive relationship between Country-Year Duration and these policies, however this relationship
loses its significance if all of the controls from Panel A are included. Innovation Box policies are
available in 15% of all country-years in the sample, 25% of Developed country-years and 6% of
emerging (non-Developed) country-years. Therefore, I removed the Developed indicator control
variable from these tests. Table 9 Column 2 shows that Innovation Box has a positive and
significant coefficient and that it has incremental significance over the other control variables. The
coefficient of 1.269 indicates that as Innovation Box increases from 0 to 1, Country-Year Duration
increases by 8% of a standard deviation which is equal to 0.5 years of equation duration, on
average, with all else held equal. Columns 3 through 5 perform the same alternative specification
tests as in Panel A. The coefficient on Innovation Box remains positive in all specifications but is
only statistically significant in Column 4.
34

Table 6 reports Tax Loss Carryforward by country-year for 2004 to 2020 and Table 8
reports Innovation Box by country-year for 2004 to 2020. For both Tables 6 and 8, 2000 through
2003 were removed due to space limitations.
In this country-year sample, Tax Loss Carryforward and Innovation Box have a correlation
of 0.14. When comparing the independent variables of interest in Table 5, 7, and 9, Governance
(WGI) has a correlation of 0.46 with Tax Loss Carryforward and 0.21 with Innovation Box. This
indicates that these three measures are capturing distinct country characteristics.

35

Chapter 6

Implications for Asset Management

6.1 Short Duration Premium

After establishing the determinants of cross country variation in duration, I shift focus to
an asset manager’s perspective and consider the short duration premium within and across
countries. I next test the empirical question of the existence of the short duration premium
internationally across various specifications. I predict that the short duration premium will hold
within countries around the world, as long as there is sufficient variation in the duration of the
underlying firms within the country. Additionally, I predict that the short duration premium will
hold on a macro level across countries. That is, if portfolios of country index funds were sorted on
Country-Year Duration, then I predict that the portfolio with the shorter duration country indices
will on average earn a higher return.

36

6.1.1 Short Duration Premium within Developed and Emerging
Markets

First, I build quintile portfolios, sorted on Firm-Year Duration, for all firms in developed
(23 countries) and emerging (26 countries) markets for the period between 2000 and 2021. In Table
10 Panel A, I find negative and significant CAPM Alpha differentials (i.e., Q5-Q1) for both the
developed (-1.02 in equal weighted and -0.57 in value weighted) and emerging portfolios (-0.72
in equal weighted and -0.67 in value weighted). While the emerging value weighted portfolio is
initially insignificant, it becomes significant if the first quarter of 2020 is dropped from the sample
as this is a period where the short duration premium is known to have reversed due to the COVID-
19 induced shutdown (Dechow et al. 2021). In general, the results in Table 10 are strengthened
when the first quarter of 2020 is excluded from the sample. When comparing the results for
developed and emerging markets, there are notable differences between these groups. In the
developed markets, MKT Beta is increasing in Firm-Year Duration, and idiosyncratic volatility
(IVol) is decreasing in Firm-Year Duration. I find the opposite result in emerging markets, that is,
they have MKT Betas that are decreasing in Firm-Year Duration and IVol is increasing in Firm-
Year Duration. This could be because emerging countries have shorter average durations, therefore
as a firm’s duration becomes longer, the firm has less in common with their country’s overall
market and any increase in risk is firm specific. The reverse would be true in developed countries,
as a firm’s duration becomes longer, they have more in common with their country’s overall
market and therefore share more of the same market risks.
The developed and emerging portfolios are then testing again in a three-factor model
following Fama and French (1993). The same results hold for the equal weighted developed
portfolios, but the negative CAPM Alpha differential becomes insignificant in the value weighted
37

developed market portfolios. For the emerging markets portfolio, the negative CAPM Alpha
differential in the equal weighted weakens but remains significant depending on the inclusion of
microcap stocks. However, the negative CAPM Alpha differential from the emerging market value
weighted portfolios becomes insignificant. In Panel B, I investigate where the CAPM Alpha
differential loses its significance in the developed markets by using the four developed markets
regional breakdowns created by Fama and French (2012): Europe, Japan, Asia Pacific excluding
Japan, and North America. I find the 3-factor alpha (FF Alpha) differential to be negative and
significant in all 4 equal weighted portfolio groupings, and negative and significant in Asia Pacific
excluding Japan and in North America for the value weighted portfolios.

6.1.2 Short Duration Premium within Countries

I continue to narrow my results to a more granular level by testing if the short duration
premium holds within individual countries as it does in the United States. In Table 11 I compute
Firm-Year Duration for all available firm-years within the MSCI ACWI between 2000 and 2021.
For this test, I create quintile portfolios sorted on Firm-Year Duration for each of the 49 countries.
I find a negative and significant CAPM Alpha differential in 44 out of the 49 countries tested when
the portfolios are equal weighted (Panel A). When the portfolios are value weighted, the CAPM
Alpha differential is negative and significant in 33 out of the 49 countries (Panel B). The CAPM
Alpha differential is not positive and significant in any of the specifications. I create a visualization
of the results from Panel A in Figure 5 and the results from Panel B in Figure 6. Countries in which
the CAPM Alpha differential is insignificant are coded as 0. These results show that the short
duration premium exists not only in the U.S., but in many other developed and emerging countries.
Additionally, short duration premium can be found not only in portfolios of equities from single
38

countries, but also in portfolios of equities created on regional and market development levels.
Finally, this table replicates the U.S. finding of MKT Beta increasing in portfolios increasing in
Firm-Year Duration, however the MKT Beta result is mixed internationally.

6.2 Determinants of the Short Duration Premium

An important detail for investors to consider with equity duration in an international setting
is which country characteristics are related to larger short duration premia. I predict that countries
with more variation in Firm-Year Duration will on average have larger short duration premia. To
answer this question, in Table 12 Panel A I regress the following equation:

𝑄 5 − 𝑄 1 𝐴𝑙𝑝 ℎ𝑎 𝑖 = 𝛽 1
𝑆𝐷
𝑖 + 𝛽 2
𝑅𝑂𝐸 𝐴𝑅
𝑖 + 𝛽 3
𝑅𝑂𝐸 𝑇 𝑖 + 𝛽 4
𝐺 𝐴𝑅
𝑖 + 𝛽 5
𝐺 𝑇 𝑖 +
𝛽 6
𝐷𝑒𝑣𝑒𝑙𝑜𝑝𝑒𝑑 𝑖 + 𝛽 7
𝑆 ℎ𝑜𝑟𝑡 𝐷𝑢𝑟 . 𝐼𝑛𝑑𝑢𝑠𝑡𝑟𝑦 %
𝑖 +
𝛽 8
𝐿𝑜𝑛𝑔 𝐷𝑢𝑟 . 𝐼𝑛 𝑑𝑠𝑢𝑡𝑟𝑦 %
𝑖 + 𝛽 9
𝐿𝑎𝑟𝑔𝑒𝑠𝑡 𝐹𝑖𝑟𝑚 %
𝑖 + 𝜀 𝑖 (6)

Q5-Q1 Alpha is the CAPM Alpha differential computed from each country’s quintile
portfolios on several country characteristics including those related to Country-Year Duration
(standard deviation of this measure, SD), the inputs of duration (the autoregressive coefficients of
returns on equity and growth, ROE AR and G AR, and the terminal values for return on equity and
growth, ROE T and G T), and country characteristics (Developed, Short Dur. Industry %, Long
Dur. Industry %, and Largest Firm %). I find that countries with the highest standard deviation in
Country-Year Duration have the highest CAPM Alpha differential between their first and fifth
portfolios. This result holds across the inclusion of all controls variables, and with the alphas
generated from value weighted portfolios (Q5-Q1 Alpha VW) and from equal weighted portfolios
39

(Q5-Q1 Alpha EW). This result complements the result found in Goncalves (2021) where the short
duration premium in the U.S. is strongest in years in which the standard deviation of duration is
highest. This is a benefit of the international study, cross sectional tests between countries can
provide new results and updates to prior findings. Next in Panel B I regress SD from Panel A on
potential country-level determinants with the following equation:

𝑆𝐷
𝑖 = 𝛽 1
𝑅𝑂𝐸 𝐴𝑅
𝑖 + 𝛽 2
𝑅𝑂𝐸 𝑇 𝑖 + 𝛽 3
𝐺 𝐴𝑅
𝑖 + 𝛽 4
𝐺 𝑇 𝑖 + 𝛽 5
𝐷𝑒𝑣𝑒𝑙𝑜𝑝𝑒𝑑 𝑖 +
𝛽 6
𝑆 ℎ𝑜𝑟𝑡 𝐷𝑢𝑟 . 𝐼𝑛𝑑𝑢𝑠𝑡𝑟𝑦 %
𝑖 + 𝛽 7
𝐿𝑜𝑛𝑔 𝐷𝑢𝑟 . 𝐼𝑛𝑑𝑠𝑢𝑡𝑟𝑦 %
𝑖 + 𝛽 8
𝐿𝑎𝑟𝑔𝑒𝑠𝑡 𝐹𝑖𝑟𝑚 %
𝑖 + 𝜀 𝑖
(7)
I find that the country-level standard deviation (SD) is greatest in countries with higher
historic costs of capital (ROE T, which is equal to R), lower historic GDP growth rates (G T), and
lower percentages of firms in long duration industries (Long Dur. Industry %). While Largest Firm
% is significant in this regression, it does not retain its significance in untabulated univariate
regressions on SD as the other variables noted do.

6.3 Short Duration Premium Across Countries

Finally, In Table 13 I create Country-Year Duration by value-weighting the Firm-Year
Duration of each stock within each country-year based on their market capitalization. I create
quintile portfolios sorted on Country-Year Duration (each portfolio contains either 9 or 10
countries) and find that the CAPM Alpha differential between portfolio 5 and 1 is negative and
significant in an equal weighted specification, but only when the first quarter of 2020 is removed
from the sample. This potential exclusion is based on prior research which found this quarter to be
one in which the short duration premium is reversed (Dechow et al. 2021). The MKT Beta for these
40

portfolios is increasing in Country-Year Duration when the market return is the MSCI ACWI
return. This result demonstrates that equity duration can be a useful risk factor for investors not
only at the firm-level, but also at the country-level. Country-Year Duration could be another
characteristic added to MSCI’s Index Factsheet for each of their country-level index funds.
Figures 7 and 8 provide a visualization of this result. Figure 7 represents the Excess Return
of the portfolio holding the indices of the ten longest duration countries in blue and the portfolio
holding the indices of the ten shortest duration countries in red during the first quarter of 2020.
This was an unusual period due to the start of the COVID-19 pandemic and the global shutdown
that followed. This event fits into the classification of a disaster followed by a quick recovery
discussed in Hasler and Marfe (2016). In these situations, short-term cash flows are riskier than
long-term cash flows. As such, this was one of the few periods where the short duration premium
was reversed, that is short duration equities had less Excess Returns than did long duration equities
(including country level index based equities). The halting of cash flows impacts short duration
firms and short duration countries more because their market values are more heavily weighted
towards near term cash flows to shareholders. Figure 8 presents the visualization of a period (2022)
when the short duration premium was stronger than usual due to rising interest rates globally.
16
As
interest rates rise, long duration equities are more negatively impacted than short duration equities
as more of their market value is weighted on cash flows to shareholders in the far future. In
environments with higher discount rates, those far off cash flows will be discounted more heavily.
During 2022, the spread in Excess Returns between the portfolio holding the 10 shortest duration
country indices and the portfolio holding the 10 longest duration country indices was
approximately 18% compared to the 4.3% annual average spread between 2000 and 2021 reported

16
This period falls outside of the main sample of this paper.
41

in Table 13 (i.e., -0.36% monthly spread). These figures show that understanding the Country-
Year Duration could be important information for asset managers that hold country-level indexes
as it represents a relevant risk factor that varies cross sectionally across countries. Overall these
results support my prediction that the short duration premium holds not only at the firm-year level
within countries, but also at the country-year level across countries.

42

Chapter 7

Robustness Tests of Alternative Discount Rates and Forecasting
Assumptions

One potential concern is that the discount rate used in computing Firm-Year Duration is
time and firm invariant (i.e., it is fixed at the country-level for the full sample period). The discount
rate used in forecasting cash flows to investors significantly influences the equity duration
computed for each firm-year. If the actual discount rate applied by market participants changes
over time, or has variation within countries, then my measure of equity duration will be less
accurate. However, with these invariants, the results will at least be consistent within countries.
To address this concern, I follow a process similar to Goncalves (2021) to compute a discount rate
at the firm-year level. To do this, I set the actual market value of equity equal to the forecasted
stream of 15-years of cash flows to shareholders plus the forecasted terminal value assumption. I
then use a root-finding algorithm to solve for each firm-year discount rate. To compare these
results to the rates I imposed, I create one new country level discount rate based on the equal
weighted firm-year rates and another based on the value weighted firm-year rates. The equal
weighted measure has a correlation of .96 and a rank correlation of .85 with my currently used
43

country-level discount rate. The value weighted measure has a correlation of .92 and a rank
correlation of .90 with my currently used country-level discount rate.
17
This result helps mitigate
concerns that the primary results in Table 5, 7, and 9 are due to my assumption of a time and firm
invariant discount rates.
A second concern is that the results in Tables 5, 7, and 9 are driven by my country-level
imposed forecasting assumptions. If the sources that I used to set the forecasting assumptions (e.g.,
G T is set equal to the median GDP growth rate for the country) for computing Firm-Year Duration
are to similar to the sources used by the World Bank to create the Governance (WGI) measures,
then my results will be biased towards finding a positive relationship between the two variables.
To address this concern I recompute Country-Year Duration using only firm-specific information.
To do this I set ROE AR, G AR, and G T equal to their global average values (i.e., 0.51, 0.14, and
0.03 respectively) and set ROE T and R equal to the discount rate obtained from the firm-year
specific root solving process described above. This contrasts with the primary computation of
Firm-Year Duration where country specific values are used for ROE AR, G AR, ROE T, G T, and
R.
I then compute the equity duration for every firm-year in my sample and create a country-
year specific measure that is an equal weighted combination of all of the firm-years in that country-
year. This new measure of country-year duration is labeled C-Y Duration Implied R because it
relies on the implicit discount rate solved for each firm-year. In Table 14 Columns 1-3 I replicate
the tests from Column 2 of Tables 5, 7, and 9, but replace the dependent variable in each regression
with C-Y Duration Implied R. I find that the coefficients on Governance (WGI), Tax Loss

17
Prior to computing the new country-level measure of the discount rate, I winsorize the firm-year discount rates at
the 1
st
and 99
th
percentiles.
44

Carryforward, and Innovation Box remain positive and statistically significant. This helps mitigate
this second concern.
One final concern is that having any variation in the discount rate, whether set at county
level or firm-year level, will create bias in the results as the actual discount rate can not be
computed with perfect accuracy. To assuage this concern, I next create another measure of firm-
year duration with even more restrictive forecasting assumptions to ensure that any differences in
firm-year duration are the result of only timing differences in the forecasted cash flows and are not
due to differences in forecasting assumptions or discount rates.
18
To do this I set ROE AR, G AR,
ROE T, and G T equal to their global average values (i.e., 0.51, 0.14, 0.09, and 0.03 respectively)
and set R equal to 0%. Without a discount rate I can no longer assume a level perpetuity for the
terminal assumption, and I instead assume that any remaining cash flows to shareholders implied
by the market price and paid out at the end of the 15 year finite forecasting period. I then compute
the equity duration for every firm-year in my sample and create a country-year specific measure
that is an equal weighted combination of all of the firm-years in that country-year. This new
measure of country-year duration is labeled C-Y Duration Fixed FA because all of the forecasting
assumptions, including the discount rate, are fixed to equal values for each firm-year when
computing duration. In Table 14 Columns 4-6 I replicate the tests from Column 2 of Tables 5, 7,
and 9 but replace the dependent variable in each regression with C-Y Duration Fixed FA. I find
that the coefficients on Governance (WGI), Tax Loss Carryforward, and Innovation Box remain
positive, but only the coefficient on Tax Loss Carryforward and Innovation Box remain
statistically significant.

18
The potential issues with using discount rates and market values when computing equity duration are discussed in
Walter and Weber (2023).
45

In summary, the results of my robustness tests show that potential concerns about my
forecasting assumptions being fixed at the country level, instead of being set at the firm-year level
or set globally, does not change my primary inferences and my predict results still hold.
Additionally, when I compute discount rates at the firm-year level, the average discount rate at the
country level is very highly correlated with the county level discount assumption in my primary
test. Collectively these results show that the predicted relationship between a country’s governance
and tax policy is robust to alternative forecasting and discount assumptions used in computing
equity duration at the firm-level. Therefore, I report the results in Table 5, 7, and 9 where Country-
Year Duration is computed using country level forecasting assumptions and discount rates as that
is a compromise between more specific firm-year assumptions and more macro globally set
assumptions.

46

Chapter 8

Additional Analysis

8.1 Components of the World Governance Indicators

The results in Table 5 show a positive relationship between the composite Governance
(WGI) measure and Country-Year Duration. However, it is important for government officials and
global organizations to understand which components of governance in the composite measure are
driving the result. This is important since if only one of the factors is related to equity duration,
then policies can be directed at that factor alone versus if all four are significant. Table 15 tests the
following regression:

𝐶𝑜𝑢𝑛𝑡𝑟𝑦 − 𝑌𝑒𝑎𝑟 𝐷𝑢𝑟𝑎𝑡𝑖𝑜𝑛 𝑖 ,𝑡 = 𝛽 1
𝐺𝑜𝑣𝑒𝑟𝑛𝑎𝑛𝑐𝑒 𝑖 ,𝑡 + 𝜃𝐶𝑜𝑛𝑡𝑟𝑜𝑙𝑠 𝑖 ,𝑡 + 𝜀 𝑖 ,𝑡 (8)

Table 15 repeats the tests in Table 5, but the primary independent variable is switched from
the composite measure (i.e., Governance (WGI)) to a placeholder (i.e., Governance) for one the
four individual governance measures that make it up, plus the fifth and sixth measures, Political
Stability and Absence of Violence/Terrorism (PV) and Voice and Accountability (VA), that were
47

excluded from the composite. I find positive and significant results for the four primary measures
of governance in the univariate tests (Panel A) and in the tests with controls (Panel B). In the
collapsed sample (Panel C) where each country has only one observation representing its average
values over the sample period, all four of the governance measures have a positive and significant
coefficient with the exception of Regulatory Quality (RQ). In each of the three specifications,
Government Effectiveness (GE) has the largest coefficient and highest R
2
. This indicates that
having an effective government that produces high quality public services and that can create and
implement new policies as necessary is the most influential governance factor tested for providing
an operating environment that is suitable for long duration investments.
I excluded Voice and Accountability (VA) from my Governance (WGI) composite measure
as I did not have a clear prediction for if allowing citizens more of a voice and vote on their form
of government and government officials would lead to better and more stable environment for long
term investments. VA is the only one of the governance measures where the coefficient flips signs
across the tests. VA is the measure that is the least correlated with the other five measures of
governance from the World Bank. It is the only measure of the six where the two countries with
the lowest score (i.e., Saudi Arabia and China) are both in the top half of the sample for Country-
Year Duration. Collectively, the relationship between VA and Country-Year Duration is not clear.
I also excluded Political Stability and Absence of Violence/Terrorism (PV) from the composite
measure to be consistent with prior research. PV has a positive and significant relationship with
Country-Year Duration in each of the 3 panels of Table 15. This indicates that more stable regimes
that experience less politically motivated violence are home to longer duration equities on average.
In summary, Government Effectiveness (GE) is the primary driver of the predicted
relationship between governance and duration. This indicates that having an effective government
48

that produces high quality public services and that can create and implement new policies as
necessary is the most influential governance factor tested for providing an operating environment
that is suitable for long duration investments.

8.2 Economic Policy Uncertainty Index

The strength of governance and the stability of the government in a country may not fully
explain the stability of the business operating environment, not even from a governmental aspect.
It is possible for a country to be strong, retain their form of government, and be able to enforce
laws and regulations fairly, but if the laws and regulations are subject to change every few years,
then they may not be a suitable environment for long duration investments. I predict that countries
with more uncertainty surrounding their future economic policies will on average have a higher
proportion of low duration firms headquartered in it. To analysis the stability of economic policy
from a manager’s perspective, I consider an additional measure called the Economic Policy
Uncertainty (EPU) index. The EPU index has been used in prior accounting research to study its
effects on information asymmetry and disclosure (Nagar, Schoenfeld, and Wellman 2019), its
ability to forecast excess returns (Brogaard and Detzel 2015), and its relationship to innovation
within a country (Cong and Howell 2021). The measure is constructed by analyzing news coverage
from major newspapers in each country to observe how often terms for economics, uncertainty,
and policy are mentioned in the same article. The data comes from: Baker, Bloom and Davis (2016)
for Australia, Brazil, Canada, France, Germany, India, Italy, Mexico, South Korea, Russia, United
Kingdom, and the United States; Cerda, Silva and Valente (2016) for Chile; Baker, Bloom, Davis
and Wang (2013) for China; Gil and Silva (2018) for Colombia; Hardouvelis, Karalas,
Karanastasis and Samartzis (2018) for Greece; Zalla (2016) for Ireland; Saxegaard, Davis, Ito and
49

Miake (2019) for Japan; Kroese, Kok and Parlevliet (2015) for the Netherlands; Davis (2016) for
Singapore; Ghirelli, Perez, and Urtasun (2019) for Spain; and Armelius, Hull, and Köhler (2017)
for Sweden. The datasets have been compiled on a website hosted by the authors of Baker, Bloom
and Davis (2016).
19

Unfortunately, this data set is only available for the full sample period for these 22 countries
out of the 49 in my sample which may limit the generalizability. Additionally, the subsample of
countries that EPU data is available skews towards Developed countries (i.e., 47% of the countries
in the full sample are Developed whereas 59% of countries with EPU data available are
Developed). I predict that as EPU increases, Country-Year Duration will be shorter on available
because having uncertainty in future economic policy will reduce managers’ willingness to invest
long term. Prior to performing any tests, I transform EPU by taking the natural log of the variable
as the measure is bounded by 0 on the left but is unbounded and right skewed on the other end.
Table 16 tests the following regression:

𝐶𝑜𝑢𝑛𝑡𝑟𝑦 − 𝑌𝑒𝑎𝑟 𝐷𝑢𝑟 𝑎𝑡𝑖𝑜𝑛 𝑖 ,𝑡 = 𝛽 1
𝐸𝑐𝑜𝑛𝑜𝑚𝑖𝑐 𝑃𝑜𝑙𝑖𝑐𝑦 𝑈𝑛𝑐𝑒𝑟𝑡𝑎𝑖𝑛𝑡𝑦 (𝐸𝑃𝑈 )
𝑖 ,𝑡 +
𝜃𝐶𝑜𝑛𝑡𝑟𝑜𝑙𝑠 𝑖 ,𝑡 + 𝜀 𝑖 ,𝑡 (9)

I predict a negative coefficient on EPU in Table 16 in Columns 1 and 2, but the results are
only negative and significant in Column 2 when the controls are in place. In Columns 3 and 4 I
limit the sample further to the 12 countries from Baker, Bloom and Davis (2016) so that the EPU
variable comes from a single source and single set of co-authors. The coefficient on EPU in
Column 3 is positive and significant because some developed countries with a high percentage of

19
https://www.policyuncertainty.com/all_country_data.html
50

long duration firms, score highly on the EPU index, such as the United Kingdom in the years
following Brexit. The United States also has a high score of EPU due to slim majorities in congress
and a shifting of power and control between parties multiple times over the sample period. There
is minimal overlap between the EPU and the WGI measures, the maximum correlation between
any of the 6 WGI measures and EPU is 0.06 for GE and -0.09 for VA (e.g., the United Kingdom
is scored as having strong governance for both of those measures). Within the United Kingdom,
recent research has shown that as the EPU increased at the country level, innovation decreased at
the firm level within the country (Nguyen and Trinh 2023). The results in Table 16 support the
prediction that EPU is another dimension of country characteristics that may be considered by
managers when making decisions on investment duration horizons, and that environments with
more uncertainty regarding economic policies are less favorable to long duration investments.

8.3 Hofstede's National Cultural Dimensions

Another important cross-sectional difference in countries, in addition to the strength of
their governance and the availability of future focused tax policy, is their culture. I predict that
cultures that are more comfortable with uncertainty and that have a more long-term perspective
will be cultures that support out long duration investments. Geert Hofstede's six dimensions of
national culture represent how a society organizes itself and they describe a society’s culture and
its citizens’ values. Hofstede developed his original model through a global survey of IBM
employees between 1967 and 1973. He has continued to develop the model, expanded country
coverage, and has added additional dimensions of culture (Hofstede 2001; Hofstede et al. 2010;
and Hofstede 2013; collectively, Hofstede). For my tests, I am using the data available on his
51

website, version 2015 12 08.
20
There is only one value available per country per dimension, but
that single value should be applicable for my entire sample period because “it is useful to think of
dimensions of culture as changing only slowly, from generation to generation.”
21

The culture dimensions have been used in prior accounting research to study the influence
of culture on accounting conservatism and risk-taking behavior (Kanagaretnam, Lim, and Lobo
2014), how managers from different cultures respond to macroeconomic uncertainty (Binz 2022),
how risk preferences vary around the world (Rieger, Wang, and Hens 2015), the impact of culture
on the pricing of extreme positive returns (Cheon and Lee 2018), and where decision rights are
held in multinational companies (Robinson & Stocken 2013). The original four dimensions were:
1. Power Distance (PDI), 2. Individualism (IDV), 3. Masculinity (MAS), and 4. Uncertainty
Avoidance (UAI). The last two cultural dimensions were not included in his original work, but
were later developed, they are: 5. Long-term Orientation (LTOWVS) and 6. Indulgence (IVR). Of
these six dimensions, I predict that UAI and LTOWVS will have the most direct relationship to the
average equity duration of the firms in a country. Table 17 tests the following regression where
Hofstede is a placeholder for one the six individual cultural dimensions:

𝐶𝑜𝑢𝑛𝑡𝑟𝑦 − 𝑌𝑒𝑎𝑟 𝐷𝑢𝑟𝑎𝑡𝑖𝑜𝑛 𝑖 ,𝑡 = 𝛽 1
𝐻𝑜𝑓𝑠𝑡𝑒𝑑𝑒 𝑖 ,𝑡 + 𝜃𝐶𝑜𝑛𝑡𝑟𝑜𝑙𝑠 𝑖 ,𝑡 + 𝜀 𝑖 ,𝑡 (10)

Power Distance (PDI) represents the extent to which less powerful members of a society
are willing to accept that power is distributed unequally (Hofstede). A low score indicates that
people question authority and attempt to distribute power (e.g., Austria and Demark), a high score
indicates that a hierarchy is clearly established (e.g., Russia and Qatar). I do not predict, and I do

20
https://geerthofstede.com/research-and-vsm/dimension-data-matrix/
21
https://geerthofstede.com/research-and-vsm/
52

not find, a strong relationship between PDI and Country-Year Duration because a low PDI culture
may give more people access to power to create innovations, but a high PDI may result in a more
stable society as citizens accept their leaders and their power. In Table 17, I find a negative and
significant relationship between PDI and Country-Year Duration in the univariate tests, but no
significant result after control variables are considered.
The spectrum related to individualism versus collectivism (IDV) describes the size and
strength of in-groups among a country’s citizens (Hofstede). In my sample, the United States is
the most individualistic (higher on scale) and Colombia is the most collectivist (lower on scale). I
do not predict, and I do not find, a strong relationship between IDV and Country-Year Duration
because it is not clear that individuals making choices compared to societies making choices will
result in an environment more suitable for long duration projects. In Table 17, I find a positive and
significant relationship between IDV and Country-Year Duration in the univariate tests, but no
significant result after control variables are considered.
Cultures that are most masculine (MAS) are defined as those in which gender roles are more
clearly distinct and competition is encouraged. On the other end of the index, in feminine societies
there is more cooperation, overlap in gender roles, and a focus on quality of life (Hofstede). Japan
and Hungary have two of the highest (masculine) scores in the sample while Norway and Sweden
have two of the lowest (feminine) scores in the sample. I do not predict, and I do not find, a strong
relationship between MAS and Country-Year Duration because it is not clear that a more
competitive and more gendered society will result in an environment more suitable for long
duration projects compared to a more caring and cooperatives society. In Table 17, I find a negative
and significant relationship between MAS and Country-Year Duration in the univariate tests, but
no significant result after control variables are considered.
53

Uncertainty avoidance (UAI) describes a culture’s tolerance for ambiguity. Do the citizens
feel anxiety and become distrustful when faced with the unknown, and do they want fixed habits
and routines (Hofstede)? Singapore scores the lowest on this measure and Greece scores the
highest. Serafeim (2015) shows that Greece scores the lowest on Innovation (European
Commission Index) and the highest on Bureaucracy (World Bank Ranking) of European Union
countries and he ties this to their high UAI score because innovation requires risk taking. There
will always be ambiguity with new products and new markets. Additionally, the higher levels of
bureaucracy are related to a desire for more structured and standardized practices. This may help
Greeks avoid uncertainty, but it increases the time spent on registrations, permitting, and general
paperwork that could have instead been spent on building new businesses. As such, I predict a
negative relationship between UAI and Country-Year Duration. I find a negative and significant
coefficient on UAI in all three specifications tested in Table 17. This indicates that countries with
higher uncertainty avoidance are home to shorter duration companies, on average, with all else
held constant.
Long-Term Orientation versus Short-Term Orientation (LTOWVS) describes not only how
a society prepares for the future, but also how they relate to their past history. A short-term oriented
culture values traditions and looks to the past for morals (Hofstede). The United States scores low
on this measure (i.e., labeled as short-term oriented) partially due to it having a relatively old
government and reliance on a constitution from the 18
th
century. China and South Korea are two
of the most long-term oriented countries in the sample. This indicates that they value being
prepared for the future, and that they are more willing to adapt and make pragmatic decisions. I
predict that there is a positive relationship between LTOWVS and Country-Year Duration. I find a
positive and significant result in the univariate test (Table 17 Panel A), but a negative and
54

significant result in the test with controls (Table 17 Panel B), and an insignificant result in the test
with the collapsed sample with only one observation per firm year (Table 17 Panel C). LTOWVS
has a correlation coefficient with High Dur. Industry % of 0.495 which indicates that countries
with long-term orientations have more of their market represented by long duration industries (e.g.,
computer software and hardware) compared to short-term oriented countries. This is the only one
of the six cultural dimensions that has a coefficient that flips signs across the tests. Overall, there
does not appear to be a strong relationship between the time orientation of a country’s culture and
their average firm duration.
Finally, indulgence versus restraint (IVR) is the newest dimension. This dimension relates
to freedom and the ability to follow your impulses on the high end, and duty and strict social norms
for the low end (Hofstede). In the sample, Mexico and Sweden have two of the highest scores and
Pakistan and Egypt have two of the lowest scores. I do not have a prediction on the relationship
between IVR and Country-Year Duration. The freedom to follow one’s impulses could lead to new
inventions and innovations, but the mindset of delayed gratification that comes with restraint may
allow for the sacrifices today that pay off in the long term. Therefore, it is not clear which
environment is more suitable for long duration projects. I find a positive and significant coefficient
on IVR in all three panels of Table 17. This is a strong result which indicates that indulgent
countries are ones that have longer duration firms, on average, with all else held equal.
In summary, only two of Hofstede’s cultural dimensions, which are based on survey data,
are related to a country’s average duration. There’s support for the prediction that cultures that are
more comfortable with uncertainty are more likely to have longer duration equities on average.
However, the cultures with longer-term orientations are not home to longer duration equities on
average because some countries that value tradition (i.e., shorter-term orientations) are home to a
55

large proportion of long duration equities as well. Finally, indulgent cultures are the second type
of culture that has a positive relationship with duration length, although this was not a predicted
result. This indicates that the freedom to follow impulses is related to preferred investment
horizons.

8.4 GDP Per Capita

Wealthier countries may have less of a need for immediate cash inflows. Therefore
I predict that managers in wealthier countries feel more confident in making investment decisions
that require a cash outflow today with the hopes of producing a large cash inflow in the future (i.e.,
delayed gratification). In Table 18 I test this prediction through the relationship between the natural
log of the gross domestic product per capita (GDP Per Capita) and Country-Year Duration. Table
18 tests the following regression:

𝐶𝑜𝑢𝑛𝑡𝑟𝑦 − 𝑌𝑒𝑎𝑟 𝐷𝑢𝑟𝑎𝑡𝑖𝑜𝑛 𝑖 ,𝑡 = 𝛽 1
𝐺𝐷𝑃 𝑃𝑒𝑟 𝐶𝑎𝑝𝑖𝑡𝑎 𝑖 ,𝑡 + 𝜃𝐶𝑜𝑛𝑡𝑟𝑜𝑙𝑠 𝑖 ,𝑡 + 𝜀 𝑖 ,𝑡 (11)

The results show that wealthier countries have longer duration stocks on average, even
after controlling for development status, relative inflation, and the presence of long and short
duration industries. Additionally, in untabulated analysis, I replace Developed with GDP Per
Capita in Tables 5, 7, and 9 and the primary results hold. The Developed variable is related to
financial market development while GDP Per Capita is more closely related to economic
development. In summary wealthier countries have a higher proportion of long duration firms than
do poorer countries. This is relationship is predicted to come from the luxury of being able to delay
gratification if there is not an immediate need for cash flows.
56

Chapter 9

Conclusion

This dissertation examines implied equity duration in a global setting. Implied equity
duration is a measure used to capture the average maturity (in years) of cash flows to shareholders
associated with a given stock. These cash flows take the form of dividends, stock repurchases, and
stock issuances. A long duration equity is one where most of its market value is based on the
forecasted expectation of a return of cash flows to shareholders in the long-term. I predict that
country characteristics such as the strength of a country’s governance and the availability of future
focused tax policies will impact equity duration because these environments encourage long
horizon investments.
I find that countries with strong governance, as defined by the Worldwide Governance
Indicators project, have longer equity durations in their mean and median firm. In particular, the
governance indicator related to the quality of policy formulation and implementation, and the
credibility of the government's commitment to such policies, has the strongest relationship with a
country’s average duration of the governance measures tested. Firms and managers in less
57

developed countries and in countries with weaker governance may need to focus on their short-
term cash flows from a practical perspective.
I also investigate whether long duration firms are more likely to be headquartered in
countries with generous tax loss carryforward rules and in countries that offer innovation box tax
incentives. I predict and find that countries with longer tax loss carryforward period are home to
longer duration equities on average. These carryforward rules allow for risk sharing between the
firm and the government. Longer carryforward periods allow for firms to continue to benefit from
this tax policy when pursuing investments that will produce a payout in the distant future and are
therefore likely to be longer duration investments. I also predict and find that countries with
innovation box tax incentives are home to longer duration equities on average. Innovation boxes
are policies that split a firm’s income into two buckets, one for innovation related income (e.g.,
income related to patents) that is taxed at a lower rate and another for all other income which is
then taxed at the statutory rate. These policies encourage the invention of new innovations that
will generate income far into the future (i.e., longer duration) as the tax benefits will generally
continue as long as the innovation produces income.
After establishing the determinants of cross country variation in duration, I next shift focus
to an asset manager’s perspective and consider the short duration premium within and across
countries. I show that the short duration premium holds across the majority of developed and
emerging countries. This result is documented at the market portfolio level (i.e., Developed and
Emerging), the regional portfolio level (e.g., Europe and North America), and at the individual
country level. The short duration premium exists not only within these groups, but also across
countries. I show that it is possible to sort countries into portfolios based on their duration, and
58

that these portfolios have CAPM alphas that decrease in duration when MSCI country-index
returns are used for country-level returns and the MSCI ACWI is used for the market return.
Many funds invest around the world so understanding risks across and within countries is
key. Understanding where the short duration premium exists is important, but it is just as important
to understand in which settings the premium is strongest. I find that the short duration premium is
strongest in countries with high standard deviations in their country level duration. This country-
level standard deviation is greatest in countries with higher historic costs of capital, lower historic
GDP growth rates, and lower percentages of firms in long duration industries.
The implied equity duration measure has been the subject of a growing area of research in
recent years not only for academics, but also for practitioners. Implied equity duration could
potentially be used in the Environmental, Social, and Governance (ESG) literature to test if firms
with a longer duration are more involved in sustainable practices as they depend on having a safe
and just operating environment in the distant future when their positive cash flows come in. While
this paper focuses on equity durations around the world for recent years, research could be done
to test the historic trends and changes in a country’s average duration over time and as their level
of development changes. For example, in the United States, as high-growth tech stocks are
increasingly representing more of the total value of the stock market, the United States equities
market is becoming a longer duration asset. This is important as cash flows from these companies
are more sensitive to interest rate movements which have frequently occurred in recent years.
My findings have implications for asset managers, executives, governments, and global
organizations. Asset managers should be aware of in which countries the short duration premium
is strongest. They may also wish to consider equity duration as another measure when diversifying
their portfolio. Executives may consider their company's goals before expanding their business
59

into a new country that has an average firm equity duration that is significantly shorter or longer
than theirs. Governments may consider ways to strengthen their governance measures and weigh
the costs and benefits of changes to tax policy if they wish to attract and encourage the
development of innovative and long duration firms. Finally, global organization such as the United
Nations, World Economic Forum, and World Bank may find this information useful to determine
which countries need the most help and are more financially impacted during different challenging
periods such as the global shutdown at the start of the COVID-19 pandemic or the raising interest
rate regimes in recent years.
60

References

Armelius, H., Hull, I. and Köhler, H.S., 2017. The timing of uncertainty shocks in a small open
economy. Economics Letters, 155, pp.31-34.

Baker, S.R., Bloom, N., Davis, S.J. and Wang, X., 2013. Economic policy uncertainty in China.
unpublished paper, University of Chicago.

Baker, S.R., Bloom, N. and Davis, S.J., 2016. Measuring economic policy uncertainty. The
Quarterly Journal of Economics, 131(4), pp.1593-1636.

Binz, O., 2022. Managerial response to macroeconomic uncertainty: implications for firm
profitability. The Accounting Review, 97(5), pp.89-117.

Brogaard, J. and Detzel, A., 2015. The asset-pricing implications of government economic policy
uncertainty. Management Science, 61(1), pp.3-18.

Cerda, R., Silva, A. and Valente, J.T., 2016. Economic policy uncertainty indices for Chile.
Economic Policy Uncertainty working paper.

Chen, S., De Simone, L., Hanlon, M. and Lester, R., 2019. The effect of innovation box regimes
on income shifting and real activity. Available at SSRN, 3486428.

Chen, S. and Li, T., 2022. A unified duration-based explanation of the value, profitability, and
investment anomalies. Available at SSRN, 3290191.

Chen, Z., 2022. Inferring Stock Duration Around FOMC Surprises: Estimates and Implications.
Journal of Financial and Quantitative Analysis, 57(2), pp.669-703.

Cheon, Y.H. and Lee, K.H., 2018. Maxing out globally: Individualism, investor attention, and the
cross section of expected stock returns. Management Science, 64(12), pp.5807-5831.

Cong, L.W. and Howell, S.T., 2021. Policy uncertainty and innovation: Evidence from initial
public offering interventions in China. Management Science, 67(11), pp.7238-7261.
61

Dai, R., 2017. International accounting databases on wrds: Comparative analysis. Available at
SSRN 2938675.

Damodaran, A., 2020a. Equity Risk Premiums: Determinants, Estimation and Implications - The
2020 Edition. NYU Stern School of Business, Available at SSRN 3550293.

Damodaran, A., 2020b. Country Risk: Determinants, Measures and Implications – The 2020
Edition. NYU Stern School of Business, Available at SSRN 3653512.

Davis, S.J., 2016. An index of global economic policy uncertainty (No. w22740). National Bureau
of Economic Research.

Dechow, P.M., Erhard, R., Sloan, R.G. and Soliman, M.T., 2021. Implied equity duration: A
measure of pandemic shut-down risk. Journal of Accounting Research, 59: 243-281.

Dechow, P.M., Sloan, R.G. and Soliman, M.T., 2004. Implied equity duration: A new measure of
equity risk. Review of Accounting Studies, 9(2), pp.197-228.

Ellahie, A. and Kaplan, Z., 2021. Show Me the Money! Dividend Policy in Countries with Weak
Institutions. Journal of Accounting Research, 59(2), pp.613-655.

Ernst & Young, 2004-2020. Worldwide Corporate Tax Guide. https://www.ey.com/en_gl/tax-
guides

Fama, E.F. and French, K.R., 1993. Common risk factors in the returns on stocks and bonds.
Journal of Financial Economics, 33(1), pp.3-56.

Fama, E.F. and French, K.R., 2012. Size, value, and momentum in international stock returns.
Journal of Financial Economics, 105(3), pp.457-472.

Fama, E.F. and MacBeth, J.D., 1973. Risk, return, and equilibrium: Empirical tests. Journal of
Political Economy, 81(3), pp.607-636.

Ghirelli, C., Pérez, J.J. and Urtasun, A., 2019. A new economic policy uncertainty index for Spain.
Economics Letters, 182, pp.64-67.

Gil, M. and Silva, D., 2018. Economic policy uncertainty indices for Colombia. Working paper
available at https://www.policyuncertainty.com/media/EPU_Colombia_Paper_GS.pdf.

Gonçalves, A.S., 2021. The short duration premium. Journal of Financial Economics, 141(3),
pp.919-945.

Gonçalves, A.S. and Leonard, G., 2023. The fundamental-to-market ratio and the value premium
decline. Journal of Financial Economics, 147(2), pp.382-405.

Gormsen, N.J. and Lazarus, E., 2023. Duration-driven returns. Journal of Finance.
62

Gow, I.D., Ormazabal, G. and Taylor, D.J., 2010. Correcting for cross‐sectional and time‐series
dependence in accounting research. The Accounting Review, 85(2), pp.483-512.

Hardouvelis, G.A., Karalas, G., Karanastasis, D. and Samartzis, P., 2018. Economic policy
uncertainty, political uncertainty and the Greek economic crisis. Political Uncertainty and the
Greek Economic Crisis (April 3, 2018).

Hasler, M. and Marfe, R., 2016. Disaster recovery and the term structure of dividend strips. Journal
of Financial Economics, 122(1), pp.116-134.

Hofstede, G., 2001. Culture’s consequences: comparing values, behaviors, institutions, and
organizations across nations. 2nd ed. Sage Publications.

Hofstede, G., Hofstede, G.J. and Minkov, M., 2010. Cultures and organizations: software of the
mind: international cooperation and its importance for survival. 3rd ed. McGraw-Hill.

Hofstede, G., 2013. Replicating and Extending Cross-National Value Studies: Rewards and
Pitfalls–An Example Middle East Studies. AIB Insights, 13(2), pp.5-7.

Hou, K., Xue, C. and Zhang, L., 2020. Replicating anomalies. The Review of Financial Studies,
33(5), pp.2019-2133.

Kanagaretnam, K., Lim, C.Y. and Lobo, G.J., 2014. Influence of national culture on accounting
conservatism and risk-taking in the banking industry. The Accounting Review, 89(3), pp.1115-
1149.

Kapons, M.M., Kelly, P., Stoumbos, R.C. and Zambrana, R., 2022. Dividends, Trust, and Firm
Value. Available at SSRN 2311512.

Kaufmann, D., Kraay, A. and Mastruzzi, M., 2010. The worldwide governance indicators:
Methodology and analytical issues. World Bank Policy Research working paper, Available at
SSRN 1682130.

Kleintop, J., 2022. Hedging Stocks Against Rising Rates.
https://www.schwab.com/learn/story/hedging-stocks-against-rising-rates

Kroese, L., Kok, S. and Parlevliet, J., 2015. Beleidsonzekerheid in nederland. Economisch
Statistiche Berichten, 4715, pp.464-467.

Langenmayr, D. and Lester, R., 2018. Taxation and corporate risk-taking. The Accounting Review,
93(3), pp.237-266.

La Porta, R., Lopez‐de‐Silanes, F., Shleifer, A. and Vishny, R.W., 1997. Legal determinants of
external finance. The Journal of Finance, 52(3), pp.1131-1150.

63

La Porta, R., Lopez‐de‐Silanes, F., Shleifer, A. and Vishny, R.W., 1998. Law and finance. Journal
of Political Economy, 106(6), pp.1113-1155.

Leff, N.H., 1964. Economic development through bureaucratic corruption. American Behavioral
Scientist, 8(3), pp.8-14.

Li, Q., Ma, M.S. and Shevlin, T., 2021. The effect of tax avoidance crackdown on corporate
innovation. Journal of Accounting and Economics, 71(2-3), p.101382.

Ljungqvist, A., Zhang, L. and Zuo, L., 2017. Sharing risk with the government: How taxes affect
corporate risk taking. Journal of Accounting Research, 55(3), pp.669-707.

Macaulay, F.R., 1938. Some theoretical problems suggested by the movements of interest rates,
bond yields and stock prices in the United States since 1856. NBER Books.

Montagna, D.M. and Bianchi, L., 2022. Equity Duration: Theoretical and Practical Analysis.
Available at SSRN 4233635.

Mukherjee, A., Singh, M. and Žaldokas, A., 2017. Do corporate taxes hinder innovation? Journal
of Financial Economics, 124(1), pp.195-221.

Nagar, V., Schoenfeld, J. and Wellman, L., 2019. The effect of economic policy uncertainty on
investor information asymmetry and management disclosures. Journal of Accounting and
Economics, 67(1), pp.36-57.

Newey, W.K. and West, K.D., 1987. A Simple, Positive Semi-Definite, Heteroskedasticity and
Autocorrelation. Econometrica, 55(3), pp.703-708.

Nguyen, M.H. and Trinh, V.Q., 2023. UK economic policy uncertainty and innovation activities:
A firm-level analysis. Journal of Economics and Business, 123, p.106093.

Nissim, D. and Penman, S.H., 2001. Ratio analysis and equity valuation: From research to practice.
Review of Accounting Studies, 6(1), pp.109-154.

Ozdagli, A.K., 2018. Financial frictions and the stock price reaction to monetary policy. The
Review of Financial Studies, 31(10), pp.3895-3936.

Rieger, M.O., Wang, M. and Hens, T., 2015. Risk preferences around the world. Management
Science, 61(3), pp.637-648.

Robinson, L.A. and Stocken, P.C., 2013. Location of decision rights within multinational firms.
Journal of Accounting Research, 51(5), pp.1261-1297.

Saxegaard, E.C.A., Davis, S.J., Ito, A. and Miake, N., 2022. Policy uncertainty in Japan. Journal
of the Japanese and International Economies, 64, p.101192.

64

Serafeim, G., 2015. If Greece Embraces Uncertainty, Innovation Will Follow, Harvard Business
Review.

Silva-Gámez, A.I., Méndez-Prado, S.M. and Arauz, A., 2022. What’s Happening with the Patent
Box Regimes? A Systematic Review. Sustainability, 14(18), p.11423.

Walter, D. and Weber, R., 2023. Is There An Equity Duration Premium?. Available at SSRN
4157585.

Weber, M., 2018. Cash flow duration and the term structure of equity returns. Journal of Financial
Economics, 128(3), pp.486-503.

Ye, D., Deng, J., Liu, Y., Szewczyk, S.H. and Chen, X., 2019. Does board gender diversity increase
dividend payouts? Analysis of global evidence. Journal of Corporate Finance, 58, pp.1-26.

Zalla, R., 2016. Economic Policy Uncertainty in Ireland." Working paper available at
https://www.policyuncertainty.com/media/Zalla_EPU_Ireland.pdf.

65

Appendix A

Variable Definitions

Variable Definition Source

Duration Variables

Country-Year Duration First, Firm-Year Duration is computed for each
firm-year in the sample as defined below. Then,
within each country-year, the firm-year durations
are value-weighted (VW) based on the market
capitalization of each firm at the start of the year.
Firms are assigned to a country based on the
country they are headquartered in. In some
specifications, the country-year duration is
alternatively based on the equal weighting of the
relevant firm-years (EW), the median duration of
the relevant firm-years (P50), or the standard
deviation of duration for the relevant firm-years
(SD).
FactSet
Fundamentals,
https://pages.ste
rn.nyu.edu/~ada
modar/, and
https://data.worl
dbank.org

C-Y Duration Implied
R
Follows a similar computation process as Country-
Year Duration, except the ROE T and R are set
equal to the implied discount rate for each firm-
year. The implied discount rate is solved for using
a root solving algorithm where the market value is
equal to the sum of the forecasted cash flows to
shareholders. Additionally, ROE AR, G AR, and G
T are set equal to their global means.
FactSet
Fundamentals
and
https://data.worl
dbank.org

66

C-Y Duration Fixed FA Follows a similar computation process as Country-
Year Duration, except R is set equal to 0%.
Additionally, ROE AR, ROE T, G AR, and G T are
set equal to their global means.
FactSet
Fundamentals
and
https://data.worl
dbank.org

Firm-Year Duration Duration is computed by using the formula created
in Dechow et al. (2004). Cash flows to
shareholders (CF) are set equal to BEt-1*(ROEt-
Gt). Book value of equity (BE) is set equal to
FF_COM_EQ. ROE is forecasted and computed
as Earningst/BEt-1. Earnings are set equal to
FF_NET_INC. Gt is the forecasted growth rate for
book equity which is assumed to follow the first-
order autoregressive process of sales growth. Sales
are set equal to FF_SALES. Market capitalization
(ME) is set equal to FF_MKT_VAL. All amounts
are reported in U.S. dollars. The finite forecast
period (T) is set equal to 15 years. Cash flows
beyond the finite forecast period are assumed to be
paid out to investors as a level perpetuity (i.e., g =
0). R is the discount rate and is defined below.
FactSet
Fundamentals,
https://pages.ste
rn.nyu.edu/~ada
modar/, and
https://data.worl
dbank.org

ROE AR The first-order autoregressive coefficient of ROE
(as defined above). This coefficient is separately
computed for each country over the period
between 1995 and 2021.
FactSet
Fundamentals

ROE T The long-run mean of ROE is set equal to the
median cost of equity (i.e., R) for each country for
the period between 2000 and 2020 so that the ROE
will converge to R.
https://pages.ste
rn.nyu.edu/~ada
modar/

G AR The first-order autoregressive coefficient of sales
growth (i.e., (Salest / Salest-1) – 1). This coefficient
is separately computed for each country over the
period between 1995 and 2021.
FactSet
Fundamentals

G T The long-run mean of G is set equal to the median
GDP growth from 2000 to 2020 for each country.
https://data.worl
dbank.org

R The cost of equity is set equal to the Total Equity
Risk Premium plus the Riskfree Rate as defined by
Damodaran (2020a) and Damodaran (2020b).
https://pages.ste
rn.nyu.edu/~ada
modar/

67

Primary Independent Variables

Governance (WGI) An equal weighted combination of 4 of the World
Bank’s World Governance Indicators:
Government Effectiveness (GE), Regulatory
Quality (RQ), Rule of Law (RL), and Control of
Corruption (CC). The value ranges from 0 to 100
where 100 is the “best”. Note that this variable is
not available for 2001.
https://info.worl
dbank.org/gove
rnance/wgi/

Tax Loss Carryforward The number of years that a tax loss can be carried
forward to offset future taxable income. If
carryforward is indefinite, then variable is set to 30
if losses can offset ≥ 50% of current year income
or 25 if losses can offset < 50% of current year
income.
EY’s
Worldwide
Corporate Tax
Guide

Innovation Box A binary indicator variable equal to Yes (1) if the
country-year has innovation box tax incentives in
place (e.g., a lower tax rate is applied to a separate
schedule of firm produced innovation related
income), it is set equal to No (0) otherwise.
Chen et al.
(2019) and
Silva-Gámez et
al. (2022)

Control Variables

Developed A binary indicator variable set equal to 1 for the
developed countries and equal to 0 for emerging
countries (as of December 2019).
The developed countries are: Australia, Austria,
Belgium, Canada, Demark, Finland, France,
Germany, Hong Kong, Ireland, Israel, Italy, Japan,
Netherlands, New Zealand, Norway, Portugal,
Singapore, Spain, Sweden, Switzerland, United
Kingdom, and the United States.
The emerging countries are: Argentina, Brazil,
Chile, China, Colombia, Czech Republic, Egypt,
Greece, Hungary, India, Indonesia, Malaysia,
Mexico, Pakistan, Peru, Philippines, Poland,
Qatar, Russia, Saudi Arabia, South Africa, South
Korea, Taiwan, Thailand, Turkey, and United
Arab Emirates.
https://www.ms
ci.com/our-
solutions/indexe
s/market-
classification

Inflation The country-year level of inflation standardized
relative to inflation in the U.S. for that year.
https://data.worl
dbank.org

68

Short Dur. Industry % The percentage of firm-years in each country that
are in short duration industries (Categories 1 and 3
of the Fama and French 12 industries).
FactSet
Fundamentals

Long Dur. Industry % The percentage of firm-years that are in long
duration industries (Categories 6 and 10 of the
Fama and French 12 industries).
FactSet
Fundamentals

Largest Firm % The percentage of market value that the top 10%
of firms within a country own.
FactSet
Fundamentals

Portfolio Variables

Total Vol The standard deviation of monthly portfolio
returns.
FactSet
Fundamentals

IVol The standard deviation of the residual of the
model. The sub-tables captioned “CAPM” are
based on the regression Return-rf = α + β1(Market-
rf) + ε. The sub-tables captioned “3-Factor” are
based on the regression Return-rf = α + β1(Market-
rf) + β2(SMB) + β3(HML) + ε.
FactSet
Fundamentals

Excess Return The portfolio return in excess of the U.S. risk-free
rate.
FactSet
Fundamentals

CAPM Alpha The intercept from a regression of monthly
portfolio returns minus the risk-free rate on the
region's value-weighted market portfolio minus
the risk-free rate. The region’s market portfolio
return is taken from the market specific MSCI
index for the CAPM regressions. The sub-tables
captioned “CAPM” are based on the regression
Return-rf = α + β1(Market-rf) + ε.
FactSet
Fundamentals

FF Alpha The intercepts from a regression of monthly
portfolio returns minus the risk-free rate on the
region's value-weighted market portfolio minus
the risk-free rate. The region’s market portfolio
return is taken from Ken French’s website for the
3-Factor regressions. The sub-tables captioned “3-
Factor” are based on the regression Return-rf = α
+ β1(Market-rf) + β2(SMB) + β3(HML) + ε.
FactSet
Fundamentals
and
http://mba.tuck.
dartmouth.edu/
pages/faculty/ke
n.french/index.h
tml
69

MKT Beta Market Beta is the β1 slope estimate of the
regression. The sub-tables captioned “CAPM” are
based on the regression Return-rf = α + β1(Market-
rf) + ε. The sub-tables captioned “3-Factor” are
based on the regression Return-rf = α + β1(Market-
rf) + β2(SMB) + β3(HML) + ε.
FactSet
Fundamentals

Additional Analysis Variables

Voice and
Accountability (VA)
One of the 6 variables from the World Bank’s
Worldwide Governance Indicator data set. The
values are scaled between 0 and 100 and are
assigned at the country-year level.
https://info.worl
dbank.org/gove
rnance/wgi/

Political Stability and
Absence of
Violence/Terrorism
(PV)
One of the 6 variables from the World Bank’s
Worldwide Governance Indicator data set. The
values are scaled between 0 and 100 and are
assigned at the country-year level.
https://info.worl
dbank.org/gove
rnance/wgi/

Government
Effectiveness (GE)
One of the 6 variables from the World Bank’s
Worldwide Governance Indicator data set. The
values are scaled between 0 and 100 and are
assigned at the country-year level.
https://info.worl
dbank.org/gove
rnance/wgi/

Regulatory Quality
(RQ)
One of the 6 variables from the World Bank’s
Worldwide Governance Indicator data set. The
values are scaled between 0 and 100 and are
assigned at the country-year level.
https://info.worl
dbank.org/gove
rnance/wgi/

Rule of Law (RL) One of the 6 variables from the World Bank’s
Worldwide Governance Indicator data set. The
values are scaled between 0 and 100 and are
assigned at the country-year level.
https://info.worl
dbank.org/gove
rnance/wgi/

Control of Corruption
(CC)
One of the 6 variables from the World Bank’s
Worldwide Governance Indicator data set. The
values are scaled between 0 and 100 and are
assigned at the country-year level.
https://info.worl
dbank.org/gove
rnance/wgi/
70

Economic Policy
Uncertainty (EPU)
An index that quantifies newspaper coverage of
policy-related economic uncertainty. For most
countries the index is the normalized volume of
news articles discussing economic policy
uncertainty. For the U.S. measure, two other data
sources are also utilized: the number of federal tax
code provisions set to expire and disagreement
among economic forecasters. EPU is set equal to
the natural log of the values reported on the linked
website.
https://www.pol
icyuncertainty.c
om/index.html

Power Distance (PDI) One of the 6 dimensions of national culture from
the Hofstede model. The values are scaled between
0 and 100 and are assigned at the country level.
https://geerthofs
tede.com/

Individualism (IDV) One of the 6 dimensions of national culture from
the Hofstede model. The values are scaled between
0 and 100 and are assigned at the country level.
https://geerthofs
tede.com/

Masculinity (MAS) One of the 6 dimensions of national culture from
the Hofstede model. The values are scaled between
0 and 100 and are assigned at the country level.
https://geerthofs
tede.com/

Uncertainty Avoidance
(UAI)
One of the 6 dimensions of national culture from
the Hofstede model. The values are scaled between
0 and 100 and are assigned at the country level.
https://geerthofs
tede.com/

Long-term Orientation
(LTOWVS)
One of the 6 dimensions of national culture from
the Hofstede model. The values are scaled between
0 and 100 and are assigned at the country level.
https://geerthofs
tede.com/

Indulgence (IVR) One of the 6 dimensions of national culture from
the Hofstede model. The values are scaled between
0 and 100 and are assigned at the country level.
https://geerthofs
tede.com/

GDP Per Capital Natural logarithm of a country’s per capita gross
domestic product (GDP).
https://data.worl
dbank.org
71

Figure 1. Equity Duration Around the World

This figure visualizes the sample average Country-Year Duration for each of the 49 countries included in the sample.
Country-Year Duration (measured in years) is calculated with the same formula used in Dechow et al. (2004), but
with the parameters defined in Table 2 of this paper. The procedure for computing Country-Year Duration is further
defined in Appendix A.

72

Figure 2. Time Series of Country-Year Duration for Developed, Emerging, and ACWI

This figure visualizes the average Country-Year Duration for the 23 developed countries, 26 emerging countries, and
the collective 49 countries included in the All Country World Index (ACWI, or World) by year. Country-Year
Duration is measured in years. The procedure for computing Country-Year Duration is further defined in Appendix
A. The average Country-Year Duration for the full sample (ACWI) primarily varies between 24 and 26 years. The
classification of countries as either developed or emerging is based on the Morgan Stanley Capital International
(MSCI) classification as of the end of 2019 (see Table 1).

73

Figure 3. World Governance Indicators Around the World

This figure visualizes the sample average Governance (WGI) for each of the 49 countries included in the sample. WGI
is calculated by the World Bank and is available on their website. The composite measure used here is an equal
weighted combination of Government Effectiveness (GE), Regulatory Quality (RQ), Rule of Law (RL), and Control of
Corruption (CC) where 100 is the “best” score.

74

Figure 4. Tax Loss Carryforwards Around the World

This figure visualizes the sample average Tax Loss Carryforward allowed for 48 of the countries included in the
sample. Tax Loss Carryforward (measured in years) is provided by EY’s Worldwide Corporate Tax Guide each year
starting in 2004 and is available on their website. Unlimited loss carryforwards that can be used to offset less than
50% of income in future years are coded as 25, unlimited loss carryforwards that can be used to offset 50% or more
of income in future years are coded as 30. The United Arab Emirates is excluded from this figure as they do not have
a corporate tax system.
75

Figure 5. Short Duration Premium by Country (Equal Weighted Portfolios)

This figure visualizes the CAPM Alpha differential that is earned for each country when using equal weighted quintile
portfolios sorted on Firm-Year Duration. The CAPM Alpha differential is equal to the CAPM Alpha of the long
duration portfolio minus the CAPM Alpha of the short duration portfolio. If the result is not statistically significant at
the 10% level, then the CAPM Alpha differential is set equal to 0. There is a negative and statistically significant
CAPM Alpha differential in 44 out of the 49 countries tested. The procedure for computing CAPM Alpha is further
defined in Appendix A.

76

Figure 6. Short Duration Premium by Country (Value Weighted Portfolios)

This figure visualizes the CAPM Alpha differential that is earned for each country when using value weighted quintile
portfolios sorted on Firm-Year Duration. The CAPM Alpha differential is equal to the CAPM Alpha of the long
duration portfolio minus the CAPM Alpha of the short duration portfolio. If the result is not statistically significant at
the 10% level, then the CAPM Alpha differential is set equal to 0. Values below -1.93 are set equal to -1.93 for visual
comparison purposes with Figure 5. Only 2 countries require this adjustment (Hungary -2.04 and Peru -2.40). There
is a negative and statistically significant CAPM Alpha differential in 33 out of the 49 countries tested. The procedure
for computing CAPM Alpha is further defined in Appendix A.

77

Figure 7. Country-Year Duration Portfolios During a Short Term Cash Flow Shock

This figure visualizes the Excess Returns of the portfolio holding MSCI indices of the 10 shortest duration countries
and of the portfolio holding MSCI indices of the 10 longest duration countries. The period represented is the first
quarter of 2020 during the onset of the COVID-19 pandemic and the global shutdown. During this time period, the
short duration premium was reversed, and long duration equities (and countries) outperformed short duration equities
(and countries) as they are less reliant on near-term cash flows.

78

Figure 8. Country-Year Duration Portfolios During Global Interest Rate Hikes

This figure visualizes the Excess Returns of the portfolio holding MSCI indices of the 10 shortest duration countries
and of the portfolio holding MSCI indices of the 10 longest duration countries. The period represented is the full year
of 2022 when interest rates were being increased around the world. During this time period, the short duration premium
was larger than normal because long duration equities (and countries) have more of their market value weighted on
future cash flows to shareholders, and as interest rates rise, these future cash flows get more heavily discounted.

79

Table 1. List of 49 Countries in Sample by Development Status and Geographic Region

Developed
Markets

Emerging
Markets

Americas
Europe & Middle
East
Pacific

Americas
Europe, Middle
East & Africa
Asia
CANADA AUSTRIA AUSTRALIA

ARGENTINA CZECH REPUB. CHINA
UNITED
STATES
BELGIUM HONG KONG

BRAZIL EGYPT INDIA

DENMARK JAPAN

CHILE GREECE INDONESIA

FINLAND NEW ZEALAND

COLOMBIA HUNGARY MALAYSIA

FRANCE SINGAPORE

MEXICO POLAND PAKISTAN

GERMANY

PERU QATAR PHILIPPINES

IRELAND

RUSSIA SOUTH KOREA

ISRAEL

SAUDI ARABIA TAIWAN

ITALY

SOUTH AFRICA THAILAND

NETHERLANDS

TURKEY

NORWAY

UAE

PORTUGAL

SPAIN

SWEDEN

SWITZERLAND

U.K.

This table reports the 49 countries that are included in the final sample. The countries are first grouped by development
status (e.g., developed or emerging) and then they are further grouped by geographic region. Both the development
status and geographic region classification are based on the Morgan Stanley Capital International classification system
as of the end of 2019. Collectively, these 49 countries represent the MSCI All Country World Index (ACWI).

80

Table 2. Equity Duration and Forecasting Assumptions Necessary to Calculate Duration
(Countries Sorted by Duration)

Country-Year Duration Country-Year Forecasting Assumptions
Country N VW EW P50 SD ROE AR ROE T G AR G T R
SINGAPORE 12,999 29.58 28.54 29.21 5.55 0.47 0.06 0.18 0.05 0.06
CANADA 23,610 27.61 26.73 27.72 7.44 0.53 0.06 0.13 0.03 0.06
UNITED STATES 140,124 27.51 25.78 26.77 6.96 0.53 0.06 0.18 0.02 0.06
SWEDEN 9,878 27.24 27.97 28.35 6.48 0.59 0.06 0.13 0.03 0.06
NETHERLANDS 3,124 27.19 25.02 26.19 5.67 0.44 0.06 0.19 0.02 0.06
FINLAND 3,143 27.05 26.64 27.43 4.80 0.52 0.06 0.16 0.03 0.06
U.K. 33,784 26.97 25.21 26.92 8.11 0.50 0.06 0.16 0.02 0.06
SWITZERLAND 5,597 26.94 24.65 25.62 6.10 0.54 0.06 0.11 0.02 0.06
DENMARK 4,027 26.91 22.92 23.62 7.69 0.58 0.06 0.11 0.02 0.06
IRELAND 845 26.89 26.07 26.74 4.87 0.46 0.07 0.18 0.05 0.07
GERMANY 17,837 26.87 25.34 26.56 6.81 0.47 0.06 0.12 0.02 0.06
FRANCE 16,755 26.26 23.77 25.46 7.84 0.51 0.06 0.19 0.02 0.06
CHINA 45,399 26.15 25.86 26.31 2.59 0.46 0.08 0.14 0.05 0.08
NORWAY 4,710 25.22 22.75 25.29 9.88 0.50 0.06 0.15 0.02 0.06
JAPAN 85,526 25.09 19.68 21.34 8.90 0.43 0.06 0.19 0.02 0.06
ISRAEL 5,773 24.92 23.01 23.78 6.27 0.47 0.07 0.10 0.04 0.07
AUSTRALIA 19,471 24.81 23.97 24.79 8.07 0.55 0.07 0.12 0.03 0.07
QATAR 662 23.95 22.77 23.30 3.49 0.51 0.08 0.11 0.05 0.08
TAIWAN 32,090 23.78 23.37 23.37 4.43 0.58 0.08 0.19 0.05 0.08
AUSTRIA 1,834 23.77 22.21 23.86 7.71 0.43 0.06 0.19 0.02 0.06
MALAYSIA 18,995 23.60 20.20 21.16 6.98 0.49 0.08 0.13 0.05 0.08
NEW ZEALAND 2,414 23.54 23.09 23.62 6.43 0.58 0.07 0.16 0.03 0.07
THAILAND 12,157 23.10 20.13 21.16 6.51 0.46 0.08 0.10 0.04 0.08
SPAIN 3,694 22.86 21.05 22.48 6.69 0.54 0.08 0.13 0.03 0.08
SAUDI ARABIA 2,325 22.32 22.95 23.00 3.49 0.59 0.08 0.12 0.03 0.08
BELGIUM 3,077 22.32 20.77 21.37 6.79 0.54 0.07 0.10 0.02 0.07
POLAND 7,688 22.01 19.69 21.36 7.86 0.43 0.08 0.11 0.04 0.08
UAE 1,601 21.93 18.73 20.02 7.09 0.47 0.08 0.17 0.04 0.08
ITALY 6,367 21.41 19.15 20.95 7.85 0.52 0.08 0.14 0.02 0.08
HONG KONG 30,920 21.26 17.33 19.47 10.10 0.50 0.08 0.16 0.03 0.08
PERU 2,242 20.70 14.33 17.21 9.81 0.59 0.09 0.12 0.04 0.09
CZECH REPUB. 588 20.36 14.58 17.69 10.21 0.48 0.08 0.16 0.03 0.08
CHILE 3,994 20.31 17.38 19.13 7.45 0.55 0.09 0.09 0.04 0.09
SOUTH KOREA 32,544 19.75 16.99 18.90 9.42 0.46 0.08 0.09 0.03 0.08
PORTUGAL 1,297 19.74 15.20 17.46 9.07 0.43 0.09 0.09 0.02 0.09
MEXICO 2,778 19.18 15.49 17.43 7.73 0.52 0.10 0.12 0.02 0.10
PHILIPPINES 4,553 17.97 14.26 16.48 8.59 0.52 0.12 0.13 0.05 0.12
81

Country-Year Duration Country-Year Forecasting Assumptions
Country N VW EW P50 SD ROE AR ROE T G AR G T R
SOUTH AFRICA 6,884 17.71 13.35 15.15 8.14 0.47 0.12 0.13 0.03 0.12
INDONESIA 9,403 17.67 12.57 14.68 8.54 0.46 0.14 0.13 0.05 0.14
INDIA 43,670 17.33 11.32 13.73 9.41 0.52 0.15 0.18 0.05 0.15
BRAZIL 6,672 16.42 12.12 14.26 8.74 0.49 0.13 0.14 0.02 0.13
HUNGARY 762 15.56 13.83 15.22 7.77 0.53 0.11 0.13 0.03 0.11
GREECE 5,327 15.46 13.48 15.76 9.78 0.59 0.11 0.18 0.02 0.11
COLOMBIA 993 15.45 12.06 14.75 9.02 0.56 0.11 0.19 0.03 0.11
EGYPT 2,728 14.53 12.22 13.45 6.90 0.59 0.16 0.14 0.05 0.16
TURKEY 6,320 14.12 12.20 13.91 7.53 0.43 0.16 0.09 0.05 0.16
RUSSIA 3,590 13.72 8.83 11.37 9.88 0.44 0.16 0.14 0.03 0.16
ARGENTINA 1,685 13.69 10.87 12.40 7.97 0.50 0.16 0.09 0.03 0.16
PAKISTAN 4,681 13.48 9.51 11.30 9.00 0.57 0.16 0.12 0.04 0.16

This table is sorted by the value weighted Country-Year Duration from longest to shortest. Unless otherwise noted,
the sample period is 2000 to 2020. See Appendix A for more detailed variable definitions. N is the number of firm-
years for each country. VW is the sample average value-weighted Country-Year Duration for each country. Country-
Year Duration is calculated with the same formula used in Dechow et al. (2004), but with the parameters defined in
this panel and in Appendix A. EW is the equal-weighted Country-Year Duration for each country. P50 is the median
Country-Year Duration for each country. SD is the standard deviation of Firm-Year Duration for each country. The
AR(1) coefficients for ROE and Growth (ROE AR and G AR) are calculated for each of the 49 countries in the sample
using country specific firm-years between 1995 and 2021. The Terminal ROE (ROE T) is set equal to the median
Total Equity Risk Premium plus the Differential Inflation Riskfree Rate as defined by Damodaran (2020a) and
Damodaran (2020b) for the sample. The Terminal Growth (G T) is set equal to the median GDP growth for the country
for the period between 2000 and 2020. The Discount Rate (R) is set equal to the Terminal ROE.

82

Table 3. Summary Statistics for Country-Year Sample

Variable N Mean SD P25 P50 P75
Country-Year Duration 1,023 19.25 5.99 14.48 20.39 24.15
Governance (WGI) 976 73.29 21.75 52.80 78.97 93.63
Tax Loss Carryforward 1,004 17.58 11.37 5.00 15.00 30.00
Innovation Box 1,004 0.15 0.36 0.00 0.00 0.00
Developed 1,023 0.47 0.50 0.00 0.00 1.00
Inflation 1,023 1.39 5.05 -0.83 0.21 1.92
Short Dur. Industry % 1,023 0.23 0.09 0.17 0.23 0.29
Long Dur. Industry % 1,023 0.12 0.09 0.05 0.12 0.19
Largest Firm % 1,023 0.68 0.12 0.61 0.71 0.77

This table provides the summary statistics for the primary variables in the country-year sample (i.e., not the firm-
year sample) used in Tables 5, 7, and 9. Governance (WGI) is a composite measure of the following four World
Governance Indicators measures: Government Effectiveness, Regulatory Quality, Rule of Law, and Control of
Corruption. Each measure of governance is scaled between 0 and 100 and is based on average values for the sample
period. The composite Governance (WGI) measure is an equal weighted average of the four measures. Tax Loss
Carryforward is the number of years that a tax loss can be carried forward to offset future taxable income.
Innovation Box is equal to Yes (1) if the country-year has innovation box tax incentives in place, it is set equal to No
(0) otherwise. Developed is based on the MSCI market development classification as of the end of 2019. Inflation is
the country-year level of inflation, downloaded from The World Bank’s website, standardized relative to inflation in
the U.S. for that year. Short Dur. Industry % is the percentage of firm-years in each country that are in short duration
industries (Categories 1 and 3 of the Fama and French 12 industries). Long Dur. Industry % is the percentage of
firm-years that are in long duration industries (Categories 6 and 10 of the Fama and French 12 industries). The
notation as a long or short duration industry is based on a comparison of the average duration in each industry in the
sample used in this paper. Largest Firm % represents the percentage of market value that the top 10% of firms
within a country own.

83

Table 4. Sample Average World Governance Indicators (Countries Sorted by Composite
WGI Measure)

Country WGI GE RQ RL CC
FINLAND 98.91 99.30 97.49 99.69 99.19
DENMARK 98.38 98.54 96.67 98.68 99.65
NEW ZEALAND 97.43 94.83 97.96 97.73 99.21
SINGAPORE 97.40 99.12 99.10 93.54 97.83
SWEDEN 97.14 96.73 95.58 98.02 98.21
SWITZERLAND 96.96 98.23 95.38 97.55 96.69
NETHERLANDS 96.56 96.71 97.55 96.03 95.95
NORWAY 96.24 97.15 91.60 99.03 97.18
CANADA 95.09 95.68 94.73 95.21 94.72
AUSTRALIA 94.85 93.71 96.32 94.76 94.61
HONG KONG 94.33 94.76 99.04 90.88 92.64
AUSTRIA 93.96 93.67 92.37 97.36 92.45
U.K. 93.94 92.41 96.22 93.42 93.71
GERMANY 92.92 91.56 93.35 93.01 93.77
IRELAND 92.15 89.59 94.95 92.56 91.51
UNITED STATES 90.69 90.98 91.34 91.38 89.06
BELGIUM 89.61 91.40 87.54 89.07 90.43
FRANCE 88.15 89.59 84.64 89.52 88.84
JAPAN 87.94 90.53 83.78 88.65 88.78
CHILE 85.87 81.15 89.53 85.29 87.52
SPAIN 82.74 83.28 82.72 84.07 80.88
ISRAEL 81.84 85.99 83.19 79.51 78.68
PORTUGAL 81.69 82.23 79.36 84.24 80.93
TAIWAN 81.30 84.22 83.76 80.93 76.28
SOUTH KOREA 77.40 81.11 77.57 80.78 70.12
CZECH REPUB. 77.38 78.66 83.06 79.75 68.07
UAE 76.18 81.86 73.69 68.51 80.66
HUNGARY 72.84 73.24 78.54 72.45 67.12
QATAR 72.68 73.05 67.49 72.15 78.05
POLAND 71.24 69.60 76.56 68.35 70.45
MALAYSIA 68.91 80.75 68.80 64.16 61.93
ITALY 68.77 68.84 76.20 64.73 65.33
GREECE 66.94 68.12 71.39 67.70 60.54
SOUTH AFRICA 60.90 63.54 64.63 54.80 60.63
SAUDI ARABIA 55.67 54.38 54.47 56.59 57.25
TURKEY 55.67 59.42 59.90 50.57 52.80
THAILAND 54.03 62.43 57.43 53.54 42.72
BRAZIL 50.12 48.22 53.82 47.37 51.06
COLOMBIA 48.97 50.52 59.70 39.27 46.38
INDIA 48.30 54.90 41.23 55.03 42.05
MEXICO 47.01 56.44 60.77 35.64 35.20

84

Country WGI GE RQ RL CC
CHINA 46.03 60.01 43.07 39.46 41.57
PERU 44.91 41.18 63.18 32.86 42.42
PHILIPPINES 44.56 56.11 52.91 37.16 32.07
ARGENTINA 39.83 52.23 29.74 33.56 43.78
INDONESIA 37.89 46.89 42.47 33.98 28.25
EGYPT 37.49 38.69 34.33 43.75 33.21
RUSSIA 30.27 41.28 40.31 21.23 18.27
PAKISTAN 24.82 30.53 26.91 22.59 19.25

This table is sorted by the Governance (WGI) composite measure from high to low. The table reports the sample
average for Governance (WGI) which is an equal weighted combination of four of the World Bank’s World
Governance Indicators: Government Effectiveness (GE), Regulatory Quality (RQ), Rule of Law (RL), and Control of
Corruption (CC). The value ranges from 0 to 100 where 100 is the “best”. Note that this variable is not available for
2001.

85

Table 5. Country-Level Governance and Country-Level Duration

Pred.
(1) (2) (3)
(4) (5)
VARIABLES Sign
Country-Year
Duration
Country-Year
Duration
Country-Year
Duration
Country-Year
Duration
Avg. Country-
Year Duration

Governance (WGI)
+
0.198*** 0.115*** 0.115*** 0.106*** 0.101**

(33.96) (11.48) (3.83) (10.38) (2.35)
Developed
+
1.748*** 1.748 1.883*** 1.626

(4.55) (1.37) (5.72) (1.06)
Inflation
-
-0.134*** -0.134* -0.241*** -0.338*

(-4.91) (-1.76) (-3.55) (-1.89)
Short Dur. Industry %
-

-6.001*** -6.001 -5.224*** -5.551

(-3.49) (-1.18) (-3.60) (-0.83)
Long Dur. Industry %
+

17.654*** 17.654** 17.419*** 16.315**

(10.98) (2.31) (23.08) (2.57)
Largest Firms %
?
-8.955*** -8.955 -8.262*** -8.913

(-6.50) (-1.30) (-8.62) (-1.66)

Observations

976 976 976 976 49
R-squared

0.542 0.621 0.621 0.669 0.715

This table tests the relationship between country-year Governance (WGI) and Country-Year Duration. See Appendix
A for more detailed variable definitions. Columns 1, 2 and 5 are standard OLS regressions with t-statistics reported in
paratheses. Column 3 reports t-statistics based on cluster-robust standard errors clustered by country. Column 4 reports
t-statistics based on Newey-West corrected Fama-MacBeth standard errors. In Column 5 the panel is collapsed to a
single observation per county by using the sample averages for each country for each variable.

86
Table 6. Statutory Tax Loss Carryback Periods by Country-Year (Countries Sorted by Sample Average)

Country Avg. 2020 2019 2018 2017 2016 2015 2014 2013 2012 2011 2010 2009 2008 2007 2006 2005 2004
AUSTRALIA 30 30 30 30 30 30 30 30 30 30 30 30 30 30 30 30 30 30
AUSTRIA 30 30 30 30 30 30 30 30 30 30 30 30 30 30 30 30 30 30
BELGIUM 30 30 30 30 30 30 30 30 30 30 30 30 30 30 30 30 30 30
CHILE 30 30 30 30 30 30 30 30 30 30 30 30 30 30 30 30 30 30
DENMARK 30 30 30 30 30 30 30 30 30 30 30 30 30 30 30 30 30 30
FRANCE 30 30 30 30 30 30 30 30 30 30 30 30 30 30 30 30 30 30
GERMANY 30 30 30 30 30 30 30 30 30 30 30 30 30 30 30 30 30 30
HONG KONG 30 30 30 30 30 30 30 30 30 30 30 30 30 30 30 30 30 30
IRELAND 30 30 30 30 30 30 30 30 30 30 30 30 30 30 30 30 30 30
ISRAEL 30 30 30 30 30 30 30 30 30 30 30 30 30 30 30 30 30 30
NEW ZEALAND 30 30 30 30 30 30 30 30 30 30 30 30 30 30 30 30 30 30
SINGAPORE 30 30 30 30 30 30 30 30 30 30 30 30 30 30 30 30 30 30
SOUTH AFRICA 30 30 30 30 30 30 30 30 30 30 30 30 30 30 30 30 30 30
SWEDEN 30 30 30 30 30 30 30 30 30 30 30 30 30 30 30 30 30 30
U.K. 30 30 30 30 30 30 30 30 30 30 30 30 30 30 30 30 30 30
PERU 28 30 30 30 30 30 30 30 30 30 30 30 30 30 30 30 30 4
NORWAY 28 30 30 30 30 30 30 30 30 30 30 30 30 30 30 30 10 10
MALAYSIA 27 7 7 30 30 30 30 30 30 30 30 30 30 30 30 30 30 30
BRAZIL 25 25 25 25 25 25 25 25 25 25 25 25 25 25 25 25 25 25
SAUDI ARABIA 25 25 25 25 25 25 25 25 25 25 25 25 25 25 25 25 25 25
UNITED STATES 22 30 30 30 20 20 20 20 20 20 20 20 20 20 20 20 20 20
COLOMBIA 22 12 12 12 12 30 30 30 30 30 30 30 30 30 30 5 8 8
SPAIN 21 30 30 30 30 30 30 18 18 18 15 15 15 15 15 15 15 15
HUNGARY 20 5 5 5 5 5 5 30 30 30 30 30 30 30 30 30 30 5
ITALY 18 30 30 30 30 30 30 30 30 30 5 5 5 5 5 5 5 5
CANADA 18 20 20 20 20 20 20 20 20 20 20 20 20 20 20 10 7 7
RUSSIA 15 30 30 30 30 10 10 10 10 10 10 10 10 10 10 10 10 10

87
Country Avg. 2020 2019 2018 2017 2016 2015 2014 2013 2012 2011 2010 2009 2008 2007 2006 2005 2004
NETHERLANDS 12 6 6 9 9 9 9 9 9 9 9 9 9 9 9 30 30 30
FINLAND 10 10 10 10 10 10 10 10 10 10 10 10 10 10 10 10 10 10
INDONESIA 10 10 10 10 10 10 10 10 10 10 10 10 10 10 10 10 10 10
MEXICO 10 10 10 10 10 10 10 10 10 10 10 10 10 10 10 10 10 10
SOUTH KOREA 8 10 10 10 10 10 10 10 10 10 10 10 5 5 5 5 5 5
TAIWAN 8 10 10 10 10 10 10 10 10 10 10 10 5 5 5 5 5 5
JAPAN 8 10 10 9 9 9 9 9 9 9 7 7 7 7 7 7 7 7
INDIA 8 8 8 8 8 8 8 8 8 8 8 8 8 8 8 8 8 8
SWITZERLAND 7 7 7 7 7 7 7 7 7 7 7 7 7 7 7 7 7 7
PORTUGAL 6 5 5 5 5 12 12 12 5 5 4 4 6 6 6 6 6 6
PAKISTAN 6 6 6 6 6 6 6 6 6 6 6 6 6 6 6 6 6 6
CHINA 6 10 10 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5
ARGENTINA 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5
CZECH REPUB. 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5
EGYPT 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5
GREECE 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5
POLAND 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5
THAILAND 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5
TURKEY 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5
QATAR 3 5 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3
PHILIPPINES 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3
UAE N/A N/A N/A N/A N/A N/A N/A N/A N/A N/A N/A N/A N/A N/A N/A N/A N/A N/A

This table is sorted by the sample average Tax Loss Carryforward from high to low, if values are tied then they the countries are sorted alphabetically. The table
reports the value of Tax Loss Carryforward (measured in years) for each country-year in the sample, with the exception of 2000-2003 which had to be removed
due to space constraints. Tax Loss Carryforward is the number of years that a tax loss can be carried forward to offset future taxable income. If carryforward is
indefinite, then variable is set to 30 if losses can offset ≥ 50% of current year income or 25 if losses can offset < 50% of current year income. The UAE does not
have a corporate tax system during the sample period.

88

Table 7. Tax Loss Carryforward Policy and Country-Level Duration

Pred.
(1) (2) (3) (4) (5)
VARIABLES
Sign
Country-Year
Duration
Country-Year
Duration
Country-Year
Duration
Country-Year
Duration
Avg. Country-
Year Duration

Tax Loss Carryforward
+
0.204*** 0.061*** 0.061 0.052*** 0.048

(13.16) (4.80) (1.29) (3.12) (0.84)
Developed
+
4.519*** 4.519*** 4.370*** 3.886***

(13.97) (3.40) (8.63) (2.89)
Inflation
-

-0.174***
-0.174* -0.351*** -0.465**

(-6.54)
(-1.83) (-3.88) (-2.67)
Short Dur. Industry %
-
-7.683*** -7.683 -6.013*** -6.868

(-4.24) (-1.03) (-3.07) (-0.96)
Long Dur. Industry %
+
25.043*** 25.043*** 23.894*** 21.726***

(15.02) (3.97) (18.74) (3.16)
Largest Firms %
?
-8.163*** -8.163 -7.374*** -8.130

(-5.67) (-1.23) (-6.54) (-1.42)

Observations

1,004 1,004 1,004 1,004 48
R-squared

0.147 0.587 0.587 0.648 0.693

This table tests the relationship between country-year Tax Loss Carryforward policy and Country-Year Duration.
See Appendix A for more detailed variable definitions. Columns 1, 2 and 5 are standard OLS regressions with t-
statistics reported in paratheses. Column 3 reports t-statistics based on cluster-robust standard errors clustered by
country. Column 4 reports t-statistics based on Newey-West corrected Fama-MacBeth standard errors. In Column 5
the panel is collapsed to a single observation per county by using the sample averages for each country for each
variable.

89
Table 8. Availability of Innovation Box Tax Policies by Country-Year (Countries Sorted by Total Years Available)

Country Total 2020 2019 2018 2017 2016 2015 2014 2013 2012 2011 2010 2009 2008 2007 2006 2005 2004
FRANCE 17 Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes
HUNGARY 17 Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes
IRELAND 17 Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes
BELGIUM 14 Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes

NETHERLANDS 14 Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes

CHINA 13 Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes

SPAIN 13 Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes

SWITZERLAND 10 Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes

U.K. 8 Yes Yes Yes Yes Yes Yes Yes Yes

PORTUGAL 7 Yes Yes Yes Yes Yes Yes Yes

ITALY 6 Yes Yes Yes Yes Yes Yes

TURKEY 6 Yes Yes Yes Yes Yes Yes

INDIA 5 Yes Yes Yes Yes Yes

ISRAEL 4 Yes Yes Yes Yes

SINGAPORE 3 Yes Yes Yes

POLAND 2 Yes Yes

ARGENTINA 0

AUSTRALIA 0

AUSTRIA 0

BRAZIL 0

CANADA 0

CHILE 0

COLOMBIA 0

CZECH REPUB. 0

DENMARK 0

EGYPT 0

FINLAND 0

90
Country Total 2020 2019 2018 2017 2016 2015 2014 2013 2012 2011 2010 2009 2008 2007 2006 2005 2004
GERMANY 0

GREECE 0

HONG KONG 0

INDONESIA 0

JAPAN 0

MALAYSIA 0

MEXICO 0

NEW ZEALAND 0

NORWAY 0

PAKISTAN 0

PERU 0

PHILIPPINES 0

QATAR 0

RUSSIA 0

SAUDI ARABIA 0

SOUTH AFRICA 0

SOUTH KOREA 0

SWEDEN 0

TAIWAN 0

THAILAND 0

UNITED STATES 0

UAE N/A N/A N/A N/A N/A N/A N/A N/A N/A N/A N/A N/A N/A N/A N/A N/A N/A N/A

This table is sorted by the total number of years that innovation box tax policies are available in the reported period (Total) from high to low, if values are tied then
they the countries are sorted alphabetically. The table reports Innovation Box for each country-year in the sample, with the exception of 2000-2003 which had to
be removed due to space constraints. A binary indicator variable equal to Yes (1) if the country-year has innovation box tax incentives in place (e.g., a lower tax
rate is applied to a separate schedule of firm produced innovation related income), it is set equal to No (0) otherwise. The UAE does not have a corporate tax system
during the sample period.

91

Table 9. Innovation Box Tax Policy and Country-Level Duration

Pred.
(1) (2) (3) (4) (5)
VARIABLES Sign
Country-Year
Duration
Country-Year
Duration
Country-Year
Duration
Country-Year
Duration
Avg. Country-
Year Duration

Innovation Box
+ 2.977*** 1.269*** 1.269 1.435** 1.850

(5.89) (3.27) (1.08) (2.80) (0.92)
Inflation
- -0.275*** -0.275** -0.522*** -0.675***

(-9.28) (-2.14) (-4.37) (-3.68)
Short Dur. Industry %
- -19.383*** -19.383** -17.353*** -14.862*

(-10.02) (-2.47) (-12.68) (-1.93)
Long Dur. Industry %
+ 29.598*** 29.598*** 27.613*** 23.281***

(16.01) (3.24) (13.88) (3.05)
Largest Firms %
? -5.524*** -5.524 -5.561*** -5.269

(-3.36) (-0.86) (-5.16) (-0.81)

Observations
1,004 1,004 1,004 1,004 48
R-squared
0.033 0.461 0.461 0.532 0.602

This table tests the relationship between country-year Innovation Box policy and Country-Year Duration. See
Appendix A for more detailed variable definitions. Columns 1, 2 and 5 are standard OLS regressions with t-statistics
reported in paratheses. Column 3 reports t-statistics based on cluster-robust standard errors clustered by country.
Column 4 reports t-statistics based on Newey-West corrected Fama-MacBeth standard errors. In Column 5 the panel
is collapsed to a single observation per county by using the sample averages for each country for each variable.
Unlike Tables 5 and 7, Developed is removed as a control variable, see Chapter 5.3.2 for further discussion.

92

Table 10. International Duration Portfolio Characteristics

Panel A: Developed Markets and Emerging Markets

Developed –
EW CAPM
1 5 Q5-Q1
p-
value

Developed –
VW CAPM
1 5 Q5-Q1
p-
value
Firm-Year Dur. 11.83 31.84 20.01

Firm-Year Dur. 13.06 30.94 17.87

Total Vol 4.23 5.73 1.51 0.00

Total Vol 5.03 5.52 0.49 0.13
IVol 2.68 2.65 -0.02 0.88

IVol 2.65 2.00 -0.65 0.00
Excess Return 1.01 0.13 -0.88 0.04

Excess Return 0.95 0.45 -0.50 0.27
CAPM Alpha 0.75 -0.27 -1.02 0.00

CAPM Alpha 0.62 0.05 -0.57 0.01
MKT Beta 0.75 1.17 0.42 0.00

MKT Beta 0.98 1.18 0.20 0.00

Developed –
EW 3-Factor
1 5 Q5-Q1
p-
value

Developed –
VW 3-Factor
1 5 Q5-Q1
p-
value
IVol 1.75 1.48 -0.27 0.01

IVol 1.76 1.41 -0.35 0.00
FF Alpha 0.51 -0.30 -0.81 0.00

FF Alpha 0.31 0.12 -0.20 0.17
MKT Beta 0.75 1.11 0.36 0.00

MKT Beta 0.99 1.12 0.13 0.00

Emerging –
EW CAPM
1 5 Q5-Q1
p-
value

Emerging –
VW CAPM
1 5 Q5-Q1
p-
value
Firm-Year Dur. 6.37 27.58 21.20

Firm-Year Dur. 6.06 27.04 20.98

Total Vol 6.36 7.03 0.67 0.10

Total Vol 7.51 7.00 -0.51 0.24
IVol 2.95 5.92 2.97 0.00

IVol 3.29 5.85 2.56 0.00
Excess Return 1.47 0.60 -0.87 0.13

Excess Return 1.46 0.55 -0.91 0.15
CAPM Alpha 1.00 0.29 -0.72 0.08

CAPM Alpha 0.90 0.23 -0.67 0.11
MKT Beta 0.96 0.65 -0.31 0.00

MKT Beta 1.15 0.66 -0.50 0.00

Emerging –
EW 3-Factor
1 5 Q5-Q1
p-
value

Emerging –
VW 3-Factor
1 5 Q5-Q1
p-
value
IVol 2.12 5.44 3.32 0.00

IVol 2.74 5.60 2.86 0.00
FF Alpha 0.58 0.27 -0.31 0.42

FF Alpha 0.33 0.34 0.01 0.99
MKT Beta 1.00 0.76 -0.24 0.00

MKT Beta 1.12 0.74 -0.39 0.00

93

Panel B: Developed Market Breakouts

Europe –
EW 3-Factor
1 5 Q5-Q1
p-
value

Europe –
VW 3-Factor
1 5 Q5-Q1
p-
value
Firm-Year Dur. 12.87 31.71 18.84

Firm-Year Dur. 14.01 30.80 16.79

Total Vol 5.63 5.81 0.18 0.61

Total Vol 6.63 5.77 -0.86 0.02
IVol 1.33 1.62 0.29 0.00

IVol 1.66 1.65 -0.01 0.94
Excess Return 0.95 0.02 -0.93 0.06

Excess Return 0.72 0.45 -0.27 0.62
FF Alpha 0.33 -0.41 -0.74 0.00

FF Alpha 0.03 0.18 0.15 0.29
MKT Beta 0.92 1.02 0.10 0.00

MKT Beta 1.08 1.06 -0.02 0.64

Japan –
EW 3-Factor
1 5 Q5-Q1
p-
value

Japan –
VW 3-Factor
1 5 Q5-Q1
p-
value
Firm-Year Dur. 8.33 29.38 21.04

Firm-Year Dur. 7.95 28.63 20.67

Total Vol 4.66 5.85 1.18 0.00

Total Vol 4.91 5.05 0.14 0.64
IVol 1.19 1.82 0.63 0.00

IVol 1.30 1.41 0.11 0.19
Excess Return 0.88 0.04 -0.84 0.07

Excess Return 0.88 0.09 -0.79 0.07
FF Alpha 0.35 -0.16 -0.52 0.00

FF Alpha 0.28 0.16 -0.11 0.35
MKT Beta 0.90 1.03 0.13 0.00

MKT Beta 0.98 0.98 0.00 0.99

Asia ex. Japan –
EW 3-Factor
1 5 Q5-Q1
p-
value

Asia ex. Japan –
VW 3-Factor
1 5 Q5-Q1
p-
value
Firm-Year Dur. 6.66 31.52 24.86

Firm-Year Dur. 8.95 30.89 21.94

Total Vol 6.30 6.14 -0.15 0.69

Total Vol 6.69 5.91 -0.79 0.04
IVol 2.33 1.82 -0.51 0.00

IVol 2.77 2.11 -0.66 0.00
Excess Return 1.17 -0.02 -1.20 0.03

Excess Return 1.23 0.44 -0.79 0.15
FF Alpha 0.45 -0.51 -0.97 0.00

FF Alpha 0.28 -0.21 -0.49 0.03
MKT Beta 0.89 0.96 0.07 0.05

MKT Beta 1.05 0.99 -0.06 0.19

North America –
EW 3-Factor
1 5 Q5-Q1
p-
value

North America –
VW 3-Factor
1 5 Q5-Q1
p-
value
Firm-Year Dur. 18.05 33.03 14.98

Firm-Year Dur. 18.16 32.20 14.04

Total Vol 5.08 6.89 1.81 0.00

Total Vol 6.00 6.74 0.73 0.06
IVol 1.62 2.18 0.56 0.00

IVol 1.74 2.17 0.43 0.00
Excess Return 1.15 0.22 -0.93 0.08

Excess Return 1.11 0.45 -0.65 0.23
FF Alpha 0.50 -0.36 -0.86 0.00

FF Alpha 0.31 -0.07 -0.39 0.02
MKT Beta 0.86 1.10 0.24 0.00

MKT Beta 1.12 1.20 0.08 0.10

94

This table presents portfolio Firm-Year Duration, Total Volatility (Total Vol), Idiosyncratic Volatility (IVol), Excess
Return, CAPM Alpha, FF Alpha, and Market Betas (MKT Beta) for a variety of portfolios sorted on Firm-Year
Duration and formed across various geographic regions and development statuses. See Appendix A for more detailed
variable definitions. Panel A: Firm-Year Duration is calculated with the same formula used in Dechow et al. (2004),
but with the parameters defined in Table 2. However, unlike Tables 5, 7, and 9, in this table Firm-Year Duration is
computed at the firm-year level instead of the country-year level. Total Vol is the standard deviation of monthly
portfolio returns. IVol is the standard deviation of the residual of the model. Excess Return is the portfolio return in
excess of the U.S. risk-free rate. CAPM Alpha and FF Alpha are the intercepts from a regression of monthly portfolio
returns minus the risk-free rate on the region's value-weight market portfolio minus the risk-free rate. The region’s
market portfolio return is taken from Ken French’s website for the 3-Factor regressions, and from the market specific
MSCI index for the CAPM regressions. MKT Beta is the β 1 slope estimate of the regression. The sub-tables captioned
“CAPM” are based on the regression Return-rf = α + β 1(Market-rf) + ε. The sub-tables captioned “3-Factor” are based
on the regression Return-rf = α + β 1(Market-rf) + β 2(SMB) + β 3(HML) + ε. β 2 and β 3 are not reported for brevity. P-
values are reported in the furthest right column of each sub-table. Any difference, Q5-Q1, that is significant at the
10% level or better is bolded. Developed includes Australia, Austria, Belgium, Canada, Demark, Finland, France,
Germany, Hong Kong, Ireland, Israel, Italy, Japan, Netherlands, New Zealand, Norway, Portugal, Singapore, Spain,
Sweden, Switzerland, United Kingdom, and the United States. Emerging includes Argentina, Brazil, Chile, China,
Colombia, Czech Republic, Egypt, Greece, Hungary, India, Indonesia, Malaysia, Mexico, Pakistan, Peru, Philippines,
Poland, Qatar, Russia, Saudi Arabia, South Africa, South Korea, Taiwan, Thailand, Turkey, United Arab Emirates. In
June of each year, I form quintile portfolios based on Firm-Year Duration using accounting information lagged three
months and market values from the end of the previous fiscal year. Firm-Year Duration, Total Vol, and Excess Return
are removed from the 3-Factor results for brevity. EW indicates that the portfolios were formed on equal weighting
while VW indicates that the portfolios were formed with a market value based weighting. Panel B: All variables are
the same unless otherwise noted. Europe includes: Austria, Belgium, Switzerland, Germany, Demark, Spain, Finland,
France, United Kingdom, Ireland, Italy, Israel (excluded in 3-Factor), Netherlands, Norway, Portugal, and Sweden.
Japan includes: Japan. Asia Pacific ex. Japan includes: Australia, Hong Kong, New Zealand, and Singapore. North
America includes: Canada and the United States. These four categories together make up the developed market
countries.

95

Table 11. Within Country Duration Portfolio Characteristics

Panel A: Equal Weighted Portfolios (Countries Sorted by Duration)

Country
Q1
Dur
Q5
Dur
Q1
Alpha
Q5
Alpha
Q5-
Q1
Alpha
p-
value
Q1
Beta
Q5
Beta
Q5-
Q1
Beta
p-
value
SINGAPORE 22.10 34.14 0.82 -0.41 -1.23 0.00 0.90 0.95 0.06 0.49
CANADA 17.97 34.09 0.74 -0.54 -1.27 0.00 0.98 1.14 0.16 0.04
UNITED STATES 18.20 32.90 0.72 -0.29 -1.01 0.00 0.91 1.28 0.36 0.00
SWEDEN 19.17 34.20 1.08 -0.25 -1.33 0.00 0.82 1.02 0.20 0.01
NETHERLANDS 17.67 30.56 0.75 -0.11 -0.85 0.01 0.94 0.97 0.03 0.63
FINLAND 20.35 31.17 1.30 0.22 -1.08 0.01 0.58 0.73 0.15 0.05
U.K. 15.09 33.16 0.75 -0.53 -1.28 0.00 0.95 0.99 0.04 0.62
SWITZERLAND 17.21 30.64 0.83 0.01 -0.81 0.00 0.79 1.07 0.28 0.00
DENMARK 12.97 31.73 0.69 -0.25 -0.94 0.01 0.71 0.92 0.21 0.01
IRELAND 20.24 30.28 1.52 0.53 -0.99 0.13 0.87 0.84 -0.03 0.82
GERMANY 16.81 32.11 0.89 -0.39 -1.28 0.00 0.72 0.69 -0.03 0.64
FRANCE 12.99 31.49 1.09 -0.23 -1.33 0.00 0.74 0.89 0.15 0.02
CHINA 23.33 28.23 1.04 1.00 -0.04 0.96 0.52 0.46 -0.06 0.60
NORWAY 9.83 32.55 0.56 -0.50 -1.07 0.00 0.85 0.94 0.09 0.15
JAPAN 8.33 29.38 0.85 0.00 -0.85 0.00 0.79 1.02 0.23 0.00
ISRAEL 16.30 29.19 1.44 0.00 -1.45 0.00 0.80 0.89 0.09 0.28
AUSTRALIA 14.03 32.15 0.76 -0.67 -1.43 0.00 1.01 1.09 0.08 0.14
QATAR 18.89 25.99 0.81 0.07 -0.74 0.12 0.63 0.98 0.35 0.00
TAIWAN 18.16 27.97 0.99 -0.29 -1.28 0.00 1.04 1.09 0.05 0.52
AUSTRIA 14.20 29.67 0.77 -0.10 -0.88 0.01 0.76 0.64 -0.12 0.06
MALAYSIA 10.90 27.60 0.59 -0.43 -1.02 0.00 1.03 1.02 -0.01 0.91
NEW ZEALAND 14.30 29.01 0.98 0.14 -0.84 0.01 0.81 0.86 0.05 0.46
THAILAND 11.85 27.02 1.29 -0.04 -1.34 0.00 0.75 0.83 0.08 0.19
SPAIN 14.48 27.32 0.78 0.01 -0.77 0.01 0.84 0.74 -0.10 0.09
SAUDI ARABIA 19.24 26.24 1.17 0.21 -0.95 0.25 0.62 0.68 0.07 0.63
BELGIUM 12.32 28.22 1.06 0.03 -1.03 0.00 0.72 0.74 0.02 0.71
POLAND 11.64 27.12 1.07 -0.25 -1.32 0.00 0.87 0.85 -0.03 0.65
UAE 9.91 25.53 0.68 -0.27 -0.95 0.07 0.69 0.52 -0.17 0.07
ITALY 8.37 26.51 0.68 -0.14 -0.82 0.00 0.95 0.91 -0.03 0.50
HONG KONG 2.02 28.02 1.01 -0.82 -1.82 0.00 0.87 0.84 -0.03 0.76
PERU 4.74 24.43 1.77 -0.15 -1.93 0.00 0.33 0.34 0.02 0.78
CZECH REPUB. 6.86 24.94 1.89 0.25 -1.64 0.01 0.48 0.68 0.20 0.03
CHILE 9.38 24.53 1.20 0.03 -1.16 0.00 0.74 0.75 0.00 0.94
SOUTH KOREA 1.60 26.67 1.04 -0.72 -1.76 0.00 0.90 0.94 0.04 0.57
PORTUGAL 5.42 24.02 0.77 -0.11 -0.88 0.03 0.86 0.80 -0.06 0.45
MEXICO 7.02 22.59 0.77 -0.17 -0.94 0.00 0.82 0.85 0.03 0.66

96

Country
Q1
Dur
Q5
Dur
Q1
Alpha
Q5
Alpha
Q5-
Q1
Alpha
p-
value
Q1
Beta
Q5
Beta
Q5-
Q1
Beta
p-
value
PHILIPPINES 1.85 22.80 0.96 -0.21 -1.17 0.00 0.79 0.74 -0.05 0.41
SOUTH AFRICA 3.80 21.51 0.74 0.07 -0.67 0.04 0.81 0.90 0.09 0.12
INDONESIA 1.25 21.65 0.90 -0.82 -1.71 0.00 0.87 0.70 -0.17 0.00
INDIA -0.38 21.07 1.06 -0.06 -1.12 0.02 1.10 0.95 -0.14 0.08
BRAZIL 1.69 21.04 1.18 -0.02 -1.20 0.00 0.87 0.92 0.04 0.35
HUNGARY 6.55 21.83 1.35 -0.26 -1.61 0.00 0.67 0.82 0.16 0.02
GREECE 1.21 23.51 1.07 -0.48 -1.55 0.00 0.83 0.68 -0.15 0.04
COLOMBIA 3.24 20.68 0.87 -0.07 -0.93 0.03 0.73 0.70 -0.03 0.69
EGYPT 3.25 19.20 1.26 0.29 -0.96 0.07 0.73 0.81 0.09 0.28
TURKEY 2.41 19.96 1.54 0.19 -1.35 0.01 1.03 0.95 -0.08 0.19
RUSSIA -4.01 19.54 1.48 0.35 -1.13 0.06 0.82 0.64 -0.19 0.10
ARGENTINA 1.24 18.87 0.70 0.19 -0.51 0.46 0.67 0.63 -0.04 0.62
PAKISTAN -1.38 19.28 1.69 0.85 -0.84 0.05 0.84 0.73 -0.11 0.11

Correlation between Q5-Q1 Duration and Q5-Q1 Alpha: -.5833

97

Panel B: Value Weighted Portfolios (Countries Sorted by Duration)

Country
Q1
Dur
Q5
Dur
Q1
Alpha
Q5
Alpha
Q5-
Q1
Alpha
p-
value
Q1
Beta
Q5
Beta
Q5-
Q1
Beta
p-
value
SINGAPORE 22.15 34.02 0.96 -0.38 -1.35 0.00 0.92 1.01 0.09 0.26
CANADA 19.01 33.68 0.66 -0.64 -1.30 0.00 1.09 1.27 0.18 0.04
UNITED STATES 18.07 32.00 0.58 -0.11 -0.69 0.02 1.14 1.32 0.18 0.04
SWEDEN 20.02 34.35 0.94 -0.14 -1.08 0.01 0.91 0.96 0.05 0.56
NETHERLANDS 18.05 30.30 0.64 -0.07 -0.71 0.06 1.01 0.93 -0.07 0.34
FINLAND 21.30 30.66 1.51 0.52 -0.99 0.02 0.64 0.80 0.16 0.06
U.K. 16.60 32.20 0.83 -0.09 -0.92 0.01 1.11 1.03 -0.07 0.38
SWITZERLAND 17.34 30.00 0.84 0.01 -0.82 0.01 0.88 1.17 0.29 0.00
DENMARK 13.11 30.49 0.92 -0.09 -1.01 0.01 0.75 0.93 0.18 0.04
IRELAND 19.87 30.26 1.29 1.32 0.03 0.98 1.17 0.83 -0.35 0.03
GERMANY 17.64 31.47 0.92 0.23 -0.69 0.03 0.93 0.80 -0.13 0.06
FRANCE 13.20 30.44 0.78 0.25 -0.53 0.07 0.97 0.91 -0.06 0.40
CHINA 23.29 28.22 0.89 1.27 0.37 0.58 0.56 0.51 -0.05 0.64
NORWAY 10.35 31.94 0.65 -0.38 -1.04 0.02 0.96 0.99 0.03 0.71
JAPAN 7.95 28.63 0.85 0.05 -0.80 0.00 0.84 1.02 0.17 0.00
ISRAEL 16.75 28.44 1.09 0.44 -0.64 0.18 0.81 0.85 0.04 0.66
AUSTRALIA 15.22 32.19 0.50 -0.34 -0.84 0.02 1.10 1.13 0.04 0.59
QATAR 19.16 25.96 0.55 -0.03 -0.58 0.37 0.85 1.04 0.19 0.15
TAIWAN 18.25 27.40 0.89 -0.42 -1.31 0.00 1.06 1.07 0.01 0.83
AUSTRIA 14.31 29.13 0.59 0.27 -0.31 0.42 0.86 0.76 -0.10 0.18
MALAYSIA 11.37 26.55 0.56 0.26 -0.30 0.26 1.09 0.93 -0.16 0.03
NEW ZEALAND 15.62 29.13 0.87 0.34 -0.53 0.16 0.87 0.87 -0.01 0.93
THAILAND 11.16 26.28 1.01 0.37 -0.64 0.05 0.91 0.85 -0.05 0.40
SPAIN 13.94 26.67 0.48 0.25 -0.24 0.50 1.05 0.72 -0.33 0.00
SAUDI ARABIA 19.68 26.84 0.42 0.28 -0.14 0.86 0.60 0.67 0.07 0.60
BELGIUM 11.50 27.32 1.15 0.42 -0.73 0.04 0.85 0.66 -0.19 0.01
POLAND 11.71 26.63 1.35 -0.32 -1.67 0.00 0.97 0.87 -0.10 0.12
UAE 10.70 25.29 1.06 0.42 -0.64 0.23 0.87 0.54 -0.33 0.00
ITALY 9.04 25.92 0.26 0.33 0.06 0.85 1.14 0.86 -0.28 0.00
HONG KONG 3.06 26.68 1.20 -0.19 -1.39 0.00 1.02 0.98 -0.05 0.53
PERU 1.26 23.69 2.31 -0.09 -2.40 0.00 0.42 0.56 0.14 0.11
CZECH REPUB. 1.57 23.87 1.98 0.35 -1.63 0.00 0.46 0.81 0.35 0.00
CHILE 8.18 23.61 1.29 0.35 -0.94 0.00 0.86 0.96 0.11 0.05
SOUTH KOREA 2.29 25.50 0.94 -0.05 -0.98 0.01 0.96 1.00 0.04 0.55
PORTUGAL 3.99 23.70 0.66 0.18 -0.48 0.29 0.88 0.86 -0.01 0.89
MEXICO 6.28 22.10 0.81 0.04 -0.78 0.03 0.93 0.91 -0.02 0.79
PHILIPPINES 1.54 21.95 1.12 0.05 -1.07 0.01 0.99 0.79 -0.20 0.03
SOUTH AFRICA 2.66 21.10 0.82 0.08 -0.73 0.04 0.90 1.03 0.13 0.03
INDONESIA -0.83 20.94 0.34 -0.37 -0.70 0.09 0.96 0.88 -0.08 0.20

98

Country
Q1
Dur
Q5
Dur
Q1
Alpha
Q5
Alpha
Q5-
Q1
Alpha
p-
value
Q1
Beta
Q5
Beta
Q5-
Q1
Beta
p-
value
INDIA -1.16 20.59 0.75 0.14 -0.61 0.17 1.19 0.98 -0.21 0.01
BRAZIL 0.62 21.04 1.11 -0.11 -1.23 0.00 1.01 0.93 -0.08 0.14
HUNGARY 5.68 21.95 1.18 -0.86 -2.04 0.00 0.87 0.92 0.05 0.54
GREECE 0.07 22.68 0.83 -0.08 -0.91 0.12 1.00 0.78 -0.21 0.01
COLOMBIA 0.05 19.85 0.54 -0.30 -0.84 0.07 0.72 0.84 0.12 0.10
EGYPT 2.03 18.92 1.31 0.18 -1.13 0.03 0.75 0.87 0.12 0.11
TURKEY 3.12 19.56 1.00 0.05 -0.94 0.07 1.14 0.94 -0.20 0.00
RUSSIA -5.03 19.06 0.85 1.35 0.51 0.48 0.97 0.81 -0.15 0.23
ARGENTINA 0.02 19.13 0.99 -0.21 -1.20 0.13 0.74 0.63 -0.11 0.21
PAKISTAN -2.36 19.01 1.28 0.39 -0.89 0.04 0.98 0.75 -0.23 0.00

Correlation between Q5-Q1 Duration and Q5-Q1 Alpha: -.2624

This table presents portfolio Firm-Year Duration (Dur), CAPM Alphas (Alpha), and MKT Betas (Beta). The countries
are sorted on the VW Country-Year Duration, which is the same sorting used in Table 2. See Appendix A for more
detailed variable definitions. Panel A: Firm-Year Duration is calculated with the same formula used in Dechow et al.
(2004), but with the parameters defined in Table 2. However, unlike Tables 5, 7, and 9, in this table Firm-Year
Duration is computed at the firm-year level instead of the country-year level. Alpha is the intercept from a regression
of monthly portfolio returns minus the risk-free rate on the country’s value-weight market portfolio minus the risk-
free rate. The country’s market portfolio return is taken from the country specific MSCI index. Beta is the slope
estimate of the regression. The results are based on the regression Return-rf = α + β 1(Market-rf) + ε. P-values are
reported to identify statistical significance. Any difference, Q5-Q1, that is significant at the 10% or better is bolded.
In this panel the portfolios were formed with equal weighting for each firm. Panel B: In this panel the portfolios were
formed with a market value based weighting.

99

Table 12. Determinants of the Short Duration Premium at the Country-Level

Panel A: Determinants of Alpha Differential

(1) (2) (3) (4) (5) (6)
VARIABLES
Q5-Q1
Alpha EW
Q5-Q1
Alpha EW
Q5-Q1
Alpha EW
Q5-Q1
Alpha VW
Q5-Q1
Alpha VW
Q5-Q1
Alpha VW

SD -0.152*** -0.227*** -0.204*** -0.096** -0.177*** -0.174**

(-5.00) (-6.54) (-5.47) (-2.05) (-3.13) (-2.59)
ROE AR

-1.033 -0.911

-2.817* -2.554

(-1.04) (-1.00)

(-1.75) (-1.56)
ROE T

6.755*** 5.424**

6.762** 7.806*

(3.39) (2.34)

(2.09) (1.87)
G AR

1.012 0.570

2.004 1.791

(0.63) (0.36)

(0.77) (0.64)
G T

-15.135*** -11.049**

-16.108* -10.970

(-2.78) (-2.02)

(-1.82) (-1.11)
Developed

0.145

0.335

(1.05)

(1.34)
Short Dur. Industry %

-0.227

-0.150

(-0.34)

(-0.12)
Long Dur. Industry %

-0.911

-1.119

(-1.28)

(-0.87)
Largest Firms % -1.338** -0.499
(-2.11) (-0.43)

Observations 49 49 49 49 49 49
R-squared 0.348 0.502 0.626 0.083 0.219 0.278

100

Panel B: Determinants of Country-Level Duration Standard Deviation

(1)
VARIABLES SD

ROE AR
-1.389

(-0.36)
ROE T
24.018**

(2.66)
G AR
10.264

(1.62)
G T
-76.854***

(-3.91)
Developed
-0.265

(-0.45)
Short Dur. Industry %
-1.442

(-0.51)
Long Dur. Industry %
-7.582***

(-2.76)
Largest Firms %
8.544***

(3.69)

Observations
49
R-squared
0.540

This table tests potential determinants of the country-level alpha differentials computed in Table 11 (i.e., the short
duration premium). See Appendix A for more detailed variable definitions. T-statistics reported in paratheses. Panel
A: Q5-Q1 Alpha EW is computed in Table 11 Panel A. Q5-Q1 Alpha VW is computed in Table 11 Panel B. All other
variables are computed the same as in Table 2 and 3, except that they are now collapsed down to the sample average
for each country (i.e., one observation per country). Panel B: This panel tests potential determinants of SD, the
standard deviation of Firm-Year Duration for each country, as reported in Table 2.

101

Table 13. Equity Duration Portfolios of Country Indices

Panel A: Equal Weighted Full Sample

Country - EW 1 2 3 4 5 Q5-Q1 p-value
Country-Year Dur. 10.55 15.43 20.26 23.49 26.46 15.91

Total Vol 5.72 5.40 4.97 4.91 5.05 -0.67 0.04
IVol 3.49 2.92 2.02 1.79 1.18 -2.31 0.00
Excess Return 0.69 0.47 0.27 0.34 0.34 -0.35 0.45
CAPM Alpha 0.38 0.16 -0.04 0.03 0.00 -0.38 0.10
MKT Beta 1.00 1.00 1.00 1.01 1.08 0.08 0.11

Panel B: Equal Weighted Excluding Q1 2020

Country - EW 1 2 3 4 5 Q5-Q1 p-value
Country-Year Dur. 10.55 15.43 20.26 23.49 26.46 15.91

Total Vol 5.59 5.31 4.87 4.85 4.95 -0.64 0.05
IVol 3.49 2.92 2.02 1.79 1.18 -2.31 0.00
Excess Return 0.81 0.58 0.38 0.42 0.44 -0.36 0.43
CAPM Alpha 0.41 0.18 -0.02 0.02 0.01 -0.40 0.08
MKT Beta 0.98 1.00 1.00 1.01 1.08 0.10 0.06

This table presents portfolios holding country-level indices and formed on Country-Year Duration. See Appendix A
for more detailed variable definitions. Panel A: Country-Year Duration is calculated with the same formula used in
Dechow et al. (2004), but with the parameters defined in Table 2. Each of the quintile portfolios holds indices for
either 9 or 10 of the 49 countries in the sample. The portfolios are equal weighted between countries. Total Vol is the
standard deviation of monthly portfolio returns. IVol is the standard deviation of the residual of the model. Excess
Return is the portfolio return in excess of the U.S. risk-free rate. CAPM Alpha is the intercept from a regression of
monthly portfolio returns minus the risk-free rate on the world (i.e., MSCI ACWI) value-weight market portfolio
minus the risk-free rate. MKT Beta is the slope estimate of the regression. The results are based on the regression
Return-rf = α + β 1(Market-rf) + ε. P-values are reported in the furthest right column of each panel. Any difference,
Q5-Q1, that is significant at the 10% level or better is bolded. Panel B: Replications the results of Panel A except Q1
2020 market data is removed as this is a period in which the short duration premium is shown to be reversed (Dechow
et al. 2021).

102

Table 14. Alternative Discount Rates and Forecasting Assumptions

(1) (2) (3) (4) (5) (6)
VARIABLES
Pred.
Sign
C-Y
Duration
Implied R
C-Y
Duration
Implied R
C-Y
Duration
Implied R
C-Y
Duration
Fixed FA
C-Y
Duration
Fixed FA
C-Y
Duration
Fixed FA

Governance (WGI)
+ 0.085***

0.003

(5.80)

(0.56)

Tax Loss Carryforward
+

0.042**

0.010*

(2.39)

(1.66)

Innovation Box
+

2.207***

0.643***

(4.30)

(4.00)
Developed
+
3.080*** 5.269***

0.697*** 0.777***

(5.49) (11.63)

(3.72) (5.12)

Inflation
-
-0.091** -0.110*** -0.214*** 0.002 0.011 -0.004

(-2.29) (-2.95) (-5.44) (0.13) (0.91) (-0.31)
Short Dur. Industry %
-
-8.528*** -9.752*** -22.358*** -2.475*** -2.296*** -4.368***

(-3.40) (-3.84) (-8.72) (-2.95) (-2.70) (-5.44)
Long Dur. Industry %
+
23.704*** 29.396*** 34.315*** 7.254*** 7.879*** 8.293***

(10.11) (12.59) (14.00) (9.25) (10.09) (10.80)
Largest Firms %
?
-18.203*** -17.398*** -14.347*** -5.573*** -5.550*** -4.897***

(-9.06) (-8.64) (-6.58) (-8.29) (-8.24) (-7.17)

Observations

976 1,004 1,004 976 1,004 1,004
R-squared

0.456 0.450 0.356 0.211 0.218 0.198

This table is designed to test the robustness of the primary results and to changes in forecasting assumptions used in
computing Firm-Year Duration and therefore Country-Year Duration. See Appendix A for more detailed variable
definitions. T-statistics reported in paratheses. Columns 1-3 replicate the results of Column 2 in Tables 5, 7, and 9
after replacing the independent variable, formerly Country-Year Duration, with C-Y Duration Implied R. This new
dependent variable follows a similar computation process as Country-Year Duration, except the ROE T and R are set
equal to the implied discount rate for each firm-year. The implied discount rate is solved for using a root solving
algorithm where the market value is equal to the sum of the forecasted cash flows to shareholders. Additionally, ROE
AR, G AR, and G T are set equal to their global means. Columns 4-6 replicate the results of Column 2 in Tables 5, 7,
and 9 after replacing the independent variable, formerly Country-Year Duration, with C-Y Duration Fixed FA. This
new dependent variable follows a similar computation process as Country-Year Duration, except R is set equal to 0%.
Additionally, ROE AR, ROE T, G AR, and G T are set equal to their global means.

103

Table 15. Country-Level WGI Components and Country-Level Duration

Panel A: Univariate Tests

(1) (2) (3) (4) (5) (6)
VARIABLES
Pred.
Sign
Country-
Year
Duration
Country-
Year
Duration
Country-
Year
Duration
Country-
Year
Duration
Country-
Year
Duration
Country-
Year
Duration

Voice and Accountability (VA)
? 0.086***

(13.82)

Political Stability and Absence of
Violence/Terrorism (PV)
+

0.121***

(23.46)

Government Effectiveness (GE)
+

0.220***

(34.36)

Regulatory Quality (RQ)
+

0.195***

(30.18)

Rule of Law (RL)
+

0.174***

(31.89)

Control of Corruption (CC)
+

0.174***

(32.95)

Observations

976 976 976 976 976 976
R-squared
0.164 0.361 0.548 0.483 0.511 0.527

104

Panel B: Tests with Control Variables

(1) (2) (3) (4) (5) (6)
VARIABLES
Pred.
Sign
Country-
Year
Duration
Country-
Year
Duration
Country-
Year
Duration
Country-
Year
Duration
Country-
Year
Duration
Country-
Year
Duration

Voice and Accountability
(VA)
? -0.043***

(-6.69)

Political Stability and
Absence of
Violence/Terrorism (PV)
+

0.041***

(7.39)

Government
Effectiveness (GE)
+

0.131***

(12.10)

Regulatory Quality (RQ) +

0.090***

(9.23)

Rule of Law (RL) +

0.087***

(9.71)

Control of Corruption
(CC)
+

0.096***

(11.42)
Developed + 6.212*** 3.643*** 1.844*** 2.658*** 2.259*** 1.850***
(16.89) (11.09) (5.00) (7.26) (5.87) (4.88)
Inflation - -0.218*** -0.187*** -0.141*** -0.135*** -0.152*** -0.156***
(-7.94) (-6.75) (-5.25) (-4.77) (-5.51) (-5.81)
Short Dur. Industry %
- -11.281*** -8.414*** -5.503*** -6.676*** -8.198*** -6.082***

(-6.51) (-4.78) (-3.21) (-3.78) (-4.77) (-3.53)
Long Dur. Industry %
+ 23.913*** 20.481*** 16.703*** 20.030*** 17.690*** 18.344***

(14.53) (12.55) (10.36) (12.45) (10.71) (11.50)
Largest Firms %
? -6.718*** -7.727*** -9.553*** -8.784*** -8.980*** -7.957***
(-4.67) (-5.42) (-6.96) (-6.24) (-6.40) (-5.79)

Observations

976 976 976 976 976 976
R-squared

0.589 0.593 0.626 0.605 0.608 0.621

105

Panel C: Collapsed Panel with Control Variables

(1) (2) (3) (4) (5) (6)
VARIABLES
Pred.
Sign
Country-
Year
Duration
Country-
Year
Duration
Country-
Year
Duration
Country-
Year
Duration
Country-
Year
Duration
Country-
Year
Duration

Voice and Accountability
(VA)
? -0.048*

(-1.94)

Political Stability and
Absence of
Violence/Terrorism (PV)
+

0.040*

(1.72)

Government
Effectiveness (GE)
+

0.130***

(2.78)

Regulatory Quality (RQ) +

0.068

(1.47)

Rule of Law (RL) +

0.074*

(1.92)

Control of Corruption
(CC)
+

0.095***

(2.73)
Developed + 5.691*** 3.064** 1.366 2.668* 2.105 1.355
(3.94) (2.36) (0.93) (1.78) (1.37) (0.91)
Inflation - -0.523*** -0.439** -0.341* -0.357* -0.378** -0.374**
(-3.14) (-2.54) (-2.00) (-1.80) (-2.10) (-2.24)
Short Dur. Industry %
- -9.380 -6.978 -4.432 -6.807 -7.402 -4.806

(-1.42) (-1.02) (-0.67) (-0.98) (-1.10) (-0.73)
Long Dur. Industry %
+ 21.227*** 18.096*** 14.813** 18.600*** 16.266** 16.293**

(3.31) (2.81) (2.35) (2.87) (2.48) (2.63)
Largest Firms %
? -6.635 -7.804 -9.657* -8.670 -8.877 -8.023
(-1.21) (-1.41) (-1.83) (-1.55) (-1.61) (-1.53)

Observations

49 49 49 49 49 49
R-squared

0.705 0.699 0.728 0.694 0.704 0.726

This table replicates the results of Columns 1, 2, and 5 in Table 5. However, the primary independent variable of
interest, formerly the composite measure Governance (WGI), is replaced by each of the 6 WGI measures individually
to determine which of them are driving the primary result. See Appendix A for more detailed variable definitions. All
WGI measures have value ranges from 0 to 100 where 100 is the “best”. Panels A, B, and C are standard OLS
regressions with t-statistics reported in paratheses. In Panel C, the variables are collapsed to a single observation per
county by using the sample averages for each country for each variable.

106

Table 16. Country-Level Economy Policy Uncertainty and Country-Level Duration

(1) (2) (3) (4)
VARIABLES
Pred.
Sign
Country-Year
Duration
Country-Year
Duration
Country-Year
Duration
Country-Year
Duration

Economic Policy Uncertainty (EPU)
-
0.625 -0.692** 1.420* -1.027***

(0.96) (-1.97) (1.75) (-2.85)
Developed
+
5.927*** 6.370***

(10.70) (8.62)
Inflation
-
-0.628*** -0.675***

(-8.56) (-8.75)
Short Dur. Industry %
-
-17.741*** -18.552***

(-5.46) (-4.96)
Long Dur. Industry %
+
23.312*** 9.405***

(9.79) (2.75)
Largest Firms %
?
-15.478*** -1.343

(-8.70) (-0.34)

Observations

459 459 252 252
R-squared

0.002 0.722 0.012 0.825

This table tests the relationship between country-year Economic Policy Uncertainty (EPU) and Country-Year
Duration. See Appendix A for more detailed variable definitions. Country-Year Duration is calculated with the same
formula used in Dechow et al. (2004), but with the parameters defined in Table 2. Economic Policy Uncertainty (EPU)
is the natural log of the economic policy uncertainty index originally developed by Baker, Bloom and Davis (2016).
All control variables have the same definitions as in Tables 5, 7, and 9. Columns 1 and 2 are the results for the full
sample of 22 countries that the EPU measure is available for. This measure is computed by different co-author groups
for some countries, so Columns 3 and 4 repeat the analysis on the subsample of 12 countries in which the EPU data
comes from a single source (i.e., Baker, Bloom and Davis 2016). T-statistics reported in paratheses.

107

Table 17. Hofstede's Country-Level Cultural Dimensions and Country-Level Duration

Panel A: Univariate Tests

(1) (2) (3) (4) (5) (6)
VARIABLES
Pred.
Sign
Country-
Year
Duration
Country-
Year
Duration
Country-
Year
Duration
Country-
Year
Duration
Country-
Year
Duration
Country-
Year
Duration

Power Distance (PDI)
?
-0.123***

(-15.58)
Individualism (IDV)
?
0.113***

(16.17)
Masculinity (MAS)
?
-0.037***

(-3.61)
Uncertainty Avoidance (UAI)
-
-0.116***

(-16.26)
Long-term Orientation (LTOWVS)
+
0.042***

(5.07)
Indulgence (IVR)
?
0.109***

(12.45)

Observations

981 981 981 981 1,006 985
R-squared
0.199 0.211 0.013 0.213 0.025 0.136

108

Panel B: Tests with Control Variables

(1) (2) (3) (4) (5) (6)
VARIABLES
Pred.
Sign
Country-
Year
Duration
Country-
Year
Duration
Country-
Year
Duration
Country-
Year
Duration
Country-
Year
Duration
Country-
Year
Duration

Power Distance (PDI)
?
0.000

(0.02)
Individualism (IDV)
?
0.000

(0.02)
Masculinity (MAS)
?
-0.011

(-1.61)
Uncertainty Avoidance (UAI)
-
-0.063***

(-12.14)
Long-term Orientation (LTOWVS)
+
-0.025***

(-3.80)
Indulgence (IVR)
?
0.054***

(8.26)
Developed
+
5.165*** 5.155*** 5.140*** 4.619*** 5.149*** 4.604***

(13.83) (14.77) (17.47) (16.63) (16.92) (14.85)
Inflation
-
-0.173*** -0.173*** -0.175*** -0.162*** -0.210*** -0.199***

(-6.35) (-6.31) (-6.46) (-6.43) (-7.77) (-7.49)
Short Dur. Industry %
-
-9.578*** -9.572*** -9.502*** -7.312*** -5.118*** -4.654***

(-6.14) (-5.93) (-6.16) (-5.04) (-3.15) (-2.91)
Long Dur. Industry %
+
22.630*** 22.620*** 22.451*** 21.482*** 24.183*** 21.155***

(13.87) (14.09) (14.02) (14.37) (13.46) (13.28)
Largest Firms %
?
-7.585*** -7.581*** -7.792*** -9.375*** -4.583*** -2.951**

(-5.14) (-5.24) (-5.38) (-6.93) (-3.06) (-1.97)

Observations

981 981 981 981 1,006 985
R-squared

0.583 0.583 0.584 0.637 0.572 0.591

109

Panel C: Collapsed Panel with Control Variables

(1) (2) (3) (4) (5) (6)
VARIABLES
Pred.
Sign
Country-
Year
Duration
Country-
Year
Duration
Country-
Year
Duration
Country-
Year
Duration
Country-
Year
Duration
Country-
Year
Duration

Power Distance (PDI)
?
-0.004

(-0.12)
Individualism (IDV)
?
0.007

(0.25)
Masculinity (MAS)
?
-0.014

(-0.52)
Uncertainty Avoidance (UAI)
-
-0.060***

(-3.00)
Long-term Orientation (LTOWVS)
+
-0.033

(-1.28)
Indulgence (IVR)
?
0.051**

(2.03)
Developed
+
4.179** 4.093*** 4.260*** 3.840*** 4.144*** 3.739***

(2.68) (2.79) (3.49) (3.44) (3.31) (2.96)
Inflation
-
-0.435** -0.443** -0.441** -0.409** -0.525*** -0.498***

(-2.45) (-2.44) (-2.50) (-2.57) (-3.08) (-2.99)
Short Dur. Industry %
-
-9.804 -9.061 -9.505 -7.187 -4.470 -4.426

(-1.44) (-1.26) (-1.41) (-1.17) (-0.60) (-0.60)
Long Dur. Industry %
+
21.597*** 21.547*** 21.471*** 20.685*** 23.798*** 19.265***

(3.03) (3.07) (3.09) (3.29) (3.12) (2.86)
Largest Firms %
?
-7.967 -8.225 -8.378 -9.855* -4.956 -3.433

(-1.35) (-1.42) (-1.45) (-1.88) (-0.84) (-0.58)

Observations

47 47 47 47 48 47
R-squared

0.692 0.692 0.694 0.748 0.698 0.712

This table tests the relationship between each of Hofstede’s six country-level dimensions of culture and Country-Year
Duration. See Appendix A for more detailed variable definitions. Country-Year Duration is calculated with the same
formula used in Dechow et al. (2004), but with the parameters defined in Table 2. Each of the six dimensions of culture
are available as a single value per country for the full sample period. Each of the six dimensions of culture are rescaled
between 0 and 100. The sample size varies between columns because the six dimensions are not all measured for each
country. All control variables have the same definitions as in Tables 5, 7, and 9. Panels A, B, and C are standard OLS
regressions with t-statistics reported in paratheses. In Panel C, the variables are collapsed to a single observation per
county by using the sample averages for each country for each variable.

110

Table 18. Country-Level GDP per Capita and Country-Level Duration

(1) (2) (3)
VARIABLES
Pred.
Sign
Country-Year
Duration
Country-Year
Duration
Country-Year
Duration

GDP Per Capita
+
3.443*** 1.624*** 1.656**

(29.76) (10.16) (2.36)
Developed
+
2.703*** 1.624

(6.96) (1.01)
Inflation
-
-0.169*** -0.447**

(-6.41) (-2.67)
Short Dur. Industry %
-
-4.159*** -3.509

(-2.68) (-0.50)
Long Dur. Industry %
+
19.172*** 19.889**

(9.45) (2.29)
Largest Firms %
?
-6.139*** -7.114

(-4.24) (-1.25)

Observations

1,002 1,002 48
R-squared

0.470 0.593 0.712

This table tests the relationship between country-year GDP Per Capita and Country-Year Duration. See Appendix A
for more detailed variable definitions. Country-Year Duration is calculated with the same formula used in Dechow et
al. (2004), but with the parameters defined in Table 2. GDP Per Capita is the natural log of the GDP per capita as
reported by the World Bank for each country-year. All control variables have the same definitions as in Tables 5, 7,
and 9. In Column 3, the variables are collapsed to a single observation per county by using the sample averages for
each country for each variable.

Abstract (if available)

Abstract I hypothesize that short and long duration firms are more prevalent in certain economies. Short duration firms return cash flows to investors more quickly than do long duration firms. Therefore, I predict the predominance of long duration firms where the risk to investors' capital is lower. I study a sample of 23 developed and 26 emerging countries and find that, in equilibrium, economies have more long duration firms where: financial markets are more developed, inflation is lower, the government is stable and promotes private sector development, agents have confidence in contract enforcement and property rights, and corruption is controlled. I also predict and find that countries with tax policies that are more future-focused (e.g., longer tax loss carryforward periods) have longer duration firms on average. I next examine whether the duration premium (i.e., the negative relation between returns and equity duration) is observable within and across countries. The results indicate that the duration premium is observable in most countries and that a country's duration premium is correlated with the range of duration in that country. Overall, these results highlight that duration is a relevant risk factor both within and across countries and that a country's institutional setting and level of development can impact corporate duration.

Linked assets

University of Southern California Dissertations and Theses

Conceptually similar

PDF

Understanding the disclosure practices of firms affected by a natural disaster: the case of hurricanes

PDF

Disclosure distance and earnings announcement returns

PDF

Essays on the firm and stakeholders relationships: evidence from mergers & acquisitions and labor negotiations

PDF

Essays on the effect of cognitive constraints on financial decision-making

PDF

Voluntary disclosure responses to mandated disclosure: evidence from Australian corporate tax transparency

PDF

Three essays on macro and labor finance

PDF

Picky customers and expected returns

PDF

Essays in empirical asset pricing

PDF

A case Study of gender gaps in the legal profession

PDF

Does mandatory adoption of International Accounting Standards reduce the cost of equity capital?

PDF

Expectation dynamics and stock returns

PDF

Changing fundamental analysis in the new economy: the case of DuPont analysis and STEM firms

PDF

Bridging the gap of gender equity in entrepreneurship: an evaluation study

PDF

Three essays in derivatives, trading and liquidity

PDF

Shareholder litigation as a disciplining device: evidence from firms' financial policies

PDF

Investment behavior of mutual fund managers

PDF

Can I borrow your footnotes? Learning and network benefits of footnote similarity

PDF

Executive compensation: the trend toward one size fits all

PDF

Social movements and access to credit

PDF

Three essays in international macroeconomics and finance

Asset Metadata

Creator Gardner, Jesse Dempsey, IV (author)

Core Title Implied equity duration: international evidence

School Marshall School of Business

Degree Doctor of Philosophy

Degree Program Business Administration

Degree Conferral Date 2023-08

Publication Date 06/14/2023

Defense Date 06/09/2023

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

Tag accounting,equity duration,financial statement analysis,governance,international,OAI-PMH Harvest,taxation

Format theses (aat)

Language English

Contributor Electronically uploaded by the author (provenance)

Advisor Dechow, Patricia (committee chair), Schmid, Lukas (committee member), Sloan, Richard (committee member)

Creator Email jdgardner4@gmail.com,jesse.gardner@usc.edu

Permanent Link (DOI) https://doi.org/10.25549/usctheses-oUC113170478

Unique identifier UC113170478

Identifier etd-GardnerJes-11957.pdf (filename)

Legacy Identifier etd-GardnerJes-11957

Document Type Dissertation

Format theses (aat)

Rights Gardner, Jesse Dempsey, IV

Internet Media Type application/pdf

Type texts

Source 20230613-usctheses-batch-1055 (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.

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