Close
About
FAQ
Home
Collections
Login
USC Login
Register
0
Selected
Invert selection
Deselect all
Deselect all
Click here to refresh results
Click here to refresh results
USC
/
Digital Library
/
University of Southern California Dissertations and Theses
/
Essays on fiscal outcomes of cities in California
(USC Thesis Other)
Essays on fiscal outcomes of cities in California
PDF
Download
Share
Open document
Flip pages
Contact Us
Contact Us
Copy asset link
Request this asset
Transcript (if available)
Content
1 Essays on Fiscal Outcomes of Cities in California A Dissertation August 2013 Cheongsin Kim University of Southern California 2 DEDICATION To my beloved mom, Geunwha Lee 3 TABLE OF CONTENTS DEDICATION 2 LIST OF TABLES 5 LIST OF FIGURES 6 ABSTRACT 7 CHAPTER 1. INTRODUCTION 10 1. RESEARCH QUESTION 10 2. OVERVIEW 12 3. CONCLUSION 14 CHAPTER 2. The Effects of Recessions on Contracting Moderated by Institutional Arrangements of Government: Evidence from California Cities, 1993-2009 16 1. INTRODUCTION 16 2. LITERATURE REVIEW AND HYPOTHESES 20 3. DESCRIPTIVE TRENDS 31 4. DATA AND METHODS 34 5. RESULTS 41 6. DISCUSSION 44 7. CONCLUSION 48 APPENDIX 2-1: Correlations 50 CHAPTER 3. The Effects of Fiscal Conditions and Intergovernmental Dependence on Non-collective Services: Evidence from California Cities, 1992-2010 51 1. INTRODUCTION 51 2. NATURE OF CITY SERVICES 53 3. LITERATURE REVIEW AND HYPOTHESES 59 4. DATA AND METHODS 66 4 5. RESULTS 73 6. DISCUSSION 76 7. CONCLUSION 78 APPENDIX 3-1: Correlations 79 CHAPTER 4. The Fiscal Effects of Term-limits, Council-manager Form, and Charters: Evidence from California Cities, 1992-2009 80 1. INTRODUCTION 80 2. LITERATURE REVIEW AND HYPOTHESES 81 3. DATA AND METHODS 87 4.RESULTS 93 5. DISCUSSION 95 6. CONCLUSION 99 APPENDIX 4-1: Correlations 101 APPENDIX 4-2: Approval Year of Term-limits on Council 102 CHAPTER 5. CONCLUSION 103 1. FINDINGS SUMMARY 103 2. THEORETICAL CONTRIBUTIONS 104 3. POLICY IMPLICATIONS 106 4. LIMITATIONS AND FUTURE RESEARCH 107 REFERENCES 109 5 LIST OF TABLES TABLE 2-1: Variables, Measures, and Sources of Data 35 TABLE 2-2: Summary Statistics of Base Variables 39 TABLE 2-3: Fixed Effects (Within) Regression Coefficients 44 TABLE 3-1: Classification of City Services 58 TABLE 3-2: Variables, Descriptions, and Sources of Data 67 TABLE 3-3: Descriptive Statistics of Individual Non-collective Services 68 TABLE 3-4: Descriptive Statistics of Base Variables 72 TABLE 3-5: Fixed Effects (Within) Regression Coefficients 76 TABLE 4-1: Variables, Descriptions, and Sources of Data 91 TABLE 4-2: Descriptive Statistics 92 TABLE 4-3: Fixed Effects (Within) Regression Coefficients 95 6 LIST OF FIGURES FIGURE 2-1: Recession and Contracting, All Governments 32 FIGURE 2-2: Council-manager vs. Mayor-council Governments 33 FIGURE 2-3: Charter vs. General-law Governments 34 FIGURE 2-4: The Effects of Recession on Contracting Moderated by Institutional Arrangements 44 FIGURE 3-1: Structure of Fund Balance 70 FIGURE 3-2: Average Trends of Major Variables 72 7 ABSTRACT This dissertation seeks to promote understanding of the critical dimensions of government fiscal behavior from the interdisciplinary rational-choice perspective. A comprehensive panel dataset containing economic, institutional, and fiscal information is constructed on California cities, beginning in the early 1990s, and then analyzed with advanced econometric techniques. This dissertation includes three main self-contained essays (Chapters 2, 3, and 4), an introduction (Chapter 1), and a conclusion (Chapter 5). Chapter 2 aims to uncover the effect of recession on cities' service contracts, moderated by two sets of institutional variables, i.e., whether the city is the council-manager or the mayor- council form, and whether the city is a general law or charter city. Drawing on the literatures of public finance and management, it is hypothesized that economic recession, the council-manager form, and the charter status all have positive effects on contracting as opposed to economic expansion, the mayor-council form, and the general-law status in that order. The hypotheses are tested on all available California cities over 17 years, from 1993 to 2009. The result indicates that recession decreases contracting in governments with the mayor-council and general-law, but the extent of the decrease is moderated if they adopt either the council-manager or the charter form of government. Finally, recession turns out to increase contracting in governments with both characteristics. This implies that the theoretical expectation of more contracting in the face of recession is based on the premise that government functions under the institutional constellation structurally capable of honoring what average residents want. Chapter 3 seeks to reveal the fiscal behaviors of local governments influenced by external or internal fiscal conditions and intergovernmental dependence. It attempts to classify 36 city services into collective (public, common-pool) and non-collective (toll, private) goods. Inspired 8 by the scholarship on collective goods, both external (state surplus/deficit) and internal (city surplus/deficit, unreserved fund balance) fiscal conditions are hypothesized to have a positive relationship on the provision of non-collective services. Drawing on the fiscal federalism literature, it is predicted that a growing intergovernmental dependence scales down such provision. Two different types of intergovernmental dependence are expected to produce differential incentive effects. The analysis of all available California cities during 1992-2010 indicates a weakly positive effect of external fiscal conditions but also an unexpected negative effect of internal conditions. This suggests that different dynamics of fiscal conditions exist. It further confirms the predicted inverse association of intergovernmental dependence and the presence of differential incentive effects. Chapter 4 explores the fiscal effects of term-limits, council-manager form, and charter status. The council-manager and the charter forms of government are hypothesized to maintain positive associations with fiscal performance, while the direction of term-limits is considered open. A rigorous empirical analysis is attempted on around 250 California cities, 1992-2009. The analysis reveals that term-limits are not associated in either direction with fiscal performance. This implies that 1) the positive and negative effects may have canceled each other out or 2) term-limits may not be a fiscal but a political institution. The analysis provides evidence for the beneficial role of the council-manager form regarding fiscal performance. The charter form shows mixed impacts. This dissertation strengthens the general public administration/management/policy and public budgeting & finance literatures by revealing the dynamics between economic/fiscal/institutional forces and localities. This project produces useful policy recommendations for numerous decision-making instances such as whether a new institution 9 must be adopted or how much intergovernmental dependence would be desirable. The geographical limitation of this project demands subsequent studies beyond California. 10 CHAPTER 1. INTRODUCTION 1. RESEARCH QUESTION This dissertation project, consisting of three independent empirical essays, aims to better understand how government fiscal behavior is influenced by a variety of forces - macroeconomic, fiscal, and institutional. While government fiscal behavior has been addressed within the scholarship of public budgeting and finance in particular and that of general public administration/management/policy, a great number of questions remain unanswered. This project strives to address some of the questions, which are theoretically informative and practically relevant, with theories from political science and economics as well as public administration as follows: First, this dissertation project seeks to better understand the mechanism of government response influenced by the external fiscal/economic environment throughout Chapters 2, 3, and 4. Local governments are vulnerable to the surrounding environment. An environmental change sometimes jeopardizes their survival but also prompts them to innovate. For example, the fiscal stress created by a declining economy may drive governments to find alternative methods of service provision, e.g., contracting. While the issues of fiscal stress are talked about sporadically in the literature, they are not particularly emphasized after Levine et al. (1981). Second, this dissertation project (Chapter 2) seeks to advance the existing research on government contracting by employing a unique body of longitudinal fiscal data, which captures the changing size of contracting precisely over time. It is correct that the theme of contracting itself is widely explored in the existing research, yet it still has substantial room for further advancement, because typical contract data are discrete and cross-sectional and this characteristic further limits the scope of applicable techniques. 11 Third, this research project (Chapter 2) seeks to reveal the moderating effect of four combined institutional arrangements, i.e., council-manager and charter, council-manager and general-law, mayor-council and charter, and mayor-council and general-law, on the relationship between the two variables. Although it is evident that contract decisions are made not out of vacuum but under a certain combined institutional arrangement, the question of how to measure its impact remains challenging. This project proposes a measurement strategy via double interaction of recession with the two basic government institutions (council-manager vs. mayor- council, charter vs. general-law). Fourth, this project (Chapter 3) strives to demonstrate that the idea of collective (public goods, common-pool resources) and non-collective goods (private goods, toll goods) (e.g., E. Ostrom, 2005, p.24) has great potential for empirical research. The conceptual distinction by two properties (excludability, subtractability) has already been discussed (e.g., E. Ostrom, 2005, p.24) but its empirical application is sparse. While an environmental shock would have not a uniform effect but rather distinguish one set of services from the other, the existing research typically considers government services as a whole (e.g., Wang and Hou, 2012) or picks up a certain set of services with little theoretical guidance (e.g., Karuppusamy and Carr, 2012). This project attempts to test the prediction that non-collective services are cut due to an environmental shock on a disproportionately large scale compared to collective services. Fifth, this project (Chapter 3) makes an effort to draw attention to the financial implications of collective and non-collective services; the latter are self-financeable at least in part, while the former largely lack this useful characteristic (Hyman, 2002, p.142). In the era of tax revolt, the strategic value of self-financing looks substantial. This logic is employed in explaining the relationship of intergovernmental fiscal dependence with non-collective services. 12 Lastly, this project (Chapters 2, 4) illuminates two local institutions which are rarely addressed so far: term-limits and charter status. The primary locus of study on term-limits is the U.S. state governments (e.g. Besley and Case, 1995). The normative significance of charter status has been widely discussed (e.g., V. Ostrom et al., 1988, p.34), but its empirical application is scant. This project builds a relevant dataset on local term-limits from various sources and on charter status from the Annual Cities Report, published by the California State Controller. 2. OVERVIEW Chapter 2 This chapter grapples with gauging the recessionary impact on contracting moderated by the four combined institutional arrangements, i.e., council-manager and charter, council-manager and general-law, mayor-council and charter, mayor-council and general-law. The four are created from intersecting the two representative institutions of local government: council- manager vs. mayor-council, charter vs. general-law. The council-manager is argued to be more suitable than the mayor-council for loyally honoring the fiscal preferences of residents. Specifically, the arrival of a recession tends to make average residents prefer the reduction of service production costs, a reduction that can be achieved via contracting (Ferris and Graddy, 1994, p.129). This fiscal preference is more likely to be respected in the former type of government than in the latter. The reason lies in the discriminating incentives to the administrative chief, the city manager in the former and the mayor in the latter. The former offers the city manager the low-powered incentive, i.e., prospects for career success, while the latter offers the mayor the high-powered incentive: reelection (Feiock and Jang, 2009, p.671). In other words, the latter is politically charged, which heightens the likelihood of disregarding the fiscal preference of the average citizen. 13 Charter governments are regarded as having a greater policy capacity to accommodate the fiscal preferences of average residents than general-law governments (Benton, 2002, p.474). Unlike the latter, in particular, the former is exempt from California state requirements to pay prevailing wages and to put through an official competitive bidding process for awarding a public works project. 1 Additionally, two political factors, Democratic governor and votes, are also assessed, inspired by the argument that Democrats are less willing to contract services out than Republicans (Clark and Ferguson, 1983, p.252). Chapter 3 This chapter identifies two categories of service, collective (18) and non-collective (18), from a total of 36 city services according to high or low excludability. It argues that an outside (state surplus/deficit) or inside (city surplus/deficit, unreserved fund balance) fiscal decline makes local governments prioritize cutting the provision of non-collective rather than collective services. The reason is that the former typically does not constitute the core responsibility of government according to the scholarship on collective services (e.g., Samuelson, 1954). On the other hand, non-collective services hold a good property, i.e., that they are self- financeable at least partly in the form of user fees and charges (Cornes and Sandler, 1996, p.34). This chapter predicts that higher dependence on the state or federal government causes less provision of non-collective services, since high dependence erodes a government's fiscal discipline (Weingast, 2009, p.287) and thus makes it slow in seeking extra revenue possibilities over time. Two forms of intergovernmental transfer are common: general and functional. General intergovernmental revenue is hypothesized to have a greater inverse impact on non-collective service provision than its functional counterpart. This is because the former is expected to 1 League of California Cities, General Law City v. Charter City, www.cacities.org/chartercities, accessed 06/16/2012. 14 enervate fiscal discipline to a greater extent than the latter due to the former's unlimited range of spending. Chapter 4 Three popular local government institutions are placed under close scrutiny: term-limits, council-manager form, and charter status. As for local term-limits, first, two competing arguments are introduced. The affirmative view claims a positive fiscal impact of term-limits because these elevate electoral competiveness (Daniel and Lott, 1997, pp.168-169) and an incumbent's resolve to make a right decision on behalf of average residents (Schelker, 2012, pp.31-32). The unfavorable view predicts a negative fiscal impact, because term-limits induce moral hazard in an incumbent and take away from residents their right to decide whether to use his expertise, augmented through tenure (Alt, Bueno de Mesquita and Rose, 2011, p.171). As the two views compete, the direction of the prediction is considered open. Drawing on the same argument from Chapter 2, council-manager governments are hypothesized to achieve better fiscal performance than mayor-council ones. Likewise, charter governments are also predicted to accomplish better fiscal performance than general-law ones. 3. CONCLUSION The research questions on government fiscal behavior will be explored thoroughly in the upcoming Chapters 2, 3, and 4. The broadly defined rational choice tradition across political science, economics, and public administration will be utilized. Rigorous quantitative analysis will be conducted on an extensive panel dataset for cities in California beginning in the early 1990s. Following the essay format of a dissertation, each chapter is designed to be independent. Each tackles substantially different fiscal aspects and is organized in a self-contained manner with its own sections of literature review, data and methods, analysis results, and discussion. 15 Chapter 5 provides an overall summary, theoretical contributions, and policy implications, and acknowledges room for subsequent research. 16 CHAPTER 2. The Effects of Recessions on Contracting Moderated by Institutional Arrangements of Government: Evidence from California Cities, 1993-2009 1. INTRODUCTION City governments have been hit hard by the recession from January 2008 through June 2009 2 , creating an urgent need for practical policy recommendations. Yet, to offer useful recommendations should be preceded by scientific research on the consequences of recent recessions for city governments. In response to the recent call for research connecting the public finance and public management literatures (Kioko et al., 2011), this paper is designed to contribute to both. The literature of public finance (for an excellent literature review, see Mullins and Pagano, 2005) has contributed to our understanding of the fiscal behaviors of sub-national governments. Its major themes include fund balances (Marlowe, 2005; Hendrick, 2006; Hou and Moynihan, 2008), tax and expenditure limitations (Mullins and Wallin, 2004; Hoene, 2004; McCubbins and Moule, 2010), and fiscal federalism (Weingast, 2009; Oates, 2008). Unfortunately, the research on contracting does not seem to be very popular in this literature, evidenced by the low visibility of contracting studies in the journal, Public Budgeting and Finance (for an exception, see Rubin, 2006). The present study aims to be a relevant addition to this literature by exploring how city governments use contracting as a financial tool for relieving fiscal stress. By contrast, contracting has been enjoying wide popularity as a major research topic in public management. It is considered one of the major indirect service arrangements (Stein, 1990, p.49), as an effect of "the transformation of governance" (Kettl, 2000, p.488), or as a primary type of privatization (Brudney et al., 2005, p.394). 2 National Bureau of Economic Research, www.nber.org/cycles/cyclesmain.html, accessed 05/12/2012 17 Although contracting has been widely studied, there exist valid reasons to demand more investigation. As for theoretical interest, most of the existing studies addressing the impact of an institution on contracting test the institutional effect independently of other explanatory factors. The particular interest of the present study is how an institution - council-manager form or charter status - moderates the impact of recession on contracting. In an empirical model, the moderating effect of an institution can be revealed by multiplying the variable accounting for recession by the variable capturing an institution (Feiock et al., 2003, p.619; Brambor et al., 2006). This study will multiply the variable of recession twice: one with the variable of the council-manager form and the other with the variable of the charter status. Regarding the characteristics of the data, much of the existing research is characterized by a discrete contract dependent variable and cross-sectional data. The variable is discrete, since it is divided into several production choices such as in-house production, public, for-profit, and nonprofit contracting. This is somewhat limiting, because contracting may not be a matter of qualitative choice in reality but rather a matter of an ingenious quantitative combination of in- house production and some form of contracting (Hirsch, 1995, p.233). A cross-sectional dataset also has a potential weakness, because a contract decision is typically made at least one year before its actual implementation. The data is not able to capture the dynamic change in contracting as a result of a government decision, but can only show the simultaneous association between a certain factor and contracting in the same year (Bel and Fageda, 2009, p.115). Research efforts to incorporate time dimensions have just begun (e.g., Lamothe et al., 2008; Picazo-Tadeo et al., 2012) and are still sparse. The current study addresses 18 these weaknesses by employing continuous contract spending with a time dimension as the dependent variable. 3 Specifically, this paper centers on recessionary effects on city contracting behaviors, varied by two institutional factors (the council-manager form of government and city charters). Two political factors are further included: the political party membership of a state governor and the share of votes to a particular party out of total votes cast. Primary research questions include: 1) Does a recession make a city government prefer contracting to in-house production? 2) Do institutional forces matter in making contract decisions during recession? 3) Do political factors matter in contracting? Relevant quantitative techniques are employed on a panel dataset of all available California cities over the past 17 years, 1993-2009. Contracting consists of public contracting, i.e., contracting with other government agencies, and private contracting, i.e., contracting with for-profit or nonprofit organizations (Levin and Tadelis, 2010, p.531). Private contracting could further be divided into for-profit contracting (contracting with for-profit firms) and nonprofit contracting (contracting with nonprofits) (Brown and Potoski, 2003, pp.449-450). Private production and public production are often used similarly. While private production may mean private contracting in the context of the contracting literature, public production is not the same as public contracting; public production actually includes in-house production as well as public contracting (Ferris and Graddy, 1994, p.131). In the face of fiscal burdens triggered by an economic recession, city governments must seek out a more efficient mode of service provision between in-house production and contracting (Ferris and Graddy, 1994, p.127). Production costs are typically higher in in-house production than contracting, whereas transaction costs are higher in contracting than in in-house production 3 The dependent variable = contract expenditure ÷ total expenditure × 100 19 (Ferris and Graddy, 1994, p.129). An economic/fiscal shock may cause city residents to prefer savings on production costs over those on transaction costs, because the former are typically more apparent than the latter in their eyes (Ferris and Graddy, 1994, p.129). As a result, city governments are expected to increase contracting in the face of recession. This preference is argued to be respected to a greater extent in council-manager cities than mayor-council cities, because the former are less politically charged (Feiock and Jang, 2009, p.671), and in charter cities than general-law cities, because the former have greater flexibilities (Benton, 2002, p.474). Contract-related flexibilities for California charter cities include exemption from state requirements on prevailing wages and competitive bidding. 4 The empirical analysis using the double interaction of recession with the council- manager form and with the charter status intriguingly indicates that recessionary impacts on contracting are moderated by four institutional constellations: the council-manager and charter, the council-manager and general-law, the mayor-council and charter, and the mayor-council and general-law. Specifically, recession is shown to reduce contracting in a government characterized by the mayor-council and general-law to the greatest extent. The introduction of either the council-manager form or the charter status, rather than the mayor-council form or the general-law status respectively, moderates the extent of the reduction. This inverse relationship between recession and contracting eventually turns positive, when a government is defined with the council-manager and charter. This paper begins with a literature review, then attempts to describe contract patterns over time according to the presence of recession and the other institutional characteristics. This will be followed by the estimation of several empirical models. Finally, considerable implications from the estimation will be discussed. 4 League of California Cities, General Law City v. Charter City, www.cacities.org/chartercities, accessed 06/16/2012. 20 2. LITERATURE REVIEW AND HYPOTHESES Economic recession Recession depresses the revenue base in general (For an explanation regarding the recessionary impact on service fees, see McCubbins and Moule, 2010, p.614). The depressed revenue base creates substantive fiscal burdens on governments (Levine, Rubin, and Wolohojian, 1981, p.12). They have to seek savings on the costs of producing services. Fiscal burdens can effectively restrain the potential opportunism of city officials shirking from their commitment to the cost savings (Levin and Tadelis, 2010, p.532). The other option is tax increases. However, these are rarely favored by tax payers especially during a recession (Stein, 1990, p.83). In the matter of production cost savings, contracting to third-party vendors such as for- profit firms, nonprofits, or other public agencies is often understood as an alternative with greater potential for savings than in-house production (Stein, 1990, p.84). The latter causes substantial labor costs, part of production costs, while the former saves that money to a great extent (Stein, 1990, p.84). Still, contracting is likely to incur the other type of costs, transaction costs (Williamson, 1981), which consist of the costs of measuring the quality of the service produced by a contractor and enforcing the contractual agreement made by the responsible government with the contractor (North, 1990, p.27). The ever-present chance of opportunistic behavior, created by the fact that a contractor's incentives may conflict with those of government officials, drives transaction costs higher. What then makes contracting more attractive than in-house production in a recession despite the hazards of transaction costs? Ferris and Graddy (1994, p.127) argue that most decisions about service contracting are made based on the consideration of production and transaction costs. Interestingly, production costs are easy to see and measure while transaction 21 costs are less so (Ferris and Graddy, 1994, p.129). Ferris and Graddy imply that the decision to save either production or transaction costs depends on which one average voters want (1994, p.129). Faced with fiscal stress, average voters typically prefer to slash production rather than transaction costs, because the former look more apparent to them (Ferris and Graddy, 1994, p.129). Let us apply this theoretical account to the context of the present study on recession and contracting. When a recession hits and triggers fiscal stress, it would be difficult to explain in detail to average voters the hazard of transaction costs, which are by nature ambiguous and less visible despite their undoubted existence. Production costs, on the other hand, can be explained to them with less difficulty than transaction costs, as evidenced by labor costs intuitively calculated by the total amount of salary paid to employees. The essential difference between the two is likely to lead average voters, frustrated by the recession, to prioritize reducing production rather than transaction costs. As a consequence, it is hypothesized that the presence of a recession is related to the increasing use of contracting. Empirical findings concerning a declining macro-economy, or its resulting fiscal burdens, and contracting are supportive of the predicted positive relationship in general. The relationship is firstly hinted at by the comparative case study of four U.S. local governments by Levine et al. (1981). Their study finds that the localities faced with fiscal stress tend to place more emphasis on production efficiencies (Levine et al., 1981, p.188). The subsequent large N quantitative studies are done with discrete or continuous dependent variables capturing contracting. Using a discrete dependent variable, Brown and Potoski (2003) support the notion that a government with a greater fiscal capacity is likely to be less interested in contracting than one with less fiscal capacity (p.463). Similarly, Levin and 22 Tadelis (2010) show that city governments carrying more debt tend to seek more contracting (p.532). The positive relationship between the presence of fiscal pressure and contracting is partly supported in Hefetz and Warner (2007, p.566). A few studies have attempted to construct continuous dependent variables for the purpose of overcoming the limitations of discrete dependent variables. For example, Hirsch (1995) measures contracting by "[p]roportion of residential solid waste disposal nonpayroll expenditures" (p.233). As predicted, the data supports the notion that a city with a poor bond rating is likely to do more contracting (Hirsch, 1995, p.236). A disadvantage of this contracting measure, however, is that it might not actually capture what it is supposed to do, the real amount of contracting. Whereas Hirsch (1995) utilizes hard data from a public finance database, Brudney et al. use soft responses to a survey item asking "the percentage of the agency's budget that is allocated to contracts for delivering services to the public (as reported in deciles by the agency heads)" (2005, p.399). Their study also finds a positive association of their fiscal need indicator with contracting at the state level (Brudney et al., 2005, p.408). Drawing on the above discussion, the following hypothesis envisaging a positive relationship between a recession and contracting can be built: Hypothesis 1: A recession is positively related to contracting. In the present study, recession will be made to interact with the council-manager form and charter status to investigate how these two institutions moderate between recession and contracting, as inspired by Feiock et al. (2003, p.619) and Brambor et al. (2006). The choice between additive and interactive approaches depends on which relationship is predicted - direct or moderated. If a certain explanatory factor (A) is predicted to cause a direct impact on a phenomenon to be explained (B) independently of an institution (C), the additive approach is 23 suitable (Feiock et al., 2003, p.619; Brambor et al., 2006). On the other hand, if A's impact on B is predicted to be moderated by the presence/absence of C, the interactive approach is considered desirable (Feiock et al., 2003, p.619; Brambor et al., 2006). The present study requires two interaction terms, because there are two moderating institutions: the council-manager and the charter. Council-manager form of government Two forms of government appear in California cities: council-manager and mayor- council (ICMA, 2005). The critical difference between the two forms lies in who does the tasks of the chief executive officer (CEO) (V. Ostrom et al., 1988, p.44). These tasks mean the everyday operations of city government agencies, including personnel appointment (V. Ostrom et al., 1988, p.45). In the typical council-manager form, the city manager takes the job (V. Ostrom et al., 1988, p.44), appointed by his or her city council (V. Ostrom et al., 1988, p.45). He is not a professional politician but a professional management expert believed to prioritize the success of his career in the management profession (Feiock and Jang, 2009, p.671). On the other hand, the city mayor is in charge of the job in the typical mayor-council form (V. Ostrom et al., 1988, p.42). The mayor is elected popularly by residents (V. Ostrom et al., 1988, p.42). The mayor is a professional politician aggressively pursuing re-election by mobilizing the support of his constituencies such as public labor unions (Feiock and Kim, 2001, p.31; Feiock and Jang, 2009, p.671). As such, the council-manager form is less politicized than its counterpart. Williamson (1985) originally devised the idea of high-powered or low-powered incentives and Frant (1996) applied it to the public sector. According to Frant (1996, p.367), the market shapes high-powered incentives since a participant in a successful deal in the market can make a direct profit from the deal. Quite differently, the organization tends to form low-powered 24 incentives because a participant in a successful deal cannot be directly rewarded for the deal with a profit from the deal (Frant, 1996, p.367). Instead, he could be rewarded in less direct ways such as a promotion in the organization (Frant, 1996, p.367). Although Williamson (1985) does not provide a complete list of high- or low-powered incentives, the conceptual distinction of incentives is useful itself in understanding that some incentives are more effective than others (Frant, 1996, p.367). Subsequent studies, notably by Feiock and his colleagues, refine and further the idea. According to them, high-powered incentives give greater motivations for efficiency but with greater risks of opportunism (Feiock and Kim, 2001, p.31). Examples include profits in the for- profit sector and re-election in the public sector (Feiock and Kim, 2001, p.31; Frant, 1996, p.370). A high-powered incentive, re-election, prevails in the mayor-council form due to the existence of the mayor as a politician (Feiock and Jang, 2009, p.671). The credibility of a promise advanced by a typical mayor on behalf of his residents is subject to whether it is within the interests of those influential groups favoring him for reelection (Feiock and Jang, 2009, p.671). Credibility tends to be threatened if the promise is not within the interests of those groups (Feiock and Jang, 2009, p.671). In contrast, low-powered incentives give less immediate motivations for efficiency but with less risk of opportunism (Feoick and Kim, 2001, p.31). A typical example is career opportunities (Feiock and Jang, 2009, p.671) like greater future jobs (Stein, 1990, p.87). The council-manager form shapes the low-powered incentive for the city manager, career opportunities in his profession (Feiock and Jang, 2009, p.671). A commitment made by a typical manager in the council-manager form on behalf of his residents tends to be more credible than one by a mayor in the mayor-council form (Feiock and Jang, 2009, p.671). The manager is well aware that he will be offered greater career opportunities in the public management profession if 25 he meets the current commitment successfully (Feiock and Jang, 2009, p.671). This expectation for this professional success effectively weakens the manager's motivation for opportunism shirking his commitment (Feiock and Jang, 2009, p.671). Drawing on the above theoretical account, let us illustrate how the high- or low- powered incentive works in the context of the present study focusing on recession and contracting. In face of recession, the typical mayor in the mayor-council can announce that he will save on production costs by increasing contracting. Regardless of how much he is sincerely determined to carry out the announcement, this is likely to be seriously undermined by his pursuit of re- election. The pursuit drives him to behave in such a way as to satisfy the major interest groups sponsoring him. Their interests may not be necessarily consistent with contracting more. For example, if the mayor is sponsored by a public labor union, his determination may well be doubted since public labor unions are known to be in general reluctant to contract work out (Hirsch, 1995, p.235). Struck by recession, on the other hand, the typical manager in the council- manager also commits himself to contracting like the mayor in the mayor-council form. Unlike his counterpart, the manager's commitment is deemed more credible, because contracting is likely to lead to a greater reputation and therefore greater probability of better jobs in the profession in the next round of his career. Based on the discussion, it is predicted that council- manager governments are more likely to increase contracting than mayor-council governments during recession. The empirical findings regarding the predicted positivity between the council-manager and contracting are somewhat mixed in the literature. Some studies are in favor of the prediction, for instance, Feiock and Jang (2009)'s hypothesis that council-manager governments are more likely to be associated with nonprofit contracting than mayor-council governments is overall 26 supported (pp.675-676). The prediction of Levin and Tadelis (2010) that council-manager governments are positively related to contracting is endorsed (p.531). On the other hand, a few studies do not find the positive regularity between the council- manager and contracting. Specifically, Stein (1990) is unable to find any statistical significance of the council-manager to contracting (p.121). Hefetz and Warner (2007) also fail to detect the association (p.568). Lastly, Lamothe et al. (2008) investigate the effect of the council-manager on five choice categories of production (p.34). Surprisingly, the analysis does not show any statistical impact of this government form (Lamothe et al., 2008, p.51). All of the aforementioned studies test the effect of the council-manager form independently of its interaction with another variable. The present study takes an interactive approach in order to discover how and to what extent the council-manager affects the positive relationship between recession and contracting. What is drawn from the literature despite the mixed empirical results is another positive relationship: that the council-manager increases contracting. As recession interacts with the council-manager, the positive relationship between recession and contracting is expected to intensify in the council-manager but not in the mayor- council: Hypothesis 2: The positive relationship between recession and contracting is stronger when a government is characterized not by the mayor-council form but by the council-manager. Charter governments Charter governments were created by the home-rule provisions installed in state constitutions (V. Ostrom et al., 1988, p.33). City charters were originally planned to grant a greater local self-government and autonomy to their respective city residents and governments than the so-called general-law cities, those without charters (V. Ostrom et al., 1988, p.34). In 27 California, charter governments are allowed superior authority over municipal affairs to the state government. 5 On the matter of municipal affairs, charter governments are not subject to provisions of the state general law, only to the state constitution. 6 Court rulings 7 indicate that extensive areas of city public finance including contract services 8 are classified as part of municipal affairs. It is noted that the difference between charter and general-law governments in California may not be as enormous as the rest of the United States, because the state constitution of California already requires the state government to comprehensively devolve many of its functions on cities regardless of charter status (Sokolow and Detwiler, 2001, p.61). 9 Nevertheless, California is not an exception from the general argument that charter governments typically have a greater policy capacity to respect what residents want than general- law governments do (Benton, 2002, p.474). The relative flexibilities of charter governments are exemplified in making payments and in choice of contractors. Charter governments have not been subject to the state law mandating the payment of prevailing wages to workers of public works projects, while it applies to general-law governments. 10 Because prevailing wages exceed market wages in most cases (Dunn et al., 2005, p.141), they may cost city governments a good deal of money. Indeed, Dunn et al. (2005, p.141) find that the 2001 imposition of prevailing wages on many California low-income housing projects caused the 9%-37% increase in construction costs, depending on model specifications. 5 League of California Cities, Charter Cities: A Quick Summary for the Press and Researchers, www.cacities.org/chartercities, accessed 09/05/2011 6 League of California Cities, Foundational Aspects of Charter Cities, www.cacities.org/chartercities, accessed 09/06/2011 7 League of California Cities, Foundational Aspects of Charter Cities, www.cacities.org/chartercities, accessed 09/05/2011 8 League of California Cities, General Law City v. Charter City, www.cacities.org/chartercities, accessed 09/05/2011 9 See also the League of Women Voters of California, "About Municipal Government," California State Government Guide to Government, www.guidetogov.org/ca/state/overview/municipal.html#2, accessed 06/16/2012 10 League of California Cities, General Law City v. Charter City, www.cacities.org/chartercities, accessed 06/16/2012. 28 To take another example, charter governments do not necessarily have to observe the formal competitive bidding procedure for public works contracts, 11 where a contract typically goes to a responsive and responsible bidder who promises the lowest price (Brown, 2010, Ch.12). General-law governments are legally required to adopt the procedure. 12 Establishing and operating a formal competitive bidding procedure may take a considerable amount of time and monetary costs on the part of city governments. That is, a relevant city should invite bids, provide relevant information to bidders, review and evaluate bids, and furthermore be prepared for possible administrative and legal challenges from unselected bidders (Brown, 2010, Ch.12). Despite the normative significance of city charters in terms of the ideal of democratic self-government (V. Ostrom et al., 1988, p.34), unexpectedly, the empirical research on its impact on fiscal outcomes is scant. The studies of Benton (2002; 2003) are helpful exceptions. Benton (2002) explores the impact of county charters on county expenditures with data tabulation. Assuming that county residents want more services from their county governments, Benton argues that governments with charters will show higher expenditures than those without, because the former is more capable of accommodating the increasing service needs than the latter (2002, p.474). The argument is consistent with what his table shows (Benton, 2002, p.474). In another study, Benton (2003) similarly finds a positive association of the presence of charters with county revenues (p.85). Earlier, it was discussed that residents tend to prefer more contracting to cut back on production costs during recession, thereby leading to the positive relationship between recessions and contracting. The contract-related flexibilities concerning prevailing wages and competitive 11 League of California Cities, General Law City v. Charter City, www.cacities.org/chartercities, accessed 06/16/2012. 12 League of California Cities, General Law City v. Charter City, www.cacities.org/chartercities, accessed 06/16/2012. 29 bidding for charter governments make the positive relationship stronger in these governments than in general-law governments: Hypothesis 3: The positive relationship between recession and contracting is stronger when a government has a charter, but not when it does not. Democratic governor and votes Along with the above economic and institutional factors, the substantive role of politics in making contract decisions is also recognized in the literature. This is because political actors are sufficiently capable of modifying fiscal outcomes. The effects of politics are examined with diverse variables such as organizational age (e.g., Levin and Tadelis, 2010) or public labor unions (e.g., Hirsch, 1995). Inspired by the literature, the present study employs two political ideological variables, Democratic governor representing an individual political orientation of a state governor and Democratic votes capturing a collective political attitude of citizens (Levin and Tadelis, 2010, p.519). Traditionally, Democrats are known to be more reluctant in contracting than Republicans (e.g., Clark and Ferguson, 1983, p.252). This is primarily because 1) Democrats favor private entities, in principle, to a lesser extent than Republicans (Clark and Ferguson, 1983, p.252) and 2) public labor unions, which dislike contracting (Hirsch, 1995, p.235), are considered to be one of the major sponsors of the Democratic Party. The Democratic negative attitude could become embodied in the municipal fiscal outcomes down from the state level and also up from city residents. Democratic governor. A governor is given power that may affect the production choices of cities, legally through regulations or fiscally through grants. A Democratic governor is 30 expected to fulfill the Democratic idea discouraging contracting throughout his tenure by making use of these legal or fiscal means. Those empirical studies particularly focused on the party membership of a governor and contracting are sparse and even mixed. Brudney et al. (2005) do not find a meaningful relationship with their political ideology measure (p.409). Still, Sundell and Lapuente (2011, published online) find evidence for the importance of political ideology; rightist governments in Sweden tend to do more contracting. Picazo-Tadeo et al. (2012) confirm the statistical significance of political ideology; Spanish governments belonging to the parties PSOE (center/left) or PP (center/right) tend to contract urban water services more than other party governments (Picazo-Tadeo et al., 2012, p.220). The discussion about the Democratic reluctance on contracting yields the subsequent hypothesis: Hypothesis 4-1: A Democratic governor is negatively related to contracting. Democratic votes. Whereas Democratic governor accounts for the political ideology of an individual governor, Democratic votes, the votes which are cast to a Democratic presidential candidate within a city, are supposed to capture the collective ideological orientation of a city (Levin and Tadelis, 2010, p.519; Clark and Ferguson, 1983, p.4). In a Democratic city, the leadership of its government may be dissuaded from contracting by pressure from its Democratic residents, even if the leadership actually favors it. In a Republican city, in contrast, leaders who are actually supportive of in-house production may be directed toward contracting by their Republican citizens. Levin and Tadelis (2010) use the percentage share of Republican votes from a county in the 2000 presidential election as a measure of political ideology (p.519, 520). They fail to find a statistically considerable relationship to city contracting (Levin and Tadelis, 2010, p.532). On the 31 other hand, Holian (2009) lends support to the hypothesis that cities in Republican counties tend to do more contracting (p.423). In the present study, the measure of political ideological orientation is improved from Levin and Tadelis (2010) and Holian (2009), because it is now variable over time and measured at the very city level. Consistent with the discussion, a negative relationship of Democratic votes with contracting is expected: Hypothesis 4-2: Democratic votes are negatively related to contracting. 3. DESCRIPTIVE TRENDS One direct way to look at the effect of recession, council-manager form, or charter status on city contracting is to plot an average share of contract expenditures out of total expenditures over time. Over the past 17 years from 1992 through 2008, two years are classified as recession years: 2001 and 2008. 13 A recession is assumed to take one year to make a real impact on contract outcomes. Drawing on the discussion of the literature so far, it is expected that the increase in contracting by all governments, council-manager governments, and charter governments will be observed one year after a year of recession. Contracting of mayor-council governments will be incongruent with that of council-manager governments in its relation with recession; that is, mayor-council contracting will be reduced one year after a year of recession. Contracting of general-law governments will be somehow distinguished from that of charter governments, too. In other words, general-law contracting will be shrunk one year after a year of recession. Recession and contracting, all governments FIGURE 2-1 shows a mixed relationship, both unexpected and expected. As expected, the mean percentage share of contracting in 2002 has increased from that in the recession year 13 National Bureau of Economic Research, NBER, www.nber.org/cycles/cyclesmain.html, accessed 05/12/2012. According to this NBER information, the year 2009 also has six months in recession. Due to the contract data scope ending in 2009, however, its consequence occurring in the next year cannot be shown in the following figures. 32 2001 (21.77% in 2001 => 22.15% in 2002). Yet, the 2008 recession betrays the prediction by reducing 2009 contracting (21.59% in 2008 =>21.23% in 2009). FIGURE 2-1: Recession and Contracting, All Governments Council-manager vs. mayor-council governments FIGURE 2-2 shows the trends of contracting grouped into council-manager and mayor- council governments. Despite the 2008 recession, unexpectedly, 2009 contracting by council- manager governments has declined by a small fraction (21.50% in 2008=> 21.33% in 2009). Expectedly, 2002 contracting has risen due to the 2001 recession (21.37% in 2001=>21.93% in 2002). On the other hand, mayor-council governments have consistently shown the expected reduction of contracting one year after the two occasions of recession: 21.26% in 2001=> 16.33% in 2002, 18.08% in 2008=>13.32% in 2009. This may imply their vulnerability to city politics as explained in the literature. The relatively higher fluctuations of mayor-council than council- manager contracting may originate from the fact that the number of mayor-council governments out of all the cities has already been very small and become even smaller over time compared to 33 its counterpart in the dataset; i.e., 46 mayor governments compared to 338 manager governments in 1992 and 17 mayor governments compared to 436 manager governments in 2009. FIGURE 2-2: Council-manager vs. Mayor-council Governments Charter vs. general-law governments FIGURE 2-3 shows the trend of contracting separated into charter and general-law governments. For charter governments, as expected, the 2008 recession has increased the 2009 contracting level (18.48% in 2008=> 19.30% in 2009). It is also true that the 2002 contracting level has increased from the level of contracting in 2001, the other recession year (15.96% in 2001=>16.54% in 2002). Yet the growth (0.58%) from 2001 to 2002 has slowed, compared to that (1.04%) from 2000 (14.92%) to 2001 (15.96%). On the other hand, general-law contracting (21.88%) in 2009 has declined as expected from that (22.61%) in 2008, whereas general-law contracting (23.26%) in 2002 has risen unexpectedly from that (23.10%) in 2001. 34 FIGURE 2-3: Charter vs. General-law Governments While the graphs are very helpful for intuitive understanding, this illustration overlooks all the cross-sectional differences of contracting across cities totaling 407 through 472 depending on the year in question. The following regression technique is employed to overcome this limitation. 4. DATA AND METHODS To rigorously test the hypotheses, an extensive panel dataset is constructed by combining data from diverse sources (TABLE 2-1). It contains economic, institutional, political, demographic, and contract information from virtually all California cities over 17 years. Specifically, the dataset covers 1993-2009 for the dependent variable, Contracting, while it spans 1992-2008 for all the independent and control variables. 35 TABLE 2-1: Variables, Measures, and Sources of Data Variables Measures Sources Contracting(%) (Contract expenditure ÷ total expenditure)×100 <California City Financial Transactions>, RAND California 14 ; originating source-Cities Financial Transactions Report Recession Dichotomous indicator, t-1. National Bureau of Economic Research 15 Council-manager Dichotomous indicator, t*; if a city is the mayor form, 0 is assigned. If the manager form, 1. ICMA, the Municipal Year Book a Charter Dichotomous indicator, t-1; if a city is subject to the state general law on municipal affairs, 0. If chartered, 1. California State Controller, Annual Cities Report Democratic Governor Dichotomous indicator, t-1; if the state governor is Republican, 0. If Democratic, 1. National Governors Association 16 Democratic votes (%) Votes to a Democratic presidential candidate out of total votes cast at city level, t-1. California Secretary of State, 1992, 1996, 2000, 2004, and 2008, Supplement to the Statement of Vote City size Logged city population, t-1. <Population and Demographic statistics>, RAND California 17 ; originating source-the U.S. Census Bureau Service demand Logged county population with age 65 or older, t-1. <Population and Demographic statistics>, RAND California 18 ; originating source-the U.S. Census Bureau Wealth Logged county personal income per capita, t-1. <Business and Economic Statistics>, RAND California 19 ; originating source- the Bureau of Economic Analysis a The Municipal Year Book in the year of time t actually shows a t-1 or even earlier information. Contracting The dependent variable measures the percentage share of contract expenditures out of total expenditures of a city government in a given year over the 17 years from 1993 to 2009. The number of cities represented in the dataset is as few as 407 in 1996 and as many as 472 in 2008. This variable is constructed by dividing contract expenditures (= public + private contract 14 ca.rand.org/stats/govtfin/cityfinance.html. The relevant items to be used are selected as follows and then merged. 1) Public contracting: For 1993-2000, "Expenditures by Major Object Classification (i.e., Major Category)"=> "Services from other governmental agencies" => "Total expenditures" and for 2001-2009, "Schedule of Total Expenditures by Major Object Classification"=> "Operating Expenditures-Contract Services, Total Expenditures, Other Governmental Agencies". 2) Private contracting: For 1993-2000, "Expenditures by Major Object Classification (i.e., Major Category)"=> "Private contract services"=> "Total expenditures" and for 2001-2009, "Schedule of Total Expenditures by Major Object Classification"=> "Operating Expenditures-Contract Services, Total Expenditures, Private". 3) Total expenditure: For 1993-2000, "Expenditures by Major Object Classification (i.e., Major Category)"=> "Total operating, capital, and debt service"=> "Total Expenditures" and for 2001-2009, "Schedule of Total Expenditures by Major Object Classification"=> "Total Expenditures, Total" 15 National Bureau of Economic Research, www.nber.org/cycles/cyclesmain.html, accessed 05/12/2012 16 National Governors Association, www.nga.org/cms/governors, accessed 05/12/2012 17 ca.rand.org/stats/popdemo/popdemo.html 18 ca.rand.org/stats/popdemo/popdemo.html 19 ca.rand.org/stats/economics/economics.html 36 expenditures) by total expenditures. Public contract expenditures are defined and illustrated as follows: "all expenditures for general or special city services performed by an outside governmental agency. For example, the cost of street work performed by a nearby city would be reported on this line" (California State Controller, 2005, p. 110). Likewise, private contract expenditures are defined and illustrated as follows: "all expenditures for general or special city services performed by private agencies. For example, the cost of garbage collection performed by a private company would be reported on this line" (California State Controller, 2005, p. 110). The data originate from the Cities Financial Transactions Report, submitted by each California city to the California State Controller. A comprehensive database, called <California City Financial Transactions>, is compiled by RAND California from the reports. While the data are reported before audit (California State Controller, 2005, p.5), they are sufficiently reliable because they are legally required of each and every city according to Government Code section 53891 (California State Controller, 2005, p.5). The <California City Financial Transactions> has two sets of data from two different periods of time: one for 1993-2000 and the other for 2001 - 2009. Relevant data for contracting covering the complete 1993-2009 are merged with caution. Recession The macro-economic condition is measured by a dichotomous variable indicating the presence of nationwide economic recession in a given year. According to the National Bureau of Economic Research (NBER), the monthly periods of recession between 1992 and 2008 are April 2001-November 2001 and January 2008-December 2008 20 . Thus, both 2001 and 2008 are considered recession years and therefore allocated 1 each, while the remaining 15 years are coded 0. 20 National Bureau of Economic Research, www.nber.org/cycles/cyclesmain.html, accessed 05/12/2012 37 Council-manager This is a dichotomous variable indicating two forms of government: council-manager and mayor-council. This variable in the dataset is dynamic over time as well as across cities. If a city adopts the council-manager form, it is given 1. If a city adopts the mayor-council form, it is given 0. The data is derived from the annual ICMA Municipal Year Book. Unlike all the other independent and control variables in the present study, council- manager does not need to employ a time-lag. The reason is that the information of council- manager of a city in a year t is already representing a form of government as of t-1 or more previous years. The ICMA Municipal Year Book in year t contains the relevant information collected through its survey conducted in the year t-1 (see ICMA, 2005, p.184). When a city does not respond to the survey, the most recent available information collected at a year earlier than t- 1 is used (see ICMA, 2005, p.184). The council-manager data for 1997, which are missing, are created by considering the continuation between 1996 and 1998, because it is very unlikely for a city to change its government structure quickly twice within 3 years. In other words, if the government form of a city in both 1996 and 1998 is commonly the council-manager, the council-manager, i.e., 1 is coded in 1997. However, if the forms in 1996 and 1998 are not the same as each other, the form remains undetermined in 1997. Council-manager is multiplied with recession, creating an interaction term, recession×council-manager, which is assigned 0 or 1. If a council-manager city is hit by a recession, recession×council-manager is assigned 1. All other possible cases are given 0. 38 Charter This variable is a dichotomous indicator, distinguishing a city with a charter from one without. The latter is called a general-law government. If a city chooses to have a charter, it is allocated 1. A general-law government is given 0. Like council-manager, charter is also variable over time and across cities. The relevant information is obtained from the Annual Cities Report, published by the California State Controller. The information from 1993 is missing and inferred based on whether a city is characterized by the same status in both 1992 and 1994. If its status in 1992 is not the same as in 1994, the status is not determined. For example, the information on Sand City indicates general- law in 1992 but charter in 1994. In this case, the status in 1993 cannot be determined. Charter is multiplied with recession, producing a new dichotomous interaction term, charter×recession. If a charter city is under recession, charter×recession is allocated 1. All the other conceivable situations are coded 0. Democratic governor Between 1992 and 2008, California had three governors; Pete Wilson (January 7, 1991 - January 4, 1999, Republican), Joseph Graham "Gray" Davis, Jr. (January 4, 1999 - November 17, 2003, Democratic), and Arnold Schwarzenegger (November 17, 2003 - January 3, 2011, Republican) 21 . Gray Davis was the only governor from the Democratic Party during this period. His tenure until he was recalled, i.e., 1999-2003, is assigned 1, while the rest of the period is allocated 0. Democratic votes The presidential election results at city level conducted in 1992, 1996, 2000, 2004, and 2008 are utilized. This continuous indicator is measured as the percentage share of the votes 21 National Governors Association, www.nga.org/cms/governors, accessed 05/12/2012 39 which a Democratic presidential candidate earns out of total votes cast in a city in a given year. It changes quadrennially. The data are gathered from the Supplement to the Statement of Vote, published by the California Secretary of State. The results of the 1992 election cover 1992, 1993, 1994, and 1995. Those of the 1996 election are employed throughout 1996 - 1999. Likewise, those of the 2000 and 2004 elections span 2000 - 2003 and 2004 - 2007, respectively. The 2008 election results cover 2008. Controls A few demographic indicators such as city population (city size), county population with age 65 or older (service demand), and county personal income per capita (wealth) are included as controls. The data on the last two are available only at county level, not city level. The data sources of these controls are specified at TABLE 2-1. TABLE 2-2 shows the descriptive statistics of the introduced variables. TABLE 2-2: Summary Statistics of Base Variables Variables (unit) Obs. Mean Standard Dev. Min. Max. Contracting (%) 7,611 22.23 17.39 0 90.21 Recession 8,053 .11 .32 0 1 Council-manager 7,552 .95 .20 0 1 Charter 8,035 .20 .40 0 1 Democratic governor 8,059 .29 .45 0 1 Democratic votes (%) 8,010 50.10 14.62 10.34 94.85 City population (person) 7,982 58,270.36 192,130.6 89 3,801,574 County population with age 65 or older (person) 8,059 257,836.1 335,249.5 579 1,024,025 County personal income per capita ($, 2005 constant) 8,059 34,562.71 11,064.7 17,452.6 86,383.49 40 Models For accurate estimation, the following treatments are conducted. First, county personal income per capita is converted to 2005 constant dollars using a GDP deflator 22 for controlling inflation. Second, all the independent and control variables take the t-1 time-lag. council- manager is not required to do so, as explained above. Third, city population, county population with age 65 or older, and county personal income per capita are transformed into natural logarithms. Their coefficients will be read as percentages. Fourth, all six models report cluster- robust standard errors adjusted for cities due to their respective presence of heteroskedasticity and autocorrelation. All the six models are estimated with the fixed-effects technique, including year dummies. This estimation is instrumental in controlling the possible effects of un-captured city-specific characteristics, which are regarded as stable over time in a given city, as well as those of un- captured year-specific qualities, which are considered more or less the same across cities in a given year (Hsiao, 2003, p.30; also MacDonald, 2008, p.462). The technique for the present study is justified on theoretical and technical grounds. According to Hsiao (2003), on theoretical grounds, it is appropriate for an analysis of a population, not a random sample (p.43). If a sample is randomly gathered from its target population, random-effects estimation would be more appropriate (Hsiao, 2003, p.43). The present study investigates not a random sample but a population itself consisting of all California cities. Technically, fixed-effects estimation is favored by a relevant F test over pooled OLS constantly throughout all the six models. The Hausman test for the comparison between fixed- and random-effects estimation prefers the former to the latter consistently across all the models. 22 Source: Bureau of Economic Analysis (www.bea.gov) => "Interactive Data" => "National Data" => "GDP & Personal Income" => "SECTION 1 - DOMESTIC PRODUCT AND INCOME" => "Table 1.1.9. Implicit Price Deflators for Gross Domestic Product," downloaded 07/05/2011 41 When an interaction term (e.g., recession×council-manager in Models 4 and 6) is included, its own constituents (recession and council-manager) are added side by side according to the recommendation of Brambor et al. (2006, p.66). The same recommendation also applies to the other interaction term, charter×recession in Models 5 and 6 . A consequence is the symptom of severe collinearity between an interaction term (recession×council-manager) and its own constituent (recession) in Models 4 and 6. Yet, both are chosen to remain in the models per Brambor et al. (2006, p.70). 5. RESULTS TABLE 2-3 shows the estimated coefficients of the models. Models 1-5 are baseline specifications to check the robustness of the results of Model 6, the complete model. In particular, Model 6 is the specification of double interactions where recession is multiplied with council- manager and charter. Model 1 consists only of controls. With these controls included, two political variables are added to Model 2, recession to Model 3, the interaction between recession and council-manager to Model 4, and the interaction between recession and charter to Model 5. In Model 6, recession is found to be significant together with its interaction terms, recession×council-manager and charter×recession. The significance of these three variables is quite robust, because each of these three also turns out to be significant in the baseline interactive Models 4 and 5: recession and recession×council-manager in Model 4, and charter×recession in Model 5. The following interpretation of Model 6 is inspired by Brambor et al. (2006, p.77, footnote 15). In this specification, the effect of recession varies depending on the respective presence (1) or absence (0) of either council-manager or charter. This is understood better with this simplified version of Model 6: Contracting = β 0 + β 1 Recession + β 2 Council-manager + 42 β 3 Charter + β 4 Recession×Council-manager + β 5 Charter×recession + ε. The effect of recession is calculated as follows: = β 1 + β 4 Council-manager + β 5 Charter. The effect of recession is differentiated by the combination of council-manager (0, 1) and charter (0, 1): first, when council-manager = 0 and charter = 0, the effect of recession = β 1, second, when council-manager = 0 and charter = 1, the effect = β 1 + β 5 , next, when council- manager = 1 and charter = 0, the effect = β 1 + β 4, and last, when council-manager = 1 and charter = 1, the effect = β 1 + β 4 + β 5 . In Model 6, β 1 means the coefficient of recession (-6.28), β 4 refers to that of recession×council-manager (5.14), and β 5 is that of charter×recession (2.07). The effect of recession moderated by the four institutional arrangements can be further elaborated as follows (FIGURE 2-4): The sole coefficient of recession (-6.28) represents the effect of recession on the contracting of those governments in the simultaneous absence of council-manager (0) and charter (0), meaning mayor-council and also general-law governments. Surprisingly, a recession turns out to bring down the level of contract expenditure as a portion of total expenditure of these governments by 6.28%. Second, the sum of the coefficient of recession (-6.28) and that of charter×recession (2.07) represents the case of the cities in the absence of council-manager (0) and the presence of charter (1), i.e. mayor-council and chartered governments. In these governments, a recession turns out to decrease the contracting level by 4.21% (=(-6.28)+2.07). Third, the sum of the coefficient of recession (-6.28) and that of recession×council- manager (5.14) describes the effect of recession on contracting of those governments in the presence of council-manager (1) and the absence of charter (0), referring to council-manager 43 and general-law governments. In these cities, a recession leads to the reduction of contracting by 1.14% (=(-6.28)+5.14). Last, the total sum of the coefficients of recession (-6.28), recession×council-manager (5.14), and charter×recession (2.07) captures the effect of recession on contracting of those governments in the simultaneous presence of council-manager (1) and charter (1), i.e. simultaneously council-manager and chartered governments. In these cities, a recession turns out to increase the level of contracting by 0.93% (=(-6.28)+5.14+2.07). All in all, Hypothesis 1's prediction of a positive relationship between recession and contracting is not confirmed. Out of the four different institutional arrangements, there is only one arrangement where the relationship turns out to be positive, council-manager (1) and charter (1), while recession shows a negative relationship with contracting in the other three arrangements. On the other hand, Hypotheses 2 and 3 considering each of council-manager and charter as a modifier between recession and contracting are supported. To be sure, neither is a modifier in terms of making a positive relationship between recession and contracting stronger, because the relationship proves to be negative. Yet each of the two does modify the negative relationship by moderating it to an extent, 5.14% (council-manager) or 2.07% (charter). Hypotheses 4-1 and 4-2 fail to be supported, because neither Democratic votes nor Democratic governor turns out to be statistically significant. Among controls, county population with age 65 or older shows a positive association with contracting significant at a 10% level. 44 TABLE 2-3: Fixed Effects (Within) Regression Coefficients Variables Dependent variable: Contracting (%) (=(contract expenditure÷total expenditure)×100) Model1 Model2 Model3 Model4 Model5 Model6 Recession t-1 -2.47 (2.23) -5.48** (2.40) -2.42 (2.51) -6.28* (3.22) Council-manager t -0.07 (2.05) -0.46 (2.14) Charter t-1 1.42 (2.02) 0.20 (2.01) Recession t-1 ×council-manager t 4.88** (1.91) 5.14*** (1.84) Charter t-1 ×recession t-1 1.79** (0.73) 2.07*** (0.72) Democratic governor t-1 2.25 (1.67) 1.33 (1.14) Democratic votes t-1 (%) -0.00 (0.07) -0.03 (0.05) LN city population t-1 0.11 (2.84) 0.04 (2.94) 0.11 (2.84) 2.14 (2.53) 0.40 (2.80) 1.89 (2.57) LN county population with age 65 or older t-1 15.87** (7.06) 16.11** (7.10) 15.87** (7.06) 9.11* (5.27) 15.89** (7.07) 9.08* (5.28) LN county personal income per capita t-1 -5.15 (5.37) -5.09 (5.32) -5.15 (5.37) -3.84 (5.02) -5.06 (5.39) -3.72 (5.00) Constant -107.93 (96.83) -112.85 (100.15) -107.93 (96.83) -65.01 (76.57) -113.01 (96.19) -62.62 (77.25) R-sq (within), % 0.74 0.77 0.74 0.73 0.88 0.86 R-sq (between), % 7.69 7.61 7.69 5.79 7.13 6.15 R-sq (overall), % 5.12 5.12 5.12 4.27 4.74 4.50 P value of F test for model fit 0.28 0.30 0.28 0.055 0.11 0.016 Number of observations 7,536 7,510 7,536 7,120 7,520 7,089 Number of cities 470 470 470 452 470 452 Note: Cluster-robust standard errors are noted in parentheses. Year dummies are included. Probability levels are denoted as follows: *** p<0.01, ** p<0.05, * p<0.1. FIGURE 2-4: The Effects of Recession on Contracting Moderated by Institutional Arrangements 6. DISCUSSION A recession resulted in decreasing contracting. This negative effect of a recession was the deepest in the mayor-council and general-law and became moderated by whether a government 45 employs a charter, council-manager form, or both. That is, the initial negative effect was more than 6%. The adoption of a charter mitigated the negative effect by about 2%. The employment of the council-manager alleviated it to a great extent by more than 5%. Surprisingly, the presence of both a charter and the council-manager was powerful enough to upset the 6% negative effect of a recession, turning it positive; in other words, a recession was shown to increase contracting in the chartered council-manager governments by around 1%. The literature (e.g., Ferris and Graddy, 1994, p.129) suggested more government contracting during a recession because local residents would typically want it. Our mostly negative empirical relationship hints that the literature implicitly presumes that governments are already equipped with institutions suitable to honor local preferences. The story of Model 6 provides an excellent case for the argument, inspired by Feiock et al. (2003, p.619), that the effect of an environmental shock such as a recession on government outcomes would be moderated by institutional characteristics of government. Government responses to a recession are not created in a vacuum. They are produced within the government context, characterized by distinctive types of institutions. To take a closer look, the council-manager form appears to facilitate making contract decisions when faced with a recession. The council-manager form, characterized by a position for professional managers rather than for politicians (V. Ostrom et al., 1988, p.42, 44), shapes a low-powered incentive, career development (Feiock and Jang, 2009, p.671). A typical manager has a good reason to honor his commitment to cost savings of service production through contracting in response to the economic shock. The reason for this is that if the manager achieves a good level of cost savings, he increases the likelihood that greater jobs will be offered to him. If he fails, the likelihood will diminish and his professional reputation as an economy rescuer 46 will be undermined. This reduces the risk of shirking the commitment. Conversely, the commitment of a mayor under the mayor-council form may not be as trustworthy as that of his counterpart under the other form, since he may easily be tempted to satisfy his powerful interest groups for re-election, a high-powered incentive (Feiock and Jang, 2009, p.671). Yet, a qualification has to be made about the empirical evidence provided for this argument, because the data is highly skewed toward the council-manager (Lamothe et al., 2008, p.51). The salience of city politics in the mayor-council form is succinctly illustrated by two instances of the City of Los Angeles. In the 2005 mayoral election, then candidate Antonio Villaraigosa was endorsed by two large unions to which approximately 16,000 city employees of Los Angeles International Airport, the Los Angeles Police Department, and the Department of Water and Power belong. 23 In a hypothetical situation where a mayor were elected with the support of such powerful groups, it would be difficult to think that a serious effort would occur to contract out airport, police, water, and electricity services. The other instance is the former mayor Richard Riordan, who began his tenure in 1993. At first, he pushed hard on an ambitious agenda to privatize substantial parts of city services via contracting, but he soon became discouraged due to the consistent opposition from unions and their supporters in the city council. 24 City charters seem to boost government contract decisions under a recession. The relevant result provides evidence that the local preference toward enhancing production efficiency, created by an economic crisis, is better satisfied by the former than by the latter. Charter governments have more flexibility in fulfilling local preferences than do general-law governments (Benton, 2002, p.474), exemplified by the exemption from the payment of 23 Patrick McGreevy, "L.A., Union at Odds Over Wage Pact; Engineers and Architects Assn. threatens to protest at LAX on Sunday. It seeks a contract similar to one reached with DWP," Los Angeles Times, November 22, 2005: p. B.1. 24 Jean Merl, "The Mayor's Midterm Exam How Far Has Richard Riordan Moved the City Since he Became Mayor? An Evaluation of L.A. Inc. After Two Years," Los Angeles Times, June 11, 1995: p.8. 47 prevailing wages and from the establishment of formal competitive bidding practice regarding contracting. 25 A specific structural advantage enjoyed by a charter city is illustrated clearly by the City of Vista, chartered by residents in June 2007, 26 just half a year before the onset of a recession. One of the rationales is known to be having the capacity to determine whether to pay prevailing wages for a locally financed public-works contract. 27 Eventually, the newly chartered city chose not to remunerate with prevailing wages, thereby saving significantly on production costs. 28 All in all, it becomes clear that the mayor-council form of government without a charter is the very institutional arrangement that is the least inclined towards contracting during recession, as supported by the largest negative effect of a recession on contracting of this type of government. By the same token, it is evident that the council-manager form of government with a charter is the very institutional arrangement that is the most inclined towards contracting during recession, as evidenced by the positive effect of a recession on contracting in this type of government. Two political variables, Democratic governor and Democratic votes, were shown to be non-significant in the present study. In particular, the impotence of Democratic votes is quite disappointing, since the measure is more refined than the traditional county-level measure (e.g., Levin and Tadelis, 2010; Holian, 2009). The votes to a particular party measure a citizens' collective political attitude in a city (Levin and Tadelis, 2010, p.519). It needs to be admitted that there is no guarantee that a citizens' collective attitude is automatically transferred into 25 League of California Cities, General Law City v. Charter City, www.cacities.org/chartercities, accessed 06/16/2012. 26 Nathan Scharn, "Vista Wins Prevailing-Wage Ruling," The San Diego Union-Tribune, July 3, 2012, www.utsandiego.com/news/2012/jul/03/tp-vista-wins-prevailing-wage-ruling/, accessed 06/11/2013. 27 Pamela MacLean, "Big California Cities Exempt from Prevailing-Pay Law," Bloomberg, July 2, 2012, www.bloomberg.com/news/2012-07-02/big-california-cities-exempt-from-prevailaing-pay-law.html, accessed 06/11/2013. 28 Nathan Scharn, "Vista Wins Prevailing-Wage Ruling," The San Diego Union-Tribune, July 3, 2012, www.utsandiego.com/news/2012/jul/03/tp-vista-wins-prevailing-wage-ruling/, accessed 06/11/2013. 48 government action. Such attitude must be strongly supported by organized groups that are able and willing to keep pushing it on government decision-makers. Accordingly, while the Democratic attitude, which favors in-house production to contracting, may be prevalent in a particular city, its full realization in government action might not be likely without the hard work of its sponsoring groups. Specifically, in the U.S. context, interest groups are well-known to be highly influential, as Terry Moe succinctly said, "[s]tructural politics is interest group politics," based on his observation of bureaucratic structure (1989, p.269). 7. CONCLUSION The moderating function of the council-manager form between economic downturns and contracting has been substantiated by the empirical findings. The literature (e.g., Feiock and Jang, 2009) on the different incentives between the council-manager and mayor-council forms has been found to be relevant in explaining the moderating function for the present study. Let us apply the different incentives indicated by Feiock and Jang (2009, p.671) to a situation where a government is hit by a recession. In the face of a recession, that is, the typical manager's will toward cost savings via contracting is more likely to be realized, since the realized efficiency would be rewarded with greater likelihood of thriving public lives ahead. In a parallel situation, however, the typical mayor's equivalent will could be undermined and even distorted by the politics in which he is involved; saving money through contracting does not necessarily help him please those interest groups which are essential for his next re-election. Similarly, the moderating effect of the chartered status between recession and contracting has been validated by the estimation results. The literature argues that the chartered status tends to create wide latitude for policies that accommodate residents' collective preferences (Benton, 2002, p.474). The present study has found specific policy flexibilities for California charter 49 governments such as the exemptions from the state requirements of prevailing wages and competitive bidding. 29 The residents' preference for contracting during a recession (Ferris and Graddy, 1994, p.129) is typically realized by charter governments utilizing these flexibilities, as opposed to general-law governments. While this study concentrates only on California cities, it could also provide useful implications for other cities outside California, as economic fluctuations are a common concern not only of California cities but also of non-California cities. How and to what extent they experience the shock may vary depending on which institutional properties they hold. A limitation of the study is the low within-R-squared value less than 1% despite the satisfactory between-R-squared of more than 6% in the main specification, Model 6 (TABLE 2- 3). In other words, temporal variations of contracting within each city were not greatly explained by the model. Part of the reason is the difficulty finding good controls at city level, varying over time. The data for the two out of the three controls, income and population with age 65 or older, was available only at county level, not at city level. Nevertheless, the present study is a positive addition to the literature by taking a time dimension into account. 29 League of California Cities, General Law City v. Charter City, www.cacities.org/chartercities, accessed 06/16/2012. 50 APPENDIX 2-1: Correlations 1.Contrac ting t 2.Recess ion t-1 3.Coun cil- manag er t 4.Chart er t-1 5.Recess ion t- 1×counci l- manager t 6.Charte r t- 1 ×recess ion t-1 7.Democ ratic governor t -1 8.Democ ratic votes t-1 9.LN city populati on t-1 10.LN county populat ion 65+ t-1 11.LNco unty personal income per capita t-1 1 1 2 -0.01 1 3 0.03* 0.01* 1 4 -0.15* 0.01 -0.04* 1 5 -0.00 0.98* 0.07* 0.00 1 6 -0.03* 0.44* -0.00 0.32* 0.43* 1 7 0.00 0.16* 0.04* -0.01 0.15* 0.04* 1 8 -0.11* 0.13* 0.00 0.12* 0.13* 0.09* 0.06* 1 9 -0.11* 0.01 0.02* 0.33* 0.01 0.10* 0.01 0.16* 1 1 0 0.22* 0.01 0.08* 0.13* 0.01 0.04* 0.00 0.15* 0.45* 1 1 1 0.02* 0.13* 0.06* 0.06* 0.12* 0.07* 0.10* 0.36* 0.18* 0.25* 1 Note: * significant at 10% level. 51 CHAPTER 3. The Effects of Fiscal Conditions and Intergovernmental Dependence on Non- collective Services: Evidence from California Cities, 1992-2010 1. INTRODUCTION Cities have been suffering enormously from the fiscal stress caused by the sluggish economy since 2008. 30 The prevalence of fiscal stress urgently necessitates a deeper understanding of local fiscal behaviors. One way of doing it is to group local services into two categories as some services are thought to be mostly preserved while others are disproportionately cut in the face of an environmental change. Nevertheless, the traditional research on government expenditure tends to focus exclusively on its total level (e.g., Stein, 1990; MacDonald, 2008; Wang and Hou, 2012) or some areas of spending chosen in a way not informed by theory (e.g., Saha, 2011; MacDonald, 2008; Karuppusamy and Carr, 2012). The present study aims to advance the existing research by grouping expenditures on local services into collective and non-collective categories by the property of exclusion. Despite the existence of a great deal of literature on collective (public, common-pool resources) and non- collective (toll, private) services, it has largely developed independently of the literature on public financial management. This study strives to connect the two by uncovering the financial implications of the collective and non-collective distinction. The primary research questions include 1) whether an external fiscal change modifies the provision of non-collective services independently or interactively with government form, i.e., council-manager vs. mayor-council; 2) whether a varying internal fiscal condition modifies the provision of non-collective services; and 3) whether fiscal dependence on an upper-level government causes the shift in the provision of non-collective services. In order to tackle these 30 For the nationwide business cycle, see the National Bureau of Economic Research, www.nber.org/cycles/cyclesmain.html 52 important research questions, the present study employs a large panel dataset containing fiscal information on approximately 470 California cities over 19 years from 1992 to 2010. Drawing on the literatures on collective services (e.g., Samuelson, 1954; Buchanan, 1965; E. Ostrom, 2005) and fiscal federalism (e.g., Weingast, 2009), it is reasoned that the provision of collective services tends to constitute the core responsibility of government as opposed to that of non-collective ones due to the lack of interest in the former on the part of private entities. Accordingly, non-collective services would become the immediate victim of spending cuts caused by the exacerbation of fiscal condition, suggesting a positive association between fiscal condition (external condition: state surplus/deficit, internal condition: city surplus/deficit, unreserved fund balance) and the provision of non-collective services. The other aspect of non-collective services is their capacity to generate revenues for themselves, typically in the form of user fees and charges (Cornes and Sandler, 1996, p.34). Spending for these services is at least partly offset by the expected revenue. High fiscal dependence on an upper government would discourage local governments from exploring further revenue possibilities by making them feel content with the status quo (Weingast, 2009, p.287; Kornai et al., 2003, p.1101). Consequently, an inverse relationship between fiscal dependence and the provision of non-collective services is hypothesized. An institutional design of intergovernmental transfer also matters. That is, general intergovernmental revenue would create more impact on local governments than functional intergovernmental revenue. This is because the former, to a greater extent than the latter, softens their budget constraints (Kornai, 1998). The empirical findings are stimulating. As expected, a modest but positive association of an external fiscal change appears with the provision of non-collective services. Yet an internal fiscal change indicates an inverse association. This hints at a possibility of two different 53 dynamics of fiscal change; that is, governments may respond to the aggravation of the overall external fiscal condition by getting rid of non-core service responsibilities but respond to the aggravation of the internal fiscal condition by seeking such responsibilities in the hope of collecting fees and charges. The results find solid evidence for the inverse relationship of fiscal dependence on the provision of non-collective services. Furthermore, general intergovernmental revenue is consistently found to produce larger negative impact on the very provision of non- collective services than functional intergovernmental revenue does. This paper begins with a discussion on the nature of city services and attempts to classify them into collective and non-collective services. Relevant literatures will be thoroughly reviewed in order to formulate valid hypotheses. Then fixed-effects estimation will be conducted, followed by a discussion of the implications of the estimated results. 2. NATURE OF CITY SERVICES Government services are not uniform in nature. This is demonstrated by the celebrated typology of services (Samuelson, 1954; E. Ostrom, 2005, p.24; V. Ostrom and E. Ostrom, 1991[1977]; V. Ostrom et al., 1961). They are classified into four categories - public services, common-pool resources, toll services, and private services - based on their varying degrees of excludability and subtractability (E. Ostrom, 2005, p.24). Excludability refers to the extent to which a potential user can be excluded from using a service (E. Ostrom, 2005, p.23). Subtractability means the extent to which a user's consumption of a service can actually reduce the part of the service which could otherwise be used by another interested user (E. Ostrom, 2005, p.23). While the terms collective services and public services are sometimes used interchangeably (e.g., Stein, 1990, p.75), E. Ostrom in a paper (1998, p.3) includes not only 54 public services but also common-pool resources as collective services. According to this usage, both private and toll services can be grouped together as non-collective services. The distinction between the two service types is made as follows; first, non-collective services are usually characterized by higher excludability than collective services (Bushouse, 2011, p.107). High excludability means that a service is measurable and quantifiable without substantive difficulty (V. Ostrom et al., 1961, p.833). Second, collective services typically constitute the essential part of government functioning and existence, because it is difficult to find a voluntary and well-performing provider of these services within the private sector, as advanced by the scholarship on collective services (e.g., Samuelson, 1954). This issue is related to those incentives generated when providing collective and non-collective services (Cornes and Sandler, 1996, p.10). That is, the property of easy exclusion, measurement, and quantification (V. Ostrom et al., 1961, p.833), typical of non- collective services, makes it possible for interested users to restrain themselves from a free-riding incentive (E. Ostrom, 2005, p.24). This property enables a private entity to provide a non- collective service at the optimal level (V. Ostrom and E. Ostrom, 1991[1977], p.167). Thus, there is no imperative for government to presume the provision responsibility, at least in this theoretical regard. The problem arises when a private entity provides a collective service. Due to the difficulty of exclusion, users of a collective service tend to have a free-riding incentive to appropriate the service without paying a due amount of money, and therefore to shirk revealing their actual preferences on the service (Tiebout, 1956, p.417). The provider has little idea of how much users truly want and as a consequence fails to provide it at the optimal level (Stiglitz, 1988, p.124). The consequence may be even worse in the case of common-pool resources which can be 55 depleted unlike public services, frequently called the "tragedy of the commons" (Hardin, 1968; E. Ostrom, 1998, p.3; E. Ostrom, 2009, p.420). These negative consequences make government provision of collective services legitimate. Collective services are considered essential for government in this sense. Third, the consumption of a non-collective service is likely to be paid for in the form of service fees and charges as an exclusion device (Cornes and Sandler, 1996, p.34). An effective exclusion device is instrumental in disclosing actual preferences and discouraging free-riding incentives (Cornes and Sandler, 1996, p.34). A consumer of a non-collective service is required to pay the price only for the size of the consumed. If he does not consume, he is not required to pay. The revenue from fees and charges is of strategic value to local governments. While this revenue is supposed to be expended for the original service from which this is earned 31 , it is also possible to spend this for another service through inter-fund transfer (Marlowe, 2006, p.367). In this way, this revenue can be utilized to finance collective as well as non-collective services. However, the present paper also assumes that inter-fund transfer is not without cost, administrative or sometimes political. A recent incident 32 succinctly illustrates how the cost can be incurred; the general fund of the City of Los Angeles was promised a surplus of $73.5 million from its Department of Water and Power. However, when its efforts to raise the electricity rates as much as it wanted were frustrated by the city council, the department threatened to break its promise. On the other hand, a typical collective service tends to be funded by taxes (Hyman, 2002, p.142), which are levied generally on all residents. Due to difficult exclusion, users of a 31 the Institute for Local Self Government, 2004, Municipal Finance Quick Reference, available at http://www.californiacityfinance.com/, as of 12/16/2012. 32 David Zahniser and Phil Willon. "Council OKs 4.5% power rate increase...," Los Angeles Times, April 15, 2010: AA.1; March 8 Endorsements. "Yes on Measures I and J," Los Angeles Times, February 18, 2011. www.latimes.com 56 collective service typically have to bear universal obligations to share its production costs regardless of the individual size of consumption. For instance, it would be difficult to levy a certain amount of fees to individual residents in proportion to their differential amounts of consumption of police, a collective service (Stein, 1990, p.74). Levying fees would occur more commonly to solid waste disposal, a non-collective service (Stein, 1990, p.74), since its consumption is easily measured by weight of the waste. The importance of exclusion is greatly underscored by the so-called club good literature (e.g., Buchanan, 1965; Cornes and Sandler, 1996). Cornes and Sandler declare, "[t]he essential difference between club goods and pure public goods depends on the existence of an exclusion mechanism ......" (1996, p.34). Club goods are defined as "excludable but partially nonrival" (Cornes and Sandler, 1996, p.9), a description which spans much of common-pool resources and toll services in the context of the current paper. Club goods can be satisfactorily provided by "clubs" (Buchanan, 1965, p.1), i.e. voluntarily established entities, to their own members with effective exclusion mechanisms (Buchanan, 1965, p.14). Who must become the primary provider of collective services? This has long been a big question in political science, economics, and public management and policy since Samuelson (1954) more than 50 years ago. Samuelson declares that the traditional market system may not work properly for collective services (1954, p.388). Tiebout proposes people's moving-out or -in on a jurisdictional government as a means for expressing their genuine preferences for a service and price (tax) package prepared by the government (1956, p.422). As long as each jurisdiction is formed to internalize probable externalities, Tiebout (1956) seems quite optimistic about government provision of collective services. 57 The club good literature implies that any governmental or private entity can assume collective service provision to the extent that the entity develops an effective exclusion device (Cornes and Sandler, 1996, p.4). Government provision remains one option (Cornes and Sandler, 1996, p.7). The Institutional Analysis and Development (IAD) framework on common-pool resources suggests the importance of self-organizing among affected stakeholders for solving the delivery problems of collective services rather than government-centered systems (E. Ostrom, 2009, p.419; E. Ostrom, 2005, p.222). For example, Tang (1991, p.42) shows greater responsiveness by community-centered establishments than bureaucracy-centered ones to the consumers of local water resources. It is true that government domination is not favored by the IAD framework. Yet it is far from an unconditional rejection of government. Rather, this framework acknowledges a positive role for government and sees local governments in particular as an important participant. Certainly, these are not the only participant but can still be an important one. What can be summarized from the discussion is that government still matters regarding the supply of collective services although it is not correct to claim that government must be the sole provider of these services. The current study makes an effort to classify the complete list of city services into collective and non-collective services according to whether an individual service is characterized by high or low excludability (TABLE 3-1). A similar effort was made by Stein about two decades ago (1990, p.74, 75). In his study, a total of 48 city services were grouped into 15 private services, 14 toll services, three common-pool resources, and 16 public services (Stein, 1990, p.74, 75). The classification of the current study is inspired by his original effort (Stein, 58 1990, p.74, 75). The service list derives from the Cities Financial Transactions Report Instructions (2005), which is given by the California State Controller to its cities. By law, all California cities are annually required to file financial information as indicated by the Instructions (California State Controller, 2005). TABLE 3-1: Classification of City Services Collective services (public services and common-pool resources) Non-collective services (private and toll services) Police Fire Emergency Medical Services Animal Regulation Weed Abatement Street Lighting Disaster Preparedness Streets/Highways/Storm Drains Street Trees and Landscaping Planning Construction and Engineering Regulation Enforcement Redevelopment Employment Community Promotion Physical and Mental Health Parks and Recreation Legislative (service) Management and Support Parking Facility Public Transit Airports Ports and Harbors Housing Hospitals and Sanitariums Solid Waste Sewers Cemeteries Marinas and Wharfs Libraries Museums Golf Courses Sports Arenas and Stadiums Community Centers and Auditoriums Water Distribution Natural Gas Distribution Electric Distribution N=18 N=18 Note: 1) Classification is based on and developed from Stein (1990, p.74, 75). 2) Source of the items of services: Schedule of Net Expenditures in California State Controller (2005). The full 36 city services are grouped into either collective (18) or non-collective (18) services. TABLE 3-1 succinctly shows that city governments are actively supplying both collective and non-collective services in practice. This classification has considerable potential for future research. For an illustration, Stein (1990, p.66) empirically shows that collective services tend to be produced in-house by government itself rather than indirect production means such as contracting, vouchers, or franchises. It turns out that these indirect means tend to be applied to non-collective services (Stein, 1990, p.66). For the purpose of the present study, this classification is utilized for 59 exploring which service type is affected more than the other by an external/internal fiscal condition and fiscal dependence. 3. LITERATURE REVIEW AND HYPOTHESES For convenience of research, this study assumes that city governments typically keep providing collective services above a certain minimum level, which ensures their very survival by barely satisfying the basic needs of residents for collective services. Below this level, they would be forced to shut down. Only after reaching this level, they may begin to provide additional amounts of services, collective or non-collective. The fact that government shutdown is very rare in reality shows that this assumption is quite reasonable. External or internal fiscal condition The traditional argument (e.g., Samuelson, 1954) that collective services are more or less essential to the existence of government enables the present study to examine the shift of the provision between non-collective and collective services, influenced by an external or internal fiscal condition. The present study employs one indicator of the fiscal condition external to cities: a state government surplus or deficit (-). A government surplus or deficit is the gap between government revenue and expenditure. When a city government is affected by a gradually worsening external fiscal situation, e.g., an accumulated state deficit, what prediction can be made? It would soon reduce the amount of monetary resources available to city governments. For some time, they could endure it. However, as the fiscal stress continues to grow over time, they would eventually have no choice but to slash spending (Levine, Rubin, and Wolohojian, 1981, p.11; Hou and Moynihan, 2008, p.145). Based on the arguments that different services are affected disproportionately and that collective services are more typically essential than non- 60 collective ones (e.g., Samuelson, 1954), the following prediction is made; a worsening fiscal environment will reduce the spending for non-collective services to a disproportionately greater extent than that for collective services. To take an example, let us assume that a government is currently providing 6 units of a non-collective and 4 units of a collective service (10 units in total). In percentage terms, this government has the composition of the 60% non-collective and the 40% collective service. 33 An exacerbation of the fiscal environment makes this government provide only 3 units of the non- collective and 3 units of the collective service (6 units in total) by giving up 3 units of the former and 1 unit of the latter. The non-collective service is reduced to a greater extent than the collective service, because the former is less typically essential than the latter according to the collective-services literature (e.g., Samuelson, 1954). This government gets a new composition of the 50% non-collective and the 50% collective service. The fiscal exacerbation resulted in the reduction of the non-collective service in percentage terms (60% => 50%), suggesting a positive relationship between a state government surplus and a non-collective service spending. Therefore, Hypothesis 1-1): higher state government surplus is positively related to higher spending for non-collective services. In this study, a state surplus is tested in not only an additive but also interactive way with government form, i.e., council-manager and mayor-council. The difference between these two is detailed at the preceding chapter on contracting. This study predicts that Hypothesis 1-1) will be observed more consistently in council-manager than mayor-council cities, because it is undermined easily to realize a rational expectation that a worsening fiscal environment will make 33 Reading in percentage terms is especially relevant for the present study, because the unit of its dependent variable is percentages. 61 a government cut non-core responsibilities, due to the dominating political motivations from mayors in the latter (Feiock and Jang, 2009, p.671). Although there is little direct research concerning the effect of government form on non- collective services, some studies confirm its effect on other fiscal outcomes. First, Saha (2011) employs data on U.S. cities from a 2002 Census of Governments survey. The analysis finds that the council-manager decreases spending for police, fire, highways, sewage, and parks by approximately 15 percent (Saha, 2011, p.162). Karuppusamy and Carr (2012) test the effect of government form on expenditure amounts aggregated from four city services in Michigan. Their results are overall modestly in favor of the effect (Karuppusamy and Carr, 2012, p.1549). On the other hand, MacDonald (2008) looking at the U. S. municipalities fails to find evidence for the effect on 1) total expenditure and 2) aggregated expenditure for police, parks and recreation, highways, fire, and sewage. Drawing on the discussion, the subsequent hypothesis is built: Hypothesis 1-1)a: higher state government surplus is positively related to higher spending for non-collective services when a government is under council-manager form, but not when it is under mayor-council form. The present study employs two indicators capturing the fiscal condition internal to cities, i.e., a city government's surplus or deficit (-) and unreserved fund balance. Fund balance is the amount of current assets less that of current liabilities (Marlowe, 2006, p.359). Unreserved fund balance, usually calculated as of the end of a fiscal year, is a monetary resource which a government can mobilize for the subsequent year (Marlowe, 2006, p.362). 34 These internal fiscal conditions are expected to work in virtually the same manner as the external fiscal condition. Given the differential importance between collective and non-collective 34 The difference in scope of unreserved fund balance exists between Marlowe (2006, p.362) and the present study. In his scope, only the undesignated share of fund balance appears to constitute unreserved fund balance (Marlowe, 2006, p.362). However, the present study considers the sum of each of the undesignated and the designated as unreserved. 62 services (e.g., Samuelson, 1954), the latter is likely to be more vulnerable to an aggravating internal fiscal condition than the former. As a consequence, the aggravation would result in a disproportionately large spending cut for non-collective services, compared to collective services. Although quantitative research on the relationship of internal fiscal conditions to collective spending for non-collective services is virtually nonexistent, two studies test the importance of internal fiscal condition on other fiscal outcomes. Marlowe (2005, p.49) finds a small expected impact of the unreserved undesignated portion of fund balance on expenditure despite the unexpected effects of the remaining portions from his analysis of 103 cities in Minnesota, 1990-2000. From her analysis about the 264 localities in the Chicago metropolitan region during 1997-2003, Hendrick discovers that the increase in unreserved fund balance is positively related to the growth of a surplus (2006, p.35). This line of reasoning leads us to the following hypothesis: Hypothesis 1-2): higher city government surplus is positively related to higher spending for non-collective services. Hypothesis 1-3): higher unreserved fund balance is positively related to higher spending for non-collective services. Fiscal federalism Three variables are employed to capture the level of intergovernmental fiscal dependence: the total amount of revenue originated from the federal government, general intergovernmental revenue from the state government, and functional intergovernmental revenue from the state government. General intergovernmental revenue is revenue given to a city government as general revenue, whereas functional intergovernmental revenue is functional revenue. 63 Their relationship with non-collective services can be approached from the recently flourishing literature on fiscal federalism (e.g., Weingast, 2009; Jin, Qian, Weingast, 2005) and soft budget constraints (e.g. Kornai, 1998). The recent fiscal federalism literature views the specific fiscal federalism of a country as an institutional arrangement. It explains fiscal or economic performances of different countries by how sound incentives each fiscal federalism produces. For example, the cases of China and Russia on economic development are compared in the way that China succeeded in giving sufficient fiscal incentives to sub-national government officials facilitating them to work hard on economic development, while Russia failed to do so (Jin et al., 2005, p.1740; Zhuravskaya, 2000, p.343). More specifically, Jin et al. employ as a measure of fiscal incentive "marginal retention rate of local revenue collection by provincial governments" (2005, p.1727). When this rate is high, according to the authors, provincial governments are regarded to have strong fiscal incentives. Utilizing a dataset on 29 provincial governments during 1970-1999, they find a significantly positive association of fiscal incentives with the development of the non-state sector and the reform of the state sector (Jin et al., 2005, pp.1721-1722). For comparison, Zhuravskaya shows that Russian regional governments reduce intergovernmental transfers to and tax shares with city governments by about 90 percent of the one-unit increase in the city own-source revenue (2000, p.357). This finding means that weak incentives are prevalent among Russian city governments, making them give up seeking their own-source revenue (Zhuravskaya, 2000, p.357). The author uses a data set on 35 major Russian cities during 1992-1997 (Zhuravskaya, 2000, p.347). This recent branch of fiscal federalism distinguishes itself from the traditional school of fiscal federalism (e.g., Oates, 1972) by naming itself Second Generation Fiscal Federalism 64 (SGFF) and by framing the traditional school as First Generation Fiscal Federalism (FGFF) (Weingast, 2009). The difference appears particularly stark on the matter of how to design intergovernmental transfer. The FGFF typically tends to acknowledge a positive role of intergovernmental transfer (Oates, 2008, p.317), implicitly assuming government as public- interested (Weingast, 2009, p.279). Two primary types of intergovernmental transfer, matching and block grants, are justified as follows; first, matching can be used for encouraging local governments to provide a collective service at an optimal level when positive externalities exist, e.g., a public park of a city enjoyed by residents of a neighboring city (Oates, 2008, p.317). Second, unconditional block grants can be given to fix fiscal inequalities among cities (Oates, 2008, p.317). In sheer contrast, the SGFF tends to underscore 1) the desirability of own-source revenue earned by local governments themselves (Weingast, 2009, p.283) and 2) the importance of becoming residual claimants on their own-source revenue (Jin et al., 2005, p.1723). The latter means that a considerable amount of their own-source revenue must be able to be kept by the very governments (Weingast, 2009, p.280). High intergovernmental transfer is worrisome due to its potential softening of the budget constraint which a local government has to face (Weingast, 2009, p.287; Kornai et al., 2003, p.1101). Developed by Kornai as a concept for explaining firm behaviors in then communist East European countries in the early 1980s (Kornai, 1998, p.11; Kornai et al., 2003, p.1095), the literature on the soft budget constraint has prospered. It is now applied to the OECD countries (e.g., Moesen and van Cauwenberge, 2000) or the United States (e.g., Duggan, 2000). The soft budget constraint is defined as follows: “the phenomenon that socialist firms are bailed out persistently by state agencies when revenues do not cover costs." (Kornai, 1998, pp.11-12). It 65 involves cultural and psychological as well as economic dimensions, sometimes called a syndrome (Kornai, 1998). This concept can be understood in terms of credible commitment (Kornai et al., 2003, p. 1101); at the beginning, the central government makes an apparently strong commitment that a failed firm will not be rescued in any event (Kornai et al., 2003, p.1101). Despite this commitment, failed firms are bailed out over time for various reasons (Kornai et al., 2003, p.1101). As other firms begin to doubt the credibility of the commitment, they come to be dis- incentivized from working hard on healthy fiscal performance such as solvency (Kornai et al., 2003, p.1101). It is worth remembering that non-collective services have the potential to produce revenues due to their ready exclusion, typically in the form of user fees and charges (Cornes and Sandler, 1996, p.34), as shown by a parking meter. The present study argues that high reliance on an upper government relaxing the budget constraint is likely to discourage a city government from providing non-collective services that generate revenue. As implied by Kornai et al. (2003, p.1104), certainly, if such high reliance is just a totally isolated one-time incident, it may not be very accurate to claim that it really is easing the budget constraint. However, if high reliance continues over time, it does ease that constraint, by making the city government 1) feel as if the transferred money were actually its own and 2) lose interest in revenue-generation for itself through non-collective services. Accordingly, the following Hypothesis 2-1) and its related three hypotheses, depending on which specific fiscal dependence measure is used, are established: Hypothesis 2-1): more intergovernmental revenue from an upper government is related to less spending for non-collective services. 66 Hypothesis 2-1)a: more intergovernmental revenue from the federal government is related to less spending for non-collective services. Hypothesis 2-1)b: more functional intergovernmental revenue from the state government is related to less spending for non-collective services. Hypothesis 2-1)c: more general intergovernmental revenue from the state government is related to less spending for non-collective services. Applying the claim of SGFF concerning intergovernmental transfer, the present study also makes a more detailed prediction that general intergovernmental revenue would discourage a city government from making its own effort toward revenue-generation to a greater extent than functional intergovernmental revenue. The former relaxes a city government's budget constraint to a greater extent than the latter, because the government can divert the former for whatever purpose it intends. The use of the latter is permitted only for a designated function at least in principle. For example, if a grant from the state government is designated for transportation, it is required to be spent only for this. Of course, the distinction between general and functional revenues can be blurred due to the practice of inter-fund transfer (Marlowe, 2006, p.367) as described earlier. Nonetheless, this prediction is still valid because inter-fund transfer is assumed to be costly. Hypothesis 2-2): the magnitude of the effect of intergovernmental revenue is greater for general than functional intergovernmental revenue. 4. DATA AND METHODS A panel dataset covering virtually all California cities over the past 19 years during 1991- 2009 for independent and control variables and 1992-2010 for the dependent variable is analyzed by fixed-effects estimation. This technique is known to be functional in weakening possible 67 omitted variable problems, since it controls those unobserved phenomena which are unique to each city and invariable over time (MacDonald, 2008, p.462). Variables TABLE 3-2 shows the exhaustive list of variables, detailed descriptions, and data sources. TABLE 3-2: Variables, Descriptions, and Sources of Data Variables Descriptions Sources %Non-collective services (sum of expenditures for housing, airports, parking facility, ports and harbors, public transit, community centers and auditoriums, golf courses, libraries, marinas and wharfs, museums, sports arenas and stadiums, cemeteries, hospitals and sanitariums, sewers, solid waste, electricity, natural gas, water) ÷ total expenditure × 100 <California City Financial Transactions>, RAND California 35 ; originating source- Cities Financial Transactions Report- Schedule of Net Expenditures %State surplus or deficit (-) t-1 lagged; (state revenue - state expenditure) ÷ state revenue × 100 FY2013 California Governor's Proposed Budget Summary 36 Council-manager (0,1) Not lagged*; 1 if a city has the council-manager form; otherwise, 0. ICMA, The Municipal Year Book %City surplus or deficit (-) t-1 lagged; (total revenue - total expenditure) ÷ total revenue × 100 <California City Financial Transactions>, RAND California; originating source- Cities Financial Transactions Report %Unreserved fund balance t-1 lagged; unreserved fund balance ÷ total revenue × 100 <California City Financial Transactions>, RAND California; originating source- Cities Financial Transactions Report %Revenue from Federal t-1 lagged; revenue from the federal government ÷ total revenue × 100 <California City Financial Transactions>, RAND California; originating source- Cities Financial Transactions Report %Functional revenue from State t-1 lagged; functional revenue from the state government ÷ total revenue × 100 <California City Financial Transactions>, RAND California; originating source- Cities Financial Transactions Report %General revenue from State t-1 lagged; general revenue from the state government ÷ total revenue × 100 <California City Financial Transactions>, RAND California; originating source- Cities Financial Transactions Report %Contracting t-1 lagged; contract expenditure ÷ total expenditure × 100 <California City Financial Transactions>, RAND California; originating source- Cities Financial Transactions Report- Schedule of Total Expenditures by Major Object Classification Population (Unit: person) t-1 lagged <Population and Demographic statistics>, RAND California 37 ; originating source-the U.S. Census Bureau $County personal income per capita t-1 lagged; 2005 constant dollars <Business and Economic Statistics>, RAND California 38 ; originating source- the Bureau of Economic Analysis %Non-collective services. The data for this dependent variable are obtained from an extensive fiscal database, called RAND California, developed by the RAND Corporation. The 35 http://ca.rand.org/stats/govtfin/cityfinance.html 36 http://www.ebudget.ca.gov/BudgetSummary/BSS/BSS.html 37 http://ca.rand.org/stats/popdemo/popdemo.html 38 http://ca.rand.org/stats/economics/economics.html 68 fiscal data in RAND California originate from the Cities Financial Transactions Report, which each and every city in California is legally required to submit to the California State Controller within a few months after the completion of a fiscal year (California State Controller, 2005, p.5). This variable represents a percentage share of the sum of relevant expenditures for non-collective services out of total expenditure of a city government. TABLE 3-3 provides useful information of how much each of the 18 non-collective services accounts for out of total expenditure. TABLE 3-3: Descriptive Statistics of Individual Non-collective Services Variable Observations Mean Standard Deviations Min. Max. %Non-collective services 9,310 27.63 18.59 0 92.25 %Housing 9,364 1.56 4.43 0 69.73 %Airports 9,364 0.46 2.14 0 55.33 %Parking facility 9,364 0.29 1.36 0 46.99 %Ports and harbors 9,366 0.26 2.10 0 45.39 %Public transit 9,362 1.35 2.79 0 43.20 %Community centers and auditoriums 9,309 0.84 2.74 0 60.61 %Golf courses 9,313 0.50 2.64 0 65.66 %Libraries 9,314 1.08 2.43 0 42.07 %Marinas and wharfs 9,312 0.10 0.97 0 41.67 %Museums 9,310 0.06 0.35 0 10.21 %Sports arenas and stadiums 9,312 0.06 0.65 0 16.44 %Cemeteries 9,312 0.02 0.20 0 6.55 %Hospitals and sanitariums 9,312 0.18 2.96 0 69.85 %Sewers 9,311 7.63 9.27 0 86.57 %Solid waste 9,309 3.10 4.83 0 51.29 %Electric 9,365 1.96 8.97 0 85.09 %Gas 9,366 0.08 1.12 0 34.12 %Water 9,365 7.94 9.10 0 76.75 Note: Ideally, the mean sum of the 18 items (27.47%) must equal the mean of %Non-collective services (27.63%). Yet here it is slightly different due to rounding and the separate process of data cleaning. That is, the data cleaning of %Non-collective services and that of the 18 non-collective service items were conducted independently. This also explains why each of %Community Centers and Auditoriums and %Solid Waste has the fewer number of observations than that of %Non-collective services. One advantage of this variable is that it can capture a variety of service delivery forms - in-house production, contracting, and vouchers among others. Nevertheless, it is also acknowledged that this variable is limited in capturing the whole complexity of non-collective service provision. That is, this is incapable of capturing a service delivery form, unless the form necessitates a certain amount of spending from the government side. As an illustration, consider: even if a city government does not spend even a cent on water distribution, it does not 69 necessarily mean that the city residents are not able to enjoy drinking water of decent quality. It is totally possible that the water is produced and provided by a third-party private firm or another separate special-purpose government. In sum, the zero-dollar expenditure for water distribution in the dataset for the current study may simply mean that a city government does not have to spend money for the service due to the existence of alternative service arrangements. %State surplus. This variable measures the external fiscal condition which influences city governments. This state-level variable holds a particular relevance for the present study, given the powerful fiscal, regulatory, and administrative tools that the state government is entitled to use on city governments. A surplus of the state government is the difference between state revenue and expenditure. The specific measure in use is a percentage share of a state surplus out of its revenue. When the amount of expenditure exceeds that of revenue, it is called a deficit. Accordingly, if %State surplus is negative, it is a state deficit in percentage terms. The data originate from FY2013 California Governor's Proposed Budget Summary. Council-manager. The details about this variable are found at the chapter on contracting. %City surplus. This measure is intended to represent an aspect of the internal fiscal condition of a city government along with %Unreserved fund balance. %City surplus is a percentage share of a surplus or deficit (-) out of total revenue. The data are obtained from the RAND California but originally from the Cities Financial Transactions Report. %Unreserved fund balance. Fund balance is made up of three portions: reserved, unreserved designated, and unreserved undesignated. Each is defined as follows: first, a reserved fund balance is "the portion of fund equity that is legally segregated for specific purposes" (California State Controller, 2005, p.127). Second, an unreserved designated fund balance refers to "the portion of fund balance segregated to indicate tentative plans for financial resource 70 utilization in a future period, such as for general contingencies or for equipment replacement. Such designations reflect tentative managerial plans or intent..." (California State Controller, 2005, p.127). Reserved fund balances and designated ones are similar to each other in terms of the fact that the use of money is held for a certain purpose. The difference between the two lies in the fact that reservations are legally expressed such as in municipal codes while designations are not (Marlowe, 2006, p.366). Thus, the reserved portion is more restricted than the designated one in terms of the leeway enjoyed by a government. Third, an unreserved undesignated fund balance means "the portion of fund balance that is unrestricted as to its use" (California State Controller, 2005, p.127). This portion is readily available for use. FIGURE 3-1 displays the structure of fund balance. FIGURE 3-1: Structure of Fund Balance Source: California State Controller, 2005, p.127. The present study utilizes the total amount of an unreserved fund balance created by merging its designated and undesignated portions. The total amount is converted to a percentage share out of total revenue. The relevant information is acquired from the RAND California, developed originally from the Cities Financial Transactions Report. %Revenue from Federal, %Functional revenue from State, %General revenue from State. %Revenue from Federal represents a percentage share of a city's revenue, which is transferred from the federal government, out of the city's total revenue irrespective of whether it 71 is general or functional revenue in its scope of use. %Functional revenue from State is a percentage share of a city's functional revenue, which is contributed by the state government, out of the city's total revenue. In a similar way, %General revenue from State is calculated by transforming a city's general revenue, which is given by the state government, as a percentage share out of the city's total revenue. The common data source of these three streams of intergovernmental revenue is RAND California and originally the Cities Financial Transactions Report. Controls include contracting (%Contracting), city population (Population), and personal income per capita by county ($County personal income per capita). %Contracting is the same measure used as the dependent variable at the preceding chapter where an elaboration on this measure appears. $County personal income per capita is expressed in 2005 constant dollars by utilizing the GDP Implicit Price deflator. 39 Their respective data origins are specified at TABLE 2. All the independent and control variables take a one-year lag. The only exception is Council-manager. TABLE 3-4 shows summary statistics of the base variables. 39 Source: Bureau of Economic Analysis (www.bea.gov) => "Interactive Data" => "National Data" => "GDP & Personal Income" => "SECTION 1 - DOMESTIC PRODUCT AND INCOME" => "Table 1.1.9. Implicit Price Deflators for Gross Domestic Product," last revised by BEA on 06/28/2012. 72 TABLE 3-4: Descriptive Statistics of Base Variables Variable Observations Mean Standard deviations Min. Max. %Non-collective services 9,310 27.63 18.59 0 92.25 Council-manager (0, 1) 8,850 0.94 0.22 0 1 %State surplus 9,460 -5.18 4.75 -15.11 3.05 %City surplus 8,898 3.47 17.21 -219.89 93.78 %Unreserved fund balance 8,898 48.08 51.30 -255.14 737.77 %Revenue from Federal 8,491 4.62 6.82 0 86.70 %Functional revenue from State 8,896 6.04 5.39 0 86.22 %General revenue from State 8,898 5.19 4.47 0.00 66.66 %Contracting 8,894 20.98 17.62 0 91.64 Population (Unit: person) 9,373 58,013.98 191,757 89 3,831,868 $County personal income per capita 9,460 34,139.43 10,904.12 17,101.46 86,405.25 In addition, the average trends of the seven major explanatory and dependent variables are shown at FIGURE 3-2: Although each variable is expressed in percentage terms, the comparison among them requires caution because they use different items as denominators. Specifically, %Non-collective services and %State surplus use city expenditure and state revenue, respectively, as denominators. The remaining five variables employ city revenue as a common denominator. FIGURE 3-2: Average Trends of Major Variables 73 Models All eight models are estimated by the fixed-effects technique. Models 1 through 6 are baseline models. Models 7 and 8 are full models, each of which is respectively established to test the effect of %State surplus independently without and interactively with Council-manager. The size of the effect of %State surplus is expected to be moderated, depending on whether a city is characterized by the mayor-council (0) or council-manager (1). The two components of %State surplus t-1 ×Council-manager t , i.e., %State surplus t-1 and Council-manager t , are included simultaneously with the interactive term in Models 6 and 8 following the guidelines of Brambor et al. (2006, p.66). All the models are reported with cluster-robust standard errors due to the common detection of both heteroskedasticity and autocorrelation throughout the models. %State surplus t- 1 ×Council-manager t and %State surplus t-1 in each of Models 6 and 8 are signaled for the existence of severe collinearity. Yet multicollinearity between the interactive term and its components does not constitute a valid reason to omit either one (Brambor et al., 2006, p.70). The F-test results for a choice between pooled OLS and fixed-effects estimation are in favor of the latter throughout all eight models. Similarly, most of the Hausman tests for either fixed- or random- effects estimation prefer fixed-effects estimation to its alternative throughout all the models except Models 3 and 6. The Hausman tests in Models 3 and 6 fail to perform properly. 5. RESULTS TABLE 3-5 shows the very interesting estimated coefficients of the eight models. First, the statistical significance of %State surplus t-1 is found at the 10% level only in Model 7, not in Model 5. In Model 7, the 1% increase in a state surplus out of state revenue relates to the 0.073% increase in spending for non-collective services out of total expenditures of a city government. 74 Neither Model 6 nor 8, which considers the differential size of its effect depending on the presence of Council-manager t , shows any statistical significance. Thus, Hypothesis 1-1) is modestly supported, and Hypothesis 1-1)a fails to be supported. Second, %City surplus t-1 reveals that the 1% increase in a city surplus out of its total revenue leads to the 0.018%, 0.019%, and 0.024% decrease in the expenditure for non-collective services out of total expenditure in Models 4, 7, and 8, respectively. All three coefficients constantly stay statistically significant. This result indicates a negative relationship as opposed to the predicted positivity of Hypothesis 1-2). Third, %Unreserved fund balance t-1 yields a similar result to that of %City surplus. In other words, a 1% increase in unreserved fund balance out of total revenue is associated with a 0.013% (Models 4 and 7) or 0.017% (Model 8) reduction in the non-collective service expenditure out of total expenditure. These keep statistically significant across the models. Again, this constant negative association unanticipated by Hypothesis 1-3) requires an explanation. Fourth, the estimates of %Revenue from Federal t-1 show a solid negative relationship with statistical significance throughout Models 3, 7, and 8. The 1% increase in the federal intergovernmental revenue of a city out of its total revenue leads to a decrease in the non- collective service expenditures out of total expenditures by 0.068% (Model 3), 0.079% (Model 7), and 0.090% (Model 8). As such, Hypothesis 2-1) and Hypothesis 2-1)a in particular are supported. Fifth, the estimates of %Functional revenue from State t-1 are statistically significant in the Models 3, 7, and 8. The 1% growth in the state-originated functional revenue of a city out of its total revenue is associated with the shrinkage in the expenditure share of non-collective services 75 by 0.123% (Model 3), 0.129% (Model 7), and 0.124% (Model 8). Thus, Hypothesis 2-1) and Hypothesis 2-1)b are confirmed. Sixth, Hypothesis 2-1) and Hypothesis 2-1)c for %General revenue from State t-1 are supported: the three models (Models 3, 7, and 8) all show a negative relationship of the state intergovernmental general revenue of a city out of its total revenue on the spending portion of non-collective services. All the relevant coefficients achieve a certain level of statistical significance. The 1% increase in %General revenue from State t-1 results in the spending cut of non-collective services by 0.208% (Model 3), 0.225% (Model 7), and 0.243% (Model 8). Seventh, the comparison between each coefficient of %Functional revenue from State t-1 and that of %General revenue from State t-1 provides evidence for Hypothesis 2-2). That is, the effect magnitude of %General revenue from State t-1 is consistently larger than that of %Functional revenue from State t-1. 76 TABLE 3-5: Fixed Effects (Within) Regression Coefficients Independent & control variables Dependent variable: %Non-collective services (= (sum of expenditures for housing, airports, parking facility, ports and harbors, public transit, community centers and auditoriums, golf courses, libraries, marinas and wharfs, museums, sports arenas and stadiums, cemeteries, hospitals and sanitariums, sewers, solid waste, electricity, natural gas, water) ÷ total expenditure × 100) Model1 Model2 Model3 Model4 Model5 Model6 Model7 Model8 %State surplus t-1 -0.022 (0.028) -0.071 (0.099) 0.073* (0.039) 0.096 (0.102) %State surplus t- 1 ×Council-manager t 0.065 (0.097) -0.009 (0.098) Council-manager t -0.826 (1.091) -0.952 (1.285) %City surplus t-1 -0.018** (0.008) -0.019** (0.008) -0.024*** (0.008) %Unreserved fund balance t-1 -0.013* (0.007) -0.013* (0.007) -0.017*** (0.004) %Revenue from Federal t-1 -0.068** (0.029) -0.079*** (0.029) -0.090*** (0.032) %Functional revenue from State t-1 -0.123*** (0.038) -0.129*** (0.038) -0.124*** (0.035) %General revenue from State t-1 -0.208*** (0.070) -0.225*** (0.072) -0.243*** (0.079) %Contracting t-1 -0.019 (0.014) -0.009 (0.014) -0.016 (0.014) Population t-1 0.000 (0.000) 0.000 (0.000) 0.000 (0.000) 0.000 (0.000) 0.000 (0.000) 0.000 (0.000) 0.000 (0.000) 0.000 (0.000) $County personal income per capita t-1 -0.000 (0.000) -0.000 (0.000) 0.000 (0.000) -0.000 (0.000) -0.000 (0.000) -0.000 (0.000) 0.000 (0.000) 0.000 (0.000) Constant 27.267*** (1.920) 27.984*** (1.995) 29.151*** (1.703) 28.102*** (2.007) 27.487*** (2.562) 28.563*** (2.787) 30.671*** (2.368) 31.813*** (2.592) R-sq (within), % 1.24 1.38 2.21 1.74 1.24 1.52 2.71 3.01 R-sq (between), % 2.83 6.76 6.89 10.16 2.83 4.36 14.97 20.94 R-sq (overall), % 2.52 5.25 5.40 7.35 2.52 3.34 10.92 14.62 P value for F test .00 .00 .00 .00 .00 .00 .00 .00 Number of observations 9,229 8,729 8,341 8,733 9,229 8,683 8,338 7,918 Number of cities 470 470 470 470 470 452 470 452 Time scope (Dependent variable) 1991- 2010 1992-2010 1992-2010 1992-2010 1991-2010 1991-2010 1992-2010 1992- 2010 Note: Cluster-robust standard errors are noted in parentheses. Year dummies are included. Probability levels are denoted as follows: *** p<.01, ** p<.05, * p<.10. 6. DISCUSSION The analysis stimulates a fruitful discussion. First, the effect of a state surplus was not strong across relevant models. Yet its positive effect appeared in the full additive Model 7. This implies that non-collective services become an immediate victim of the fiscal environmental aggravation because local governments prioritize preserving collective services such as police, core government responsibilities (e.g., Samuelson, 1954), over non-collective ones. 77 In addition to their vulnerability to external fiscal conditions, local governments are subject to internal fiscal condition. Consistent with the logic applied to external fiscal situations, a positive association was initially proposed; its exacerbation would make local governments concentrate on collective services by giving up the existing level of non-collective ones. Very intriguingly, the empirical analysis indicates the opposite, an inverse association; its deterioration led to a greater supply of non-collective services. This may suggest opposite mechanisms on the part of the external and internal fiscal conditions. Despite the lack of a definitive answer, a speculation is worth: the inverse association makes us think twice about the revenue-generation aspect of non-collective services, typically in user fees and charges, as an exclusion mechanism (Cornes and Sandler, 1996, p.34). Lacking this property, traditionally, the financing of collective services relies on taxes (Hyman, 2002, p.142). The problem is that levying a tax is not a convenient option given the popular sentiment against it. Therefore, it is possible that local governments without sufficient fiscal resources are tempted to increase the non-collective service provision, expecting more fees and charges. Although the revenue from fees and charges is considered functional revenue, it could be moved to the general fund to support collective services through inter-fund transfer (Marlowe, 2006, p.367). All three factors of fiscal dependence compellingly make the case for the recent fiscal federalism literature (e.g., Weingast, 2009). The relevant findings give support for the argument that high fiscal dependence on an upper government discourages local governments from seeking their own-source revenue opportunities by softening their budget constraints (Weingast, 2009, p.287; Kornai et al., 2003, p.1101). Another intriguing argument, empirically supported here, is the differential incentive effects caused by different types of intergovernmental transfer (Weingast, 2009, p.284). General 78 intergovernmental revenue had a stronger effect than its functional counterpart in making local governments give up further income in fees and charges. This implies that the former without any limitation in its area of use tends to relax the budget constraints to a greater extent than the latter. This finding also suggests that the separation between general and functional revenues is effective to an extent as an institutional barrier despite the frequent utilization of inter-fund transfer (Marlowe, 2006, p.367). If functional revenue could have been directed, through the transfer without any substantive cost, to any area of service, there should have been no discernible difference between their magnitudes of effect. 7. CONCLUSION The research on collective and non-collective services has traditionally been disconnected from that on public financial management. In an effort to connect each other, the present study argued that the service classification holds considerable financial implications. Simply put, collective services are typically funded by taxes (Hyman, 2002, p.142) due to free-riding incentives originating from difficult exclusion (Tiebout, 1956, p.417), while non-collective services tend to be financed by user fees and charges due to relatively easy exclusion (Cornes and Sandler, 1996, p.34). Spending for non-collective services can be rewarded as future revenue, while that for collective ones cannot. Local governments appear to find ways to finance local services considering the exclusion property. In general, the empirical findings provide support for the utility of the classification. As expected, though weak, it was suggested that non-collective services are the direct target of budget cuts in the face of a deteriorating external condition. Intriguingly, a government with a deteriorating internal fiscal condition appeared to escalate the provision of non-collective services, possibly in anticipation of creating extra revenues. Certainly, subsequent research is 79 needed to investigate why the external and internal fiscal conditions work in such opposite ways. It was well demonstrated that high fiscal dependence typically dis-incentivizes city governments from pursuing future revenue prospects (Weingast, 2009, p.287; Kornai et al., 2003, p.1101). How intergovernmental transfer is designed was also shown to matter. The geographical focus of the present study is California cities. In consequence, it is not promoted to statistically generalize these findings to localities outside California. Nevertheless, it is still intuitively conceivable that the findings could provide a useful ground for understanding the fiscal behaviors of non-California localities, as any local government in the United States tends to provide a package of non-collective as well as collective services. Besides, they are likely to be affected to an extent by such explanatory factors as the external or internal fiscal condition and intergovernmental fiscal dependence. Another limitation of the present study is, as indicated in the variable description section, that the dependent variable is unable to capture the alternative provision of a non-collective service by another separate entity which does not require government spending. APPENDIX 3-1: Correlations 1 2 3 4 5 6 7 8 9 10 11 12 1.%Non-collective services 1 2.%State surplus t-1 .01 1 3.%State surplus t-1 × Council-manager t .03* .95* 1 4.Council-manager t -.04* .00 -.22* 1 5.%City surplus t-1 -.06* .09* .09* .00 1 6.%Unreserved fund balance t-1 -.28* -.04* -.06* .02* .10* 1 7.%Revenue from Federal t-1 .11* .03* .04* -.06* -.01* -.08* 1 8.%Functional revenue from State t-1 -.12* -.10* -.09* -.01 -.00 .06* .08* 1 9.%General revenue from State t-1 -.28* .12* .13* -.00 .02* .07* .02* .18* 1 10.%Contracting t-1 -.29* -.01 -.02* .04* .10* .21* -.01* .14* .26* 1 11.Population t-1 .12* -.00 .04* -.22* -.02* -.08* -.00 -.08* -.07* -.06* 1 12.$County personal income per capita t-1 -.26* -.10* -.10* .05* .01 .09* -.18* -.09* -.19* .05* .04* 1 Note: * significant at 10% level. 80 CHAPTER 4. The Fiscal Effects of Term-limits, Council-manager Form, and Charters: Evidence from California Cities, 1992-2009 1. INTRODUCTION This study is one of the first quantitative research studies testing the fiscal effect of term- limits at the local level. Most of the existing research (e.g. Besley and Case, 1995) tends to focus on U.S. state governments. This is unfortunate because local governments are a fertile ground to evaluate competing arguments regarding term-limits. One apparent reason for this oversight is the lack of local term-limits data. The present study aims to contribute to the literature by utilizing a unique term-limits dataset on approximately 250 California cities during 1992-2009. In addition, the present study employs two additional institutional variables: government form (council-manager vs. mayor-council) and charter status (charter vs. general-law). These are known to be fundamental local institutions which regulate the functioning of government. Thus, it would be intriguing to explore their influence on fiscal performance. Consistent with the accepted assumption on the fiscal preferences of average voters in the relevant term-limits literature (e.g., Besley and Case, 1995, p.781; Alt, Bueno de Mesquita and Rose, 2011, p.176), the present study also assumes average voters to be "fiscal conservatives" who like less revenue and less expenditure (Alt, Bueno de Mesquita and Rose, 2011, p.176) under the ceteris paribus condition that all other factors are held constant. This study further assumes that average voters prefer more government savings. It could generally be agreed that average voters are typically pleased when the government provides a certain quality of services by spending less money (less revenue and less spending) and, if possible, by saving money (more fund balance). 81 This study introduces two competing arguments around term-limits; the first maintains that term-limits cause beneficial fiscal performance because they enhance electoral competitiveness (Daniel and Lott, 1997, pp.168-169) and somehow give a council member the courage to make a right decision (Schelker, 2012, pp.31-32). The second argument discredits the first by arguing that term-limits make an incumbent behave in a morally hazardous way and take away the voters' capacity to offer one additional term if an incumbent is really good (Alt, Bueno de Mesquita and Rose, 2011, p.171). The empirical finding lends support to neither of these arguments. No fiscal effect of term-limits suggests either that the positive effect may have canceled out the negative one or that term-limits may have been meant from the outset not as a fiscal institution but as a political one (Friedman and Wittman, 1995, p.81). It is argued that the council-manager form respects the fiscal preferences of average voters by effectively protecting them from city politics (Feiock and Jang, 2009, p.671). The empirical analysis finds evidence in favor of this argument. The charter is believed to be an institution which creates substantive opportunities for accommodating average fiscal preferences (Benton, 2002, p.474); the finding turns out to be mixed. The present study begins by describing two contrasting perspectives on term-limits, followed by the brief discussion on the difference between the council-manager and mayor- council forms and then that between the charter and general-law statuses. The hypotheses formulated from the literatures will be rigorously tested and the meaning of the results discussed. 2. LITERATURE REVIEW AND HYPOTHESES Term-limits While the research on term-limits has been plentiful at the U.S. state level, no agreement is yet reached concerning their relationship with fiscal performance. Two contrasting literatures 82 compete with each other. One is affirmative on term-limits; the other isn't. The present study introduces each of the two in order. The first body of literature (Daniel and Lott, 1997; Schelker, 2012) typically affirms a desirable function of term-limits; Daniel and Lott (1997, pp.167-168) argue that term-limits typically discount the utility of a political agency for which candidates compete. This reduced utility dis-incentivizes candidates from spending monetary resources to win an election (Daniel and Lott, 1997, p.168). Term-limits boost the extent of electoral competition, as they force out long-time incumbents with established popularity among voters (Daniel and Lott, 1997, p.168, 169). The authors measure the extent of competition by the number of major candidates and the spread of votes cast for the highest two vote-getters (Daniel and Lott, 1997, p.181). They predict an inverse relationship with campaign spending and a positive relationship with competition. Their empirical results provide support for these two predictions with the elections data on the Assembly and Senate of California during 1976-1994 (Daniel and Lott, 1997, p.166, 174, 179). Second, Schelker (2012, p.31) introduces one perspective for and one against term-limits. According to the perspective for term-limits, these measures incentivize political officials to implement a policy which is socially beneficial in the long run even if it is politically unpopular today for the same reason advanced by Daniel and Lott (1997), i.e., the disappearing usefulness of a political position (Schelker, 2012, p.31-32). He does not take sides in the term-limits debate and accordingly does not hypothesize the direction of term-limits on fiscal performance (Schelker, 2012, p.34). However, his analysis of the U.S. states during 1990-1999 indicates that term-limits on auditors tend to improve bond ratings (Schelker, 2012, p.28, 38, 41). Overall, this body of literature is directly applicable to the present study. Term-limits would remove senior members from the city council, thereby enhancing electoral 83 competitiveness (Daniel and Lott, 1997, p.168, 169). This enhanced competitiveness would facilitate the production of information regarding candidates' abilities, which helps lower the risk of adverse selection and raise the likelihood of electing a more capable candidate (Alt, Bueno de Mesquita and Rose, 2011, p.171). Having a more capable council member naturally leads one to expect decent fiscal performance. In fact, the positive relationship between electoral competitiveness and fiscal performance is demonstrated by Baker (2003, p.340). In addition, the depreciated worth of a council membership due to term-limits would give a term-limited member the courage to enact a fiscal policy that is socially benign in the long run even if it faces current resistance from a powerful faction (Schelker, 2012, p.31-32). On the other hand, the body of literature against term-limits (Besley and Case, 1995; 2003; Alt, Bueno de Mesquita, and Rose, 2011; Erler, 2007) tends to worry that term-limits deprive incumbents of the election opportunity itself through which incumbents are disciplined (Alt, Bueno de Mesquita and Rose, 2011, p.171). The prospect of reelection encourages an incumbent to work hard for his or her constituency, thereby lessening moral hazard (Alt, Bueno de Mesquita and Rose, 2011, p.171). Elections also help voters sort out more competent candidates from the rest by providing the large quantity of information on their qualities revealed during an electoral campaign, lowering the probability of adverse selection (Alt, Bueno de Mesquita and Rose, 2011, p.171). Alt, Bueno de Mesquita and Rose (2011, p.171) call these effects the "accountability effect" and "competence effect" respectively. One study by Besley and Case (1995) is a frequently cited work which investigates the significance of gubernatorial term-limits on state fiscal outcomes during 1950-1986 (p.773). The authors argue against term-limits by emphasizing the possibility that term-limits create moral hazard for incumbents as they will be deposed soon. They do not have to worry about the quality 84 of their performance because voters can no longer discipline them through the ballot box (Besley and Case, 1995, p.772). If they are entitled to run for reelection, in contrast, they will surely care about their performance (Besley and Case, 1995, p.773). Therefore, it is hypothesized that the fiscal performance of a term-limited governor is not as good as that of a non-term-limited governor (Besley and Case, 1995, p.773). The empirical results provide evidence for this hypothesis (Besley and Case, 1995, p.780). Specifically, out of seven models with fiscal dependent variables (sales taxes, income taxes, corporate taxes, total taxes, state expenditure, state minimum wage, maximum weekly benefits), four confirm the significance of term-limits (sales taxes, income taxes, state expenditure, state minimum wage) (Besley and Case, 1995, p.780). Interestingly, the significance of term-limits partly disappears in the follow-up study expanding the time scope to 1950-1997 (Besley and Case, 2003, p.55). The effect on state spending still appears, but that on taxes is not detected any more (Besley and Case, 2003, p.55). In pursuit of an explanation for the mysterious disappearance, Alt, Bueno de Mesquita, and Rose (2011, p.173) develop a measure which distinguishes term-limits on and also term- lengths of a governor. This measure is more sophisticated than that of Besley and Case (1995; 2003), which considers only whether an incumbent is in his last term or not (Besley and Case, 1995, p.774). They reason that the negative effect of term-limits disappears in Besley and Case (2003, p.55) possibly due to the "competence effect" (Alt, Bueno de Mesquita and Rose, 2011, p.176). This means that a term-limited incumbent typically faces the temptation of a moral hazard due to term-limits but also nurtures meritorious expertise during his incumbency (Alt, Bueno de Mesquita and Rose, 2011, p.179). This reasoning is empirically backed up (Alt, Bueno de Mesquita and Rose, 2011, p.179). In addition, their second empirical model indicates that a 85 term-limited incumbent has a tendency to show higher taxes and higher borrowing cost than a non-term-limited one does (Alt, Bueno de Mesquita and Rose, 2011, p.182). From the data on the U.S. states during 1977-2001 (p.482), Erler (2007) confirms that term-limits are associated with larger spending (p.485), consistent with the body of scholarship against term-limits. This body of scholarship is immediately transferrable to the local context of the present study. That is, term-limits would tend to remove an incentive to work diligently from current council members and strip well-performing incumbents of an additional opportunity (Alt, Bueno de Mesquita and Rose, 2011, p.171). Consequently, it is expected that term-limits result in mediocre fiscal performance (Besley and Case, 1995). All in all, if the present study adopts the first body of scholarship acknowledging a beneficent role of term-limits, it is hypothesized that term-limits are associated with good fiscal performance: less expenditure, less revenue, and more savings. If, on the other hand, this study employs the second body typically discrediting the beneficial role, it is expected that term-limits will be associated with poor fiscal performance: more expenditure, more revenue, and less savings. Because the direction of the effect caused by term-limits is not yet determined in the literature, the present study is open to either possibility, as argued by Schelker (2012, p.34). Council-manager form and charter status Council-manager vs. mayor-council. The present chapter does not go into all the details concerning the different features of the council-manager and mayor-council forms, as they are already elaborated at the preceding Chapter 2 on government contracting. For the purpose of the present study, it is sufficient to reiterate that the council-manager form is more effective in isolating city administration from the influence of city politics than is 86 the mayor-council form (Feiock and Jang, 2009, p.671). This implies that the fiscal outcomes preferred by average voters are less likely to be distorted by city politics in the council-manager form than in the mayor-council form. Accordingly, the present study predicts that the council- manager form is associated with decent fiscal performance; in other words, Hypothesis 1: council-manager form is positively related to fiscal performance. Hypothesis 1-1): council-manager form is inversely related to total expenditure. Hypothesis 1-2): council-manager form is inversely related to total revenue. Hypothesis 1-3): council-manager form is inversely related to revenue from taxes. Hypothesis 1-4): council-manager form is inversely related to revenue from fees and charges. Hypothesis 1-5): council-manager form is positively related to unreserved fund balance. Charter vs. general-law. The difference between charter and general-law governments is not described in great detail here, as it is found at the preceding Chapter 2. What is relevant to the present study is that a charter provides its own government with great flexibility to carry out what average voters desire within the areas of municipal affairs (Benton, 2002, p.474). A substantial part of fiscal policy falls into these areas. 40 Whereas the council members in a general-law government cannot raise their own salaries beyond the cap imposed by state law, for example, those in a charter government are not restrained by the cap. 41 Thus, it is conceivable to reason that a desired fiscal performance is more often observed in charter than general-law governments; that is, Hypothesis 2: a charter is positively related to fiscal performance. Hypothesis 2-1): a charter is inversely related to total expenditure. 40 League of California Cities, General Law City v. Charter City, www.cacities.org/chartercities, accessed 09/05/2011 41 League of California Cities, General Law City v. Charter City, www.cacities.org/chartercities, accessed 09/05/2011 87 Hypothesis 2-2): a charter is inversely related to total revenue. Hypothesis 2-3): a charter is inversely related to revenue from taxes. Hypothesis 2-4): a charter is inversely related to revenue from fees and charges. Hypothesis 2-5): a charter is positively related to unreserved fund balance. 3. DATA AND METHODS In order to test the impact of these institutional forces, the present study analyzes a large panel dataset on approximately 250 California cities where more than 25,000 people resided during 1991-2008 for independent and control variables and 1992-2009 for dependent variables. The number of included cities is about half the total number of cities in California. Variables Dependent variables. All five per-capita indicators of fiscal performance are utilized: total expenditure ($), total revenue ($), revenue from taxes ($), revenue from service fees and charges ($), and unreserved fund balance ($). The data are obtained from RAND California, a statewide city fiscal database, but ultimately originated in the Cities Financial Transactions Report. Each of these indicators captures a vital fiscal aspect of government operation commonly observed across governments: how much they earn (total revenue) by which means (revenue from taxes, revenue from service fees and charges), how much they spend (total expenditure), and how much they save for future use (unreserved fund balance). Each of these has been utilized through operationalization as a dependent variable in the existing research: for example, revenue from taxes and total expenditure in Besley and Case (1995, p.780), revenue from fees and charges in Gordon (2009, p.46), total revenue in Crain and Tollison (1993, p.157), and unreserved fund balance in Hendrick (2006, p.31). 88 Term-limits. There are two known styles of constructing a measure for term-limits. The first style is that a dichotomous value is given on an individual official basis, depending on whether an incumbent official is forced by term-limits to give up the upcoming election at a time t or not (e.g., Besley and Case, 1995, pp.773-774; 2003; Alt, Bueno de Mesquita and Rose, 2011, p.177). The second is that a dichotomous value is given on a jurisdiction basis, based on whether a jurisdiction has term-limits as of a time t or not (e.g., Erler, 2007, p.483; Crain and Oakley, 1995, p.7; Crain and Tollison, 1993, p.156). The present study employs this second style, considering data characteristics and availability. More specifically, it is collected and coded in the following manner: Data collection began with the survey results 42 conducted by the City of Thousand Oaks 43 , which contained cross-sectional information on term-limits for all 481 cities in California as of the year 2011. Those cities, which were reported as being without term-limits in the Thousand Oaks document, are given 0 throughout the whole time span for the present study, 1991-2008; that is, this study assumes that those cities were without term-limits throughout this time span. This assumption makes sense because institutions have tendency to linger and not to be easily repealed, especially within a given period of time, although repeal is not entirely impossible. 44 As for those cities which were reported to have term-limits in the Thousand Oaks document, this study has conducted a careful and extensive search for when each of the term- limits was first approved by a city's voters through all the available sources such as the municipal code of ordinances, the charter, the county registrar of voters, California Elections Data Archive 42 The author appreciates the generosity of the City of Thousand Oaks in allowing him to utilize the information on term-limits for this research project. 43 The City of Thousand Oaks, April 12, 2011, "State of California Councilmember Term Limits Survey by City in Alphabetical Order," as Attachment #4 of the Memorandum, Thousand Oaks Councilmembers Term Limits Initiative Impact Analysis Report & Scheduling for November 6, 2012 Election. The survey results can be found at: http://www.ci.thousand- oaks.ca.us/government/depts/city_clerk/weblink.asp => CTO => CITY COUNCIL/REDEVELOPMENT AGENCY (RDA) => Agenda Packets => 2011 => 04/12/2011. 44 Term-limits in the city of Palmdale were approved in 2001and then repealed in 2009. 89 (CEDA) 45 , Smart Voter 46 , Ballotpedia 47 , newspapers (especially the Los Angeles Times), Petracca and O'Brien (1999), and/or personal inquires via email with the city clerk. The search continued until this author of the present study became confident of the accuracy of the information. Any city for which this author was not able to determine its approval year of term- limits with confidence remains unknown, i.e., missing in the dataset. As a result, a total 94 cities are identified as currently having term-limits as of March 2013. 48 APPENDIX 2 shows their respective approval years. Each of these 94 cities is assigned the value 1, beginning one year after the year of popular approval of initial term-limits on its city council members; otherwise, the city is assigned the value 0. To clarify: first, only term-limits on city council members are included, so term- limits on other offices such as mayor or commissions/boards are not considered. Second, the revision of the existing term-limits is not counted - e.g., revision from three term limits to two term limits. Third, only legally binding term-limits are included, not advisory term-limits such as term-limits in general-law cities before 1996. According to California Government Code Section 36502, a general-law city can adopt term-limits 1) beginning in 1996 and 2) by the voting approval of its own residents. Fourth, this research considers not the year term-limits are proposed in the form of a council ordinance/resolution or voters' initiative (Government Code Section 36502), but the year of their popular approval through the ballot. Either form must go through the ballot to become binding. Term-limits can be installed in the municipal code of a general-law or charter city (Government Code Section 36502), or in the charter of a charter city (Elections Code Section 9255). 45 California State University at Sacramento (Institute for Social Research, Center for California Studies) 1995-2011, http://www.csus.edu/isr/reports/california_elections, accessed April 07, 2013. 46 League of Women Voters of California Education Fund, http://www.smartvoter.org, accessed April 07, 2013. 47 The Lucy Burns Institute, http://ballotpedia.org, accessed April 07, 2013. 48 Here, the term-limits of the city of Palmdale repealed in 2009 are included. 90 Council-manager. The data for this measure is obtained from the Municipal Year Book, published by the International City/County Management Association (ICMA). Details are found in the preceding Chapter 2. Charter. The data source for this variable is the Annual Cities Report, published by the California State Controller. Its data characteristics are described in Chapter 2. Controls. The present study adds a total of six controls: three socioeconomic indicators (city population, population density, and unemployment rate) and three streams of intergovernmental revenue from federal, state, and county governments (per capita, $). It should be noted that the empirical model of this study is constructed explicitly with that of Besley and Case (1995) in mind where income, young and elderly portions of population, and total population are included at the state level (Besley and Case, 1995, p.777). Among these, only the data on total population is published annually at the city level. Thus, each of population density (persons per square mile) and unemployment rate (annual, %) is included as a proxy for the portions of population and income, respectively. Population density accounts for service needs like the population portions and unemployment rate does for a city's general economy like income. Three streams of intergovernmental revenue are added, given the substantive reliance of cities on the upper-level governments. These controls have already been used in the existing literature (e.g., Erler, 2007, p.484; Gordon, 2009, p.45). The data for city population (persons) is acquired from RAND California (originating source: U.S. Census Bureau). The immediate data source for population density is also RAND California (originating source: U.S. Census Bureau). The data for unemployment rate is available on cities with populations above 25,000 from the Bureau of Labor Statistics. Lastly, the data for 91 the three flows of intergovernmental revenue are obtained from RAND California (originating source: Cities Financial Transactions Report). All monetary data are converted to 2005 constant dollars to correct for inflation. All independent variables and controls except council-manager are lagged by t-1 year. TABLE 4-1 specifies how each of these variables is operationalized with a source of data. TABLE 4-2 displays their respective descriptive statistics. TABLE 4-1: Variables, Descriptions, and Sources of Data Variables (unit) Descriptions Sources Total expenditure ($) Per capita; 2005 constant dollar <California City Financial Transactions>, RAND California 49 ; originating source-Cities Financial Transactions Report Total revenue ($) Per capita; 2005 constant dollar <California City Financial Transactions>, RAND California; originating source-Cities Financial Transactions Report Revenue from taxes ($) Per capita; 2005 constant dollar <California City Financial Transactions>, RAND California; originating source-Cities Financial Transactions Report Revenue from fees and charges ($) Per capita; 2005 constant dollar <California City Financial Transactions>, RAND California; originating source-Cities Financial Transactions Report Unreserved fund balance ($) Per capita; 2005 constant dollar; Unreserved fund balance (UFB) = Designated + Undesignated UFB <California City Financial Transactions>, RAND California; originating source-Cities Financial Transactions Report Term-limits (0, 1) t-1 lagged; 1, if a city has term-limits on its city council; 0, otherwise. The City of Thousand Oaks, April 12, 2011, "State of California Councilmember Term Limits Survey by City in Alphabetical Order," as Attachment #4 of the Memorandum, Thousand Oaks Councilmembers Term Limits Initiative Impact Analysis Report & Scheduling for November 6, 2012 Election; municipal code of ordinances; charter; county registrar of voters; California Elections Data Archive (CEDA); Smart Voter; Ballotpedia; newspapers (esp., the Los Angeles Times); Petracca and O'Brien, 1999; and/or email inquires with city clerks Council-manager (0, 1) Not lagged; 1, if council-manager; 0, if mayor-council. ICMA, Municipal Year Book Charter (0, 1) t-1 lagged; 1, if chartered; 0, if general- law. California State Controller, Annual Cities Report City population (persons) t-1 lagged. <Population and Demographic statistics>, RAND California 50 ; originating source-the U.S. Census Bureau 49 http://ca.rand.org/stats/govtfin/cityfinance.html 92 Population density (persons per square mile) t-1 lagged. <Community statistics>, RAND California 51 ; originating source-the U.S. Census Bureau Unemployment rate (%) t-1 lagged; annual rate; information available on cities above 25,000 populations; not seasonally adjusted <Local Area Unemployment Statistics>, Bureau of Labor Statistics 52 Intergovernmenta l revenue from Federal ($) t-1 lagged; per capita; 2005 constant dollar < California City Financial Transactions>, RAND California; originating source-Cities Financial Transactions Report Intergovernmenta l revenue from State ($) t-1 lagged; per capita; 2005 constant dollar < California City Financial Transactions>, RAND California; originating source-Cities Financial Transactions Report Intergovernmenta l revenue from County ($) t-1 lagged; per capita; 2005 constant dollar < California City Financial Transactions>, RAND California; originating source-Cities Financial Transactions Report TABLE 4-2: Descriptive Statistics Variables (unit) Observations Mean Standard Deviation Min. Max. Total revenue (per capita, $) 4,468 1,207.08 900.31 87.28 8,497.18 Total expenditure (per capita, $) 4,468 1,160.47 877.36 77.63 8,367.73 Revenue from taxes (per capita, $) 4,468 444.35 295.49 39.84 3,707.36 Revenue from fees and charges (per capita, $) 4,466 394.93 493.03 2.04 4,714.73 Unreserved fund balance (per capita, $) 4,468 464.02 459.96 -753.28 5,391.71 Term-limits (0, 1) 4,473 .19 .39 0 1 Council-manager (0, 1) 4,508 .96 .17 0 1 Charter (0, 1) 4,535 .29 .45 0 1 City population (persons) 4,484 100,712.5 255,421.1 11,419 3,801,574 Population density (persons per square mile) 4,464 5,499.12 4,395.81 275 78,777 Unemployment rate (%) 4,545 6.36 3.85 1.01 38.69 Intergovernmental revenue from federal (per capita, $) 4,429 46.08 67.77 0 1,191.93 Intergovernmental revenue from state (per capita, $) 4,467 97.02 64.39 13.17 1,083.40 Intergovernmental revenue from county (per capita, $) 4,420 7.68 21.43 0 528.67 50 http://ca.rand.org/stats/popdemo/popdemo.html 51 http://ca.rand.org/stats/community/community.html 52 http://www.bls.gov/data/ 93 Models The current study employs the estimation technique of Besley and Case (1995) as a benchmark: fixed-effects estimation including year effect (p.773). Yet, unlike theirs which utilizes Huber standard errors (p.780), i.e., heteroskedasticity-robust standard errors (Cameron and Trivedi, 2010, p.334), this study utilizes cluster-robust standard errors, which are known to be effective in mitigating the problems of heteroskedasticity and autocorrelation (Cameron and Trivedi, 2010, p.85). 4.RESULTS TABLE 4-3 shows the estimation results. Within R-squared values are typically higher than Between R-squared ones in their respective models. Given that the three institutional variables tend to have the property of perpetuity after their installment, each of their effects appears to be identified mainly from the difference between before and after the installation of each institution within a city rather than from the cross-sectional difference of each institution between cities at a time t (Besley and Case, 1995, p.778). Term-limits do not show any statistically significant coefficients across all the five models. Council-manager shows statistically significant negative coefficients in Models 1, 3, and 4. A governmental transformation from mayor-council to council-manager typically brings down total expenditure per capita by about 58 dollars (Model 1), revenue from taxes per capita by about 36 dollars (Model 3), and revenue from fees and charges per capita by about 62 dollars (Model 4). This negative direction is as expected in Hypothesis 1 in general and Hypothesis 1-1), Hypothesis 1-3), and Hypothesis 1-4) specifically. Charter displays positive coefficients with statistical significance in Models 1 and 5. Once a government becomes chartered, its level of per-capita spending is likely to increase by 94 around 109 dollars (Model 1) and the stock of unreserved fund balance per capita typically goes up by approximately 171 dollars (Model 5). Its positive direction in Model 5 is consistent with Hypothesis 2 in general and Hypothesis 2-5) in particular, while such a direction in Model 1 is unexpected from Hypothesis 2 and specifically Hypothesis 2-1). Among controls, city population is not able to reach statistical significance in any models. Population density turns out to be statistically significant in Models 1, 2, and 5. Unemployment rate turns out to be significant in Models 1 and 3 where it is positively associated with the respective dependent variables, i.e., total expenditure and revenue from taxes. All three intergovernmental revenues have a positive relationship with total expenditure (Model 1) and total revenue (Model 2) with statistical significance in common. 95 TABLE 4-3: Fixed Effects (Within) Regression Coefficients Model1 Model2 Model3 Model4 Model5 Total expenditure (per capita, $) Total revenue (per capita, $) Revenue from taxes (per capita, $) Revenue from fees and charges (per capita, $) Unreserved fund balance (per capita, $) Term-limits t-1 -43.127 (36.243) -31.513 (38.136) 11.295 (18.576) -20.604 (15.436) 55.513 (49.069) Council-manager t -57.918* (33.357) -56.036 (50.223) -36.337** (16.456) -61.967*** (21.773) -16.511 (45.738) Charter t-1 108.573* (59.803) 100.903 (64.307) 13.466 (29.365) 32.152 (30.600) 170.805** (81.853) City population t-1 (persons) 0.000 (0.001) 0.000 (0.001) -0.000 (0.000) -0.001 (0.000) -0.001 (0.001) Population density t-1 (persons per square mile) -0.005** (0.002) -0.009*** (0.002) -0.001 (0.001) -0.002 (0.001) -0.005*** (0.001) Unemployment rate t-1 (%) 15.398** (7.676) 8.007 (7.485) 8.599*** (1.805) 0.524 (4.622) 9.363 (6.881) Intergovernmental revenue from Federal t-1 (per capita, $) 0.694*** (0.259) 0.648*** (0.230) 0.152* (0.085) 0.042 (0.132) 0.113 (0.192) Intergovernmental revenue from State t-1 (per capita, $) 0.578*** (0.194) 0.909*** (0.215) 0.086 (0.084) 0.217*** (0.065) 0.075 (0.382) Intergovernmental revenue from County t-1 (per capita, $) 0.856** (0.430) 0.930*** (0.296) 0.254 (0.197) 0.037 (0.140) -0.138 (0.286) Constant 766.455*** (117.846) 842.203*** (112.702) 323.874*** (37.542) 405.695*** (61.678) 235.983* (130.224) R-sq (within), % 37.1% 33.4% 56.3% 16.3% 18.5% R-sq (between), % 21.1% 30.7% 1.7% 2.7% 1.2% R-sq (overall), % 21.1% 22.1% 1.7% 1.2% 5.5% P value of F test for model fit .000 .000 .000 .000 .000 Number of observations 4,276 4,276 4,276 4,274 4,276 Number of cities 254 254 254 254 254 Time scope (dependent variable) 1992-2009 1992-2009 1992-2009 1992-2009 1992-2009 Note: cluster-robust standard errors are noted in parentheses. Year dummies are included. Monetary data are converted to 2005 constant dollars. Probability levels are denoted as follows: *** p<.01, ** p<.05, * p<.10. 5. DISCUSSION The above results show that term-limits have no meaningful association with fiscal outcomes. While the present study is not able to present a conclusive explanation to this, two 96 possibilities may deserve consideration. First, it is conceivable that both positive and negative effects did exist but were offset in the estimation. That is, it may be true that term-limits in a city led to decent fiscal performance because of possibly increased competitiveness (Daniel and Lott, 1997, p.168, 169) and the possible courage provoked by the diminishing worth of a council membership (Schelker, 2012, p.31-32). At the same time, it may equally be true that term-limits resulted in poor fiscal performance because term-limits strip incumbents of the prospect of reelection as a powerful discipline mechanism (Alt, Bueno de Mesquita and Rose, 2011, p.171). The second possibility is inspired by the view of term-limits as a product of pork-barrel politics, not one of benevolent care for fiscal performance (Buchanan and Congleton, 1994; Friedman and Wittman, 1995). According to this view, politicians compete with each other to win a special-benefits project to their respective districts (Buchanan and Congleton, 1994, p.48). A lawmaker with higher seniority is known to be more effective than one with lower seniority in winning such a project (Friedman and Wittman, 1995, p.70). Friedman and Wittman (1995) argue that those constituents in a district are likely aware of this seniority advantage and therefore endorse term-limits only if the lawmaker whom they elected is junior in term-length (p.71). They can expect greater special rewards from their own lawmaker with comparative seniority caused by the effect of term-limits in removing seniors (Friedman and Wittman, 1995, p.71). By the same token, it is not likely that they support term-limits if they already have a senior lawmaker representing their district (Friedman and Wittman, 1995, p.71). The authors conclude, "term limits are not imposed to make the legislature or executive work 'better' but to redistribute power......" (Friedman and Wittman, 1995, p.81). This particular view is enlightening to the present study, since it underscores the sheer political motivations around term-limits. Unlike the state- or federal- level lawmakers and term- 97 limits on which this view typically focuses, however, city council members are predominantly elected at large, not by district in California (Hajnal, Lewis, and Louch, 2002, p.25). Thus, the politics around local term-limits would not involve a special project for a district. The likely characteristic of local politics around term-limits would be the political competition between sitting council members and interested contenders outside the council to obtain a council membership itself. Term-limits may be introduced as a result of this political pursuit. Yet, first of all, is a local council membership really worth pursuing? The answer of the present study is a yes, especially given the fact that the job is designed to be part-time in most cities. The average level of regular salaries for council members of California cities may not be very high but the salaries tend to come with additional rewards such as health benefits in over two-thirds of all the cities in California 53 and/or other benefits (e.g., payment for participating in government commissions/boards, auto allowance, or business-cost allowance) in some cities 54 . Together with the material dimension of compensation, the job earns a resident the honor of representing the whole city. Even more, it is an excellent opportunity to practice leadership skills, appealing especially to those young people who plan to develop a leadership career in politics or government on a city, state, or even national level. Local term-limits are likely to be advanced by the contenders outside the council who have interest in the membership. Obviously, incumbents would not support term-limits, because they would become the immediate victims of those limits. Term-limits are highly advantageous to outside contenders, because this institution will depose incumbents and create new vacancies. During the campaign for public approval, the outsiders may claim that term-limits are conducive to good fiscal performance, while the incumbents may claim that such limits are detrimental to it. 53 Catherine Saillant. "Many small cities pay officials hefty health benefits," Los Angeles Times, July 04, 2011. 54 Abby Sewell. "City councils cautious on raising their pay," Los Angeles Times, August 02, 2011. 98 Due to the utterly political motivations concerning term-limits, however, it is difficult to believe that this institution is associated with fiscal performance in any way. As for government form, the empirical results provided support for the beneficial role of council-manager over mayor-council. Less spending, less tax revenue, and less revenue from fees and charges, all of which are preferred by average voters, were observed in the council- manager form to a greater extent than in the mayor-council form. This is consistent with what the literature implies; the council-manager form offers a structural advantage in the effort of city administration to loyally carry out public preferences compared to the mayor-council form which makes the administration vulnerable to the distortional force of city politics (Feiock and Jang, 2009, p.671). Charter status was expected as another institution to honor what average voters want (Benton, 2002, p.474). The empirical results suggest that the impact direction of establishing a charter is less straightforward. In other words, while it did lead to a greater amount of savings (unreserved fund balance), it also led to more spending, not less. It is worth recalling that one rationale behind the charter grant by voters is to create more flexibility for their own government, so that it can become more responsive to their needs (Benton, 2002, p.474). Then, one possibility would be that the flexibility may have been taken advantage of by local political officials, defeating the rationale. It is interesting to note that council members are typically paid more in charter governments than in general-law governments, according to the Los Angeles Times 55 . The positive relationship of city unemployment rates with total expenditures and also tax revenues warrants a fair amount of explanation. Unemployment rate is used as a proxy for income and the general economy. The rise in tax revenue can be viewed as a local effort to secure revenue in face of a worsening economy (Hou and Moynihan, 2008, p.145). In such a 55 Abby Sewell. "City councils cautious on raising their pay," Los Angeles Times, August 02, 2011. 99 case, which specific tax was more likely to be raised? Personal income is not taxed by cities in California 56 (California State Controller, 2005, p.81-83) and property tax is restrained by the well-known Proposition 13 passed in 1978 (Hoene, 2004, p.63). Accordingly, the tax increase may have occurred via another taxing option such as special taxes, sales and use taxes, transient occupancy taxes, business license taxes, and/or utility users taxes. 57 The observed rise in total spending is not surprising as the collected tax revenue is going to be expended in the end. 6. CONCLUSION Term-limits did not have any positive or negative fiscal consequences. This could be either because positive and negative effects neutralized each other or because fiscal performance is not something that advocates or opponents actually care about from the beginning, but rather a rhetorical claim to influence support for this institution (Friedman and Wittman, 1995, p.81). If the former is the case, there is a need for a more refined measure of term-limits to distinguish the positive from the negative impact. If the latter is the case, term-limits may not be a substantive fiscal institution any more. Yet the result makes it possible to develop a simple theory concerning political competition involving term-limits toward a council membership between an incumbent council member and a challenger outside the council. Term-limits are strategically advanced by the latter. The incumbent makes every effort to defeat the passage of term-limits. Term-limits will be approved at ballot when the challenger becomes stronger than the incumbent. Surely, the validity of this theory will have to be proven by subsequent empirical research. The present study adds local term-limits as a new viable focus of research beyond the traditional state-level focus. 56 Institute of Local Government. 2008. Understanding the Basics of County and City Revenues. Retrieved from www.ca- ilg.org/revenueguide, October 03, 2011. 57 Institute of Local Government. 2008. Understanding the Basics of County and City Revenues. Retrieved from www.ca- ilg.org/revenueguide, October 03, 2011. 100 It has been demonstrated that the council-manager form of government is associated with better fiscal performance than the mayor-council form. The relative deactivation of politics under the former (Feiock and Jang, 2009, p.671) typically makes it a more trustworthy form of government than the latter in executing public preferences. The importance of a charter as a fiscal institution was supported as well. This works in a mixed way; it honors a public preference (more savings) but also betrays one (more spending). It implies that the greater scope of policy opportunities allowed for charter than general-law governments (Benton, 2002, p.474) may not always work for residents. This reinforces the need for further empirical studies on the municipal charter. Unemployment rate, which accounted for the overall economy, was associated with higher tax revenue. A sluggish economy makes city governments respond by ensuring revenue in the subsequent year (Hou and Moynihan, 2008, p.145) - for example, by applying higher tax rates or enlarging its own tax base. 101 APPENDIX 4-1: Correlations 1.Tot al expe nditu re 2.T otal rev enu e 3.Re venu e from taxes 4.Rev enue from fees& charge s 5.Unr eserve d fund balanc e 6.T erm - limi ts t-1 7.Co uncil - man ager 8.C hart er t-1 9.Cit y popu latio n t-1 10.Pop ulation densit y t-1 11.Unem ploymen t rate t-1 12.Intergo vernmenta l revenue from Federal t-1 13.Intergo vernmenta l revenue from State t-1 14.Intergo vernmenta l revenue from County t-1 1 1.00 2 .96* 1.0 0 3 .76* .76 * 1.00 4 .86* .87 * .46* 1.00 5 .29* .35 * .40* .22* 1.00 6 .22* .22 * .24* .18* .15* 1.0 0 7 -.18* - .16 * -.11* -.11* .04* - .04 * 1.00 8 .39* .38 * .31* .34* .06* .46 * -.09* 1.00 9 .23* .21 * .12* .20* -.05* .21 * -.41* .25* 1.00 1 0 .02 .00 .03* -.04* -.07* .08 * -.16* .01 .06* 1.00 1 1 -.10* - .12 * -.26* -.01 -.24* - .18 * -.03* .00 .01 .11* 1.00 1 2 .49* .47 * .33* .36* -.03* .10 * -.19* .29* .20* .19* .08* 1.00 1 3 .44* .42 * .36* .26* .00 .11 * -.24* .15* .12* .08* -.00 .43* 1.00 1 4 .14* .13 * .14* .05* .00 .09 * -.07* .05* .03* .05* -.08* .14* .21* 1.00 Note: * significant at 10% level. 102 APPENDIX 4-2: Approval Year of Term-limits on Council a No. City Approval Year No. City Approval Year 1 Alameda 1977 48 Menifee 2010 2 Albany 1981 49 Merced 1968 3 Alhambra 1976 50 Millbrae 1997 4 Anaheim 1992 51 Milpitas 1996 5 Arcadia 1968 52 Mission Viejo 1998 6 Buena Park 1996 53 Modesto 2003 7 Campbell 2000 54 Monte Sereno 1998 8 Capitola 2002 55 Mountain View 1975 9 Cerritos 1986 56 Murrieta 2010 10 Chula Vista 1973 57 Newport Beach 1992 11 Commerce 2011 58 Orange 1996 12 Coronado 2002 59 Pacific Grove 1990 13 Costa Mesa 1996 60 Pacifica 2010 14 Culver City 1994 61 Palmdale 2001 approved; 2009 repealed 15 Cupertino 1997 62 Palo Alto 1991 16 Cypress 1977 63 Piedmont Unknown b 17 Dana Point 1996 64 Pinole 2008 18 Downey 1993 65 Placentia 2002 19 Dublin 1996 66 Pleasanton 1996 20 Eureka Unknown b 67 Rancho Palos Verdes 2003 21 Foster City 1996 68 Redondo Beach 1975 22 Fountain Valley 2004 69 Redwood City 1968 23 Fremont 1996 70 Roseville 1974 24 Fresno Unknown b 71 San Diego 1992 25 Fullerton 2010 72 San Francisco 1990 26 Garden Grove 1996 73 San Jose 1990 27 Gardena 1999 74 San Leandro 1974 28 Grover Beach 1996 75 San Luis Obispo 1978 29 Hemet 2010 76 San Marcos 1998 30 Huntington Beach 1978 77 San Mateo 1983 31 Huntington Park 2003 78 Santa Ana 1966 32 Indian Wells 2010 79 Santa Barbara 1990 33 Irvine 1986 80 Santa Clara Unknown b 34 Kerman 1996 81 Santa Cruz Unknown b 35 La Palma 1996 82 Seal Beach 1974 36 Laguna Hills 2010 83 Stockton 2000 37 Laguna Niguel 1996 84 Sunnyvale 1975 38 Livermore 2003 85 Temple City 1992 39 Long Beach 1992 86 Thousand Oaks 2012 40 Loomis 2010 87 Torrance 1992 41 Los Alamitos 1998 88 Tracy 2008 42 Los Altos 1999 89 Tustin 1994 popularly approved; 1997 confirmed by the council. 43 Los Altos Hills 1998 90 Union City 1996 44 Los Angeles 1993 91 Vallejo 1977 45 Lynwood 2008 92 Villa Park 1998 46 Malibu 2000 93 Watsonville 1989 47 Manhattan Beach 1996 94 Yorba Linda 1996 Source: see TABLE 4-1. Note: a. The year of the popular approval of the initial term-limits on city council members. b. Unknown means that a city currently has term-limits but their approval year is not able to be determined. 103 CHAPTER 5. CONCLUSION 1. FINDINGS SUMMARY The highly stimulating findings are summarized as follows. Chapter 2 The importance of macro-economic and institutional forces was clearly supported. In the simultaneous absence of the council-manager and the charter, recession was associated with a decrease in contracting. When one of the two was present, the decreasing magnitude was moderated to an extent. Their concurrent presence made the inverse recessionary association eventually turn positive. Namely, recession turned out to be related with more contracting only in chartered council-manager governments. These findings indicate that our expectation of greater contracting under recession is actually based on an implicit assumption that governments would have a suitable institutional arrangement capable of carrying out what their citizens demand. Chapter 3 Albeit weakly, the findings showed that non-collective services were the direct victim of an aggravating fiscal environment. The reason for this is that the provision of non-collective services does not constitute a core government responsibility (e.g., Samuelson, 1954). Yet further research is required concerning why the aggravating internal fiscal condition did not decrease but increased the non-collective service provision. Higher intergovernmental dependence made governments content with the current state of affairs and passive in pursuing additional revenue from non-collective services. Lastly, whether intergovernmental revenue was received as general or functional was significant; general intergovernmental revenue consistently displayed a larger negative impact on the non-collective service provision than functional revenue. This is because the former typically corrodes fiscal discipline more than the latter. 104 Chapter 4 Term-limits showed insignificant impacts on fiscal performance. The potentially positive and negative effects of term-limits may have canceled each other out in each of the estimations. Alternatively, term-limits may have come out as a product of sheer political competition to grab a council membership, not as a benevolent human invention that cares about fiscal performance (refer to Friedman and Wittman, 1995, p.81). The positive relationship between the council- manager and fiscal performance was consistent with the argument formulated at Chapter 2. The effects of the charter status were less straightforward over fiscal performance. The macro-economy, measured by a city's unemployment rate, turned out to matter: higher unemployment rates led to higher tax revenues and therefore higher total spending. As for taxes, the raised item may not have been the property tax, due to the restriction put in place by Proposition 13 in 1978 (Hoene, 2004, p.63), but other taxes such as sales and use taxes, or business license taxes. 58 2. THEORETICAL CONTRIBUTIONS This dissertation project intends to contribute to the research on public budgeting and finance specifically and that of public administration/management/policy broadly. Five potential contributions may be noted. First, this project enhances our understanding of how the external economic/fiscal environment affects local governments and in turn how they respond to it. Each empirical chapter has an indicator to account for the environment: nationwide recession in Chapter 2, statewide state surplus/deficit in Chapter 3, and city-specific unemployment rate in Chapter 4. Local governments tend to increase contracting under a certain institutional 58 Institute of Local Government. 2008. Understanding the Basics of County and City Revenues. Retrieved from www.ca- ilg.org/revenueguide, October 03, 2011. 105 arrangement (Chapter 2), decrease non-collective service provision (Chapter 3), and raise extra tax revenues (Chapter 4) in face of a poor external economic/fiscal environment. Second, this project is an effort to connect the scholarship of general public administration/management/policy to that of public budgeting and finance. While the two have great potential to benefit from each other, they have developed independently (Kioko et al., 2011). To take an example, contracting as a research theme is traditionally popular in the former scholarship, but the research is somewhat limited in terms of data. On the other hand, the theme has low presence in the latter. This project contributes to the former scholarship by using continuous fiscal data with a time dimension and also to the latter by underscoring the value of contracting as a financial means to save service production costs (see Stein, 1990, p.84). Another case in point is the conceptual typology of collective vs. non-collective goods, which is a well-known idea justifying government existence in both literatures. Nonetheless, it is rarely employed for empirical research. This project contributes to both literatures by providing empirical supports showing the importance of the typology. Third, the use of double interaction in Chapter 2 offers an opportunity to measure the effect of combined institutional arrangements (e.g., the council-manager and the charter at once), moving beyond the measurement of one institutional arrangement at a time (e.g., the council- manager). This interaction approach better describes reality than does the traditional approach, since government in reality is characterized not by a single institution but by the concurrent presence of multiple institutions. Fourth, this project widens the horizon of the existing research on term-limits. Although term-limits are a highly popular theme in research at the U.S. state level (e.g. Besley and Case, 1995), local term-limits are not often examined, primarily due to a lack of data. This project is 106 one of the first attempts to analyze them at the local level by constructing an original longitudinal dataset. Lastly, this project has analyzed how several key public institutions affect government behavior: council-manager (Chapters 2, 3, 4), charter (Chapters 2, 4), term-limits (Chapter 4), and two types of intergovernmental transfers (Chapter 3). Most of them proved their relevance in the respective chapters. These findings contribute to the existing scholarship. 3. POLICY IMPLICATIONS This dissertation project is designed to be a study that is relevant to the practice of local government. Local governments are the infrastructure of the political and administrative system in a country. There is no doubt about the desirability of keeping them healthy and prosperous. Although government often appears too big and too powerful in the public eye, local governments are in fact challenged by numerous forces, including unfavorable economic/fiscal environments, inappropriate institutional arrangements, or high intergovernmental dependence. Two immediate practical implications from this study are illustrated below. Whether to adopt a new institution Adopting a new institution is a major policy tool by which residents in a city can modify their government behavior in a preferred way. Yet it has to be the right institution that is adopted. It takes a considerable amount of time to find out whether a new institution will bring about the intended impact. Chapter 2 demonstrated that the right institutional arrangement may counteract the negative effect of an environmental shock. On the other hand, Chapter 4 failed to show any significant impact of term-limits on fiscal performance. This suggests that the claimed effect of term-limits is somewhat exaggerated. In general, residents need to be able to discern whether an argument for or against a new institution is sound or not. 107 High intergovernmental dependence When a system of intergovernmental relations calls for reform, a difficult issue would be how much fiscal (in)dependence is appropriate. Chapter 3 found that higher dependence made local governments reduce the provision of non-collective services. It does not offer the incentive for them to turn active and creatively seek extra fiscal opportunities from non-collective service provision. 4. LIMITATIONS AND FUTURE RESEARCH This dissertation research was conducted and discussed in the context of California. Accordingly, it is not statistically possible to generalize the findings beyond California to the rest of the U.S. states or to other countries. In particular, the findings are not directly transferrable to other countries, because their municipalities operate in highly different social, political, and cultural environments. To take an example, municipalities in the Republic of Korea have a very short history of self-government, starting only approximately two decades ago. Although municipal council members have been elected by popular vote since, the Korean system of intergovernmental relations is still highly centralized, especially when compared to U.S. federalism. These limitations call for subsequent research in different national contexts. Nevertheless, this research project offers useful information and insights for other countries. Municipalities in California and those in other countries are often subject to such common forces as economic/fiscal environments, institutional constraints, and intergovernmental dependence. In response, they would utilize contracting as a financial tool and adjust the provision level of (non-)collective services. For instance, one heated issue in Korean public administration has been whether to decentralize the intergovernmental system. The discussion on 108 intergovernmental dependence in Chapter 3 may be a good source of information and provide relevant insights for this issue. 109 REFERENCES Alt, James, Ethan Bueno de Mesquita, and Shanna Rose. 2011. "Disentangling Accountability and Competence in Elections: Evidence from U.S. Term Limits," The Journal of Politics, 73(01), 171-186. Baker, S. H. 2003. "The Tax Revolt and Electoral Competition," Public Choice, 115(3/4), 333- 345. Bel, G. and X. Fageda. 2009. "Factors explaining local privatization: a meta-regression analysis," Public Choice, 139(1), 105-119. Benton, J. E. 2002. "County Service Delivery: Does Government Structure Matter?" Public Administration Review, 62(4), 471-479. Benton, J. E. 2003. "County Government Structure and County Revenue Policy: What's the Connection?" State and Local Government Review, 35(2), 78-89. Besley, T. and A. Case. 1995. "Does Electoral Accountability Affect Economic Policy Choices? Evidence from Gubernatorial Term Limits," The Quarterly Journal of Economics, 110(3), 769-798. Besley, T. and A. Case. 2003. "Political Institutions and Policy Choices: Evidence from the United States," Journal of Economic Literature, 41(1), 7-73. Brambor, Thomas, William Roberts Clark, and Matt Golder. 2006. "Understanding Interaction Models: Improving Empirical Analyses," Political Analysis, 14(1), 63-82. Brown, E. C. 2010. "The Bidding Process," California Infrastructure Projects: A Guide to Successful Contracting and Dispute Resolution, Chapter 12. Ernest Brown & Company, LLC. Retrieved from http://constructionlawyers.com/CM/Custom/Book.htm, 06/19/2012. 110 Brown, T. L. and M. Potoski. 2003. "Transaction Costs and Institutional Explanations for Government Service Production Decisions," Journal of Public Administration Research and Theory, 13(4), 441-468. Brudney, J. L., S. Fernandez, J. E. Ryu, and D. S.Wright. 2005. "Exploring and Explaining Contracting Out: Patterns among the American States," Journal of Public Administration Research and Theory, 15(3), 393-41. Buchanan, J. M. 1965. "An Economic Theory of Clubs," Economica, 32(125), 1-14. Buchanan, J. and R. Congleton. 1994. "The incumbency dilemma and rent extraction by legislators," Public Choice, 79(1-2), 47-60. Bushouse, B. K. 2011. "Governance Structures: Using IAD to Understand Variation in Service Delivery for Club Goods with Information Asymmetry," Policy Studies Journal, 39(1), 105-119. California State Controller. 2005. Cities Financial Transactions Report Instructions. Retrieved from <http://www.sco.ca.gov/ard_locinstr_cities_forms.html>, May 14 th 2011. Cameron, A. Colin and Pravin K. Trivedi. 2010. Microeconometrics Using Stata, revised edition. College Station, Texas: Stata Press. Clark, T. N. and L. C. Ferguson. 1983. City Money: Political Processes, Fiscal Strain, and Retrenchment. New York: Columbia University Press. Cornes, R. and T. Sandler. 1996. The Theory of Externalities, Public Goods and Club Goods. New York, NY: Cambridge University Press. Crain, W. M. and L. K. Oakley. 1995. "The Politics of Infrastructure," Journal of Law and Economics, 38(1), 1-17. 111 Crain, W. M. and R. D. Tollison. 1993. "Time inconsistency and fiscal policy: Empirical analysis of U.S. States, 1969-89," Journal of Public Economics, 51(2), 153-159. Daniel, K. and J. J. R. Lott. 1997. "Term limits and electoral competitiveness: Evidence from California's state legislative races," Public Choice, 90, 165-184. Duggan, M. G. 2000. "Hospital Ownership and Public Medical Spending," The Quarterly Journal of Economics, 115(4), 1343-1373. Dunn, S., J. M. Quigley, and L. A. Rosenthal. 2005. "The Effects of Prevailing Wage Requirements on the Cost of Low-Income Housing," Industrial and Labor Relations Review, 59(1), 141-157. Erler, H. A. 2007. "Legislative term limits and state spending," Public Choice, 133(3-4), 479-494. Feiock, R. C. and J.-H. Kim. 2001. "Form of Government, Administrative Organization, and Local Economic Development Policy," Journal of Public Administration Research and Theory, 11(1), 29-50. Feiock, Richard C. and Hee Soun Jang. 2009. "Nonprofits as Local Government Service Contractors," Public Administration Review, 69(4), 668-680. Feiock, R. C., M.-G. Jeong, and J. Kim. 2003. "Credible Commitment and Council-Manager Government: Implications for Policy Instrument Choices," Public Administration Review, 63(5), 616-625. Ferris, J. M. and E. Graddy. 1994. "Organizational Choices for Public Service Supply," Journal of Law, Economics, & Organization, 10(1), 126-141. Frant, H. 1996. "High-Powered and Low-Powered Incentives in the Public Sector," Journal of Public Administration Research and Theory, 6(3), 365-381. 112 Friedman, D. and D. Wittman. 1995. "Why voters vote for incumbents but against incumbency: A rational choice explanation," Journal of Public Economics, 57(1), 67-83. Gordon, T. 2009. "Bargaining in the shadow of the ballot box: causes and consequences of local voter initiatives," Public Choice, 141(1-2), 31-48. Hajnal, Zoltan L., Paul G. Lewis, and Hugh Louch. 2002. Municipal elections in California: turnout, timing, and competition. San Francisco, CA: Public Policy Institute of California. Hardin, G. 1968. "The Tragedy of the Commons," Science, 162(3859), 1243-1248. Hefetz, A. and M. Warner. 2007. "Beyond the market versus planning dichotomy: Understanding privatisation and its reverse in US cities," Local Government Studies, 33, 555-572. Hendrick, R. 2006. "The Role of Slack in Local Government Finances," Public Budgeting & Finance, 26(1), 14-46. Hirsch, W. Z. 1995. "Factors Important in Local Governments' Privatization Decisions," Urban Affairs Review, 31(2), 226-243. Hoene, C. 2004. "Fiscal Structure and the Post-Proposition 13 Fiscal Regime in California's Cities," Public Budgeting & Finance, 24(4), 51-72. Holian, M. J. 2009. "Outsourcing in US cities, ambulances and elderly voters," Public Choice, 141(3), 421-445. Hou, Y. and D. P. Moynihan. 2008. "The Case for Countercyclical Fiscal Capacity," Journal of Public Administration Research and Theory, 18(1), 139-159. Hsiao, C. 2003. Analysis of Panel Data. West Nyack, NY: Cambridge University Press. Hyman, D. N. 2002. Public Finance, A Contemporary Application of Theory to Policy. Mason, Ohio: South-Western. 113 ICMA, International City Management Association. 2005. The Municipal Year Book 2005. Washington, D.C.: ICMA. Jin, Hehui, Yingyi Qian and Barry R. Weingast. 2005. "Regional decentralization and fiscal incentives: Federalism, Chinese style," Journal of Public Economics, 89, 1719-1742. Karuppusamy, S. and J. B. Carr. 2012. "Interjurisdictional Competition and Local Public Finance: Assessing the Modifying Effects of Institutional Incentives and Fiscal Constraints," Urban Studies, 49(7), 1549-1569. Kettl, D. F. 2000. "The Transformation of Governance: Globalization, Devolution, and the Role of Government," Public Administration Review, 60(6), 488-497. Kioko, S. N., J. Marlowe, D. S. T. Matkin, M. Moody, D. L. Smith, and Z. J. Zhao. 2011. "Why Public Financial Management Matters," Journal of Public Administration Research and Theory, 21(suppl. 1), i113-i124. Kornai, János. 1998. "The Place of the Soft Budget Constraint Syndrome in Economic Theory," Journal of Comparative Economics, 26(1), 11-17. Kornai, János, Eric Maskin, and Gérard Roland. 2003. "Understanding the Soft Budget Constraint," Journal of Economic Literature, 41(4), 1095-1136. Lamothe, S., M. Lamothe, and R. C. Feiock. 2008. "Examining Local Government Service Delivery Arrangements Over Time," Urban Affairs Review, 44(1), 27-56. Levin, Jonathan and Steven Tadelis. 2010. "Contracting for Government Services: Theory and Evidence from U.S. Cities," Journal of Industrial Economics, 58(3), 507-541. Levine, Charles H., Irene S. Rubin, and George G. Wolohojian. 1981. The Politics of Retrenchment: How Local Governments Manage Fiscal Stress. Beverly Hills, CA: Sage Publications. 114 MacDonald, L. 2008. "The impact of government structure on local public expenditures," Public Choice, 136(3), 457-473. Marlowe, J. 2005. "Fiscal Slack and Counter-Cyclical Expenditure Stabilization: A First Look at the Local Level," Public Budgeting & Finance, 25(3), 48-72. Marlowe, J. 2006. "Fund Balance, Working Capital, and Net Assets," in H. A. Frank (ed.), Public Financial Management, pp.357-381. New York: CRC Press. McCubbins, M. D. and E. Moule. 2010. "Making Mountains of Debt out of Molehills: The Pro- cyclical Implications of Tax and Expenditure Limitations," National Tax Journal, 63(3), 603-622. Moe, T. M. 1989. "The Politics of Bureaucratic Structure," in J. E. Chubb and P. E. Peterson (eds.), Can the Government Govern? Washington, D.C.: Brookings. Moesen, W. and P. van Cauwenberge. 2000. "The Status of the Budget Constraint, Federalism and the Relative Size of Government: A Bureaucracy Approach," Public Choice, 104(3), 207-224. Mullins, D. R. and M. A. Pagano. 2005. "Local Budgeting and Finance: 25 Years of Developments," Public Budgeting & Finance, 25(4s), 3-45. Mullins, D. R. and B. A. Wallin. 2004. "Tax and Expenditure Limitations: Introduction and Overview," Public Budgeting & Finance, 24(4), 2-15. North, Douglass C. 1990. Institutions, Institutional Change, and Economic Performance. Cambridge, UK: Cambridge University Press. Oates, W. E. 1972. Fiscal Federalism. U.S.A: Harcourt Brace Jovanovich, Inc. Oates, W. E. 2008. "On the Evolution of Fiscal Federalism: Theory and Institutions," National Tax Journal, 61(2), 313-334. 115 Ostrom, E. 1998. "The Comparative Study of Public Economies," The American Economist, 42(1), 3-17. Ostrom, E. 2005. Understanding Institutional Diversity. Princeton, New Jersey: Princeton University Press. Ostrom, E. 2009. "A General Framework for Analyzing Sustainability of Social-Ecological Systems," Science, 325, pp.419-422. Ostrom, Vincent, Robert Bish, and Elinor Ostrom. 1988. Local Government in the United States. San Francisco, CA: ICS Press, Institute for Contemporary Studies. Ostrom, V. and E. Ostrom. 1991. "Public Goods and Public Choices: The Emergence of Public Economies and Industry Structures," in V. Ostrom, The Meaning of American Federalism, pp.163-197. San Francisco, CA: ICS Press. (Originally appeared in E. S. Savas, ed., 1977. Alternatives for Delivering Public Services: Toward Improved Performance. Boulder, Colorado: Westview Press) Ostrom, Vincent, Charles M. Tiebout, and Robert Warren. 1961. "The Organization of Government in Metropolitan Areas: A Theoretical Inquiry," The American Political Science Review, 55(4), 831-842. Petracca, Mark P. and Karen M. O'Brien. 1999. "The Impact of Municipal Term Limits," in Roger L. Kemp (ed.) Local Government Election Practices: A Handbook for Public Officials and Citizens, pp.433-445, Jefferson, North Carolina: McFarland & Company, Inc. Picazo-Tadeo, A., F. González-Gómez, J. Wanden-Berghe, and A. Ruiz-Villaverde. 2012. "Do ideological and political motives really matter in the public choice of local services 116 management? Evidence from urban water services in Spain," Public Choice, 151(1), 215- 228. Rubin, I. 2006. "Budgeting for Contracting in Local Government," Public Budgeting & Finance, 26(1), 1-13. Saha, S. 2011. "City-level analysis of the effect of political regimes on public good provision," Public Choice, 147(1), 155-171. Samuelson, P. A. 1954. "The Pure Theory of Public Expenditure," The Review of Economics and Statistics, 36(4), 387-389. Schelker, M. 2012. "The influence of auditor term length and term limits on US state general obligation bond ratings," Public Choice, 150(1-2), 27-49. Sokolow, A. D. and P. M. Detwiler. 2001. "California," in D. Krane, P. N. Rigos, M. B. Hill jr. (eds.), Home rule in America: a fifty-state handbook. Washington, D.C.: CQ Press. Stein, Robert M. 1990. Urban Alternatives: Public and Private Markets in the Provision of Local Services. Pittsburgh: University of Pittsburgh Press. Stiglitz, J. E. 1988. Economics of the Public Sector, 2nd ed., New York: W. W. Norton & Company. Sundell, A. and V. Lapuente. 2011. Published online. "Adam Smith or Machiavelli? Political incentives for contracting out local public services," Public Choice, doi: 10.1007/s11127- 011-9803-1. Tang, Shui Yan 1991. "Institutional Arrangements and the Management of Common-Pool Resources," Public Administration Review, 51(1), 42-51. Thousand Oaks City, April 12, 2011, "State of California Councilmember Term Limits Survey by City in Alphabetical Order," as Attachment #4 of the Memorandum, Thousand Oaks 117 Councilmembers Term Limits Initiative Impact Analysis Report & Scheduling for November 6, 2012 Election. Retrieved on May 27, 2013 from: http://www.ci.thousand- oaks.ca.us/government/depts/city_clerk/weblink.asp => CTO => CITY COUNCIL/REDEVELOPMENT AGENCY (RDA) => Agenda Packets => 2011 => 04/12/2011. Tiebout, C. M. 1956. "A Pure Theory of Local Expenditures," Journal of Political Economy, 64(5), 416-424. Wang, W. and Y. Hou. 2012. "Do Local Governments Save and Spend Across Budget Cycles? Evidence From North Carolina," The American Review of Public Administration, 42(2), 152-169. Weingast, B. R. 2009. "Second generation fiscal federalism: The implications of fiscal incentives," Journal of Urban Economics, 65(3), 279-293. Williamson, O. E. 1981. "The Economics of Organization: The Transaction Cost Approach," American Journal of Sociology, 87(3), 548-577. Williamson, O. E. 1985. The economic institutions of capitalism: firms, markets, relational contracting. NY: Free Press. Zhuravskaya, E. V. 2000. "Incentives to provide local public goods: fiscal federalism, Russian style," Journal of Public Economics, 76(3), 337-368.
Abstract (if available)
Linked assets
University of Southern California Dissertations and Theses
Conceptually similar
PDF
Civic associations, local governance and conflict prevention in Indonesia
PDF
The effects of interlocal collaboration on local economic performance: investigation of Korean cases
PDF
Three essays on the causes and consequences of China’s governance reforms
PDF
The risks and rewards of city management: how city managers evaluate the nature of the job and compensation
PDF
Why go green? Cities' adoption of local renewable energy policies and urban sustainability certifications
PDF
Essays on mitigation, adaptation, and resilience to urban fiscal crises
PDF
Governance networks: internal governance structure, boundary setting and effectiveness
PDF
The economy-environment trade-off in China's air pollution policies
PDF
Does collaborative R&D policy work? The effect of collaborative R&D policy on innovative activities of firms in Korea
PDF
How does collaborative governance work? The experience of collaborative community-building practices in Korea
PDF
The functions of the middleman: how intermediary nonprofit organizations support the sector and society
PDF
Governing regional collaboratives: institutional design, management and leadership
PDF
Local votes and outside money: campaign contribution geographic origins and their impact on Los Angeles City Council election outcomes
PDF
A study of Chinese environmental NGOs: policy advocacy, managerial networking, and leadership succession
PDF
Goal setting and performance in federal agencies
PDF
A view from below: the development and role of organizational social capital in neighborhood regeneration in Los Angeles
PDF
Institutional analysis of factors critical to facilitating project management and implementation of Metro ExpressLanes
PDF
Quango reforms and challenges in South Korea: social relations, informal networks, and hidden actions
PDF
The social making of authoritarian environmentalism: protest-litigation nexus and policy changes in China
PDF
Governing public goods: how representation and political power in local and regional institutions shape inequalities
Asset Metadata
Creator
Kim, Cheongsin
(author)
Core Title
Essays on fiscal outcomes of cities in California
School
School of Policy, Planning and Development
Degree
Doctor of Philosophy
Degree Program
Policy, Planning, and Development
Publication Date
07/29/2015
Defense Date
05/14/2013
Publisher
University of Southern California
(original),
University of Southern California. Libraries
(digital)
Tag
charters,council-manager form,fiscal federalism,Local government,OAI-PMH Harvest,public finance,term limits
Format
application/pdf
(imt)
Language
English
Contributor
Electronically uploaded by the author
(provenance)
Advisor
Tang, Shui Yan (
committee chair
), Sellers, Jefferey M. (
committee member
), Suarez, David (
committee member
)
Creator Email
cheongsk@usc.edu,goldspirit@daum.net
Permanent Link (DOI)
https://doi.org/10.25549/usctheses-c3-306493
Unique identifier
UC11292627
Identifier
etd-KimCheongs-1890.pdf (filename),usctheses-c3-306493 (legacy record id)
Legacy Identifier
etd-KimCheongs-1890.pdf
Dmrecord
306493
Document Type
Dissertation
Format
application/pdf (imt)
Rights
Kim, Cheongsin
Type
texts
Source
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 a...
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
Tags
council-manager form
fiscal federalism
public finance
term limits