Managing Fiscal Volatility: An Empirical Analysis of California County Governments’ Saving Behavior

Abstract

The increase in the importance of countercyclical behavior has expanded the research on fiscal saving behavior to local governments. In particular, the Great Recession has shown that local governments are not immune to economic shocks, spurring interest in local savings behavior. County governments are particularly vulnerable to negative economic shocks, as they rely more on intergovernmental revenues. With a focus on the determinants of fiscal slack, we empirically examined the relationship between tax revenue volatility and unassigned fund balance in 57 California counties over the period of 2004 to 2014. Employing spatial regression models, our empirical analysis revealed that revenue volatility is positively associated with general unassigned fund balance in California counties, and revenue diversification has partially positive effects on the fund balance. We infer that tax revenue volatility threatens the stabilized delivery of local services, which suggests that local governments should look to the factors that potentially affect revenue stability to improve their capacity for financial management. The spillover effects from the findings suggest that spatial effects need to be taken into account in analyzing the determinants of local fiscal slack.

Keywords

fiscal volatility local saving behavior California County governments unassigned fund balance spillover effects

Introduction

Economic stabilization has been the main role of the national government; however, subnational governments have recognized the need for countercyclical fiscal capacity over the economic rollercoaster of boom and recession (Gramlich, 1987). Economic shocks have had relatively less impact on local governments due to their heavy reliance on property taxes, which are often considered the most stable revenue source (Wolkoff, 1987). However, economic recessions, especially that of 2008, have had a negative effect on housing markets and the fiscal condition of local governments; thus, local governments are no longer seen as immune to a recession (Alm et al., 2011; Chernick et al., 2011; Paulais, 2009).

As a response, public finance scholars have turned their attention to fiscal slack and the countercyclical behavior of local governments. The extant literature has expanded the subject of analysis from general-purpose governments (Hendrick, 2006) to special-purpose governments and school districts (Arapis et al., 2017; Duncombe & Hou, 2014; Moulick & Taylor, 2017). In addition, methodological approaches to this issue have been developed (Guo & Wang, 2017; Stewart, 2009). Studies have consistently shown that the variations in the effects of fiscal slack result from organizational efforts to cope with risk and uncertainty; furthermore, they have empirically proved the effects on diverse fiscal conditions in local governments, such as long- and short-term debt levels, governing conditions, and operating deficits (Arapis et al., 2017; Arapis & Reitano, 2018; Duncombe & Hou, 2014; Hendrick, 2006; Stewart, 2009, 2011; Stewart et al., 2013; Wang & Hou, 2012).¹

However, slight differences within this literature suggest that the unique fiscal environments of individual governments should be considered. Empirical models could be more experimental by including variables that are theoretically valid but have not been previously considered. In spite of the efforts in this research area, relatively few studies focus on county governments (Stewart, 2009, 2011; Stewart et al., 2013; Wang & Hou, 2012),² and less attention has been paid to potential spillover effects in local fiscal slacks (Guo & Wang, 2017). In keeping with the research trend, this study aims to explore the factors that affect the level of fiscal slack in local governments to contribute to the growing literature on local fiscal slack. For an empirical analysis, this study focuses on California counties from 2004 to 2014 and employs a spatial Durbin model (SDM) that can estimate the potential spillover effects in fiscal slack accumulation. This study confirms the potential spillover effects in fiscal slack accumulation, as well as the role of fiscal structures in counties in fiscal slack accumulation.

The remainder of this article is structured as follows. In the next sections, we identify the factors believed to motivate local governments to accumulate slack resources based on findings from the existing fiscal savings literature and provide an overview of fiscal rules and propositions in California county governments. We then present the research methods used for our empirical analysis including the data, variables, and model specification and the results of the empirical analysis. In the last sections, we discuss the implications of the findings and conclude with the limitations of our study and suggestions for future research.

Literature Review

In the real world, economic issues such as inflation and unemployment are inevitable as the economy moves through the business cycle from expansion to recession to expansion again (Stock & Watson, 1999). The classical school of economics and the Keynes school have different views about the business cycle. Keynesian economics views that economic fluctuations are not the natural flow of a market economy but rather a symptom of market failure; therefore, government intervention in the economy is necessary (Keynes, 1936). Citing the instability of private sector spending, such as investment in public facilities and housing, and the demand for durable goods during the Great Depression, Keynes argued that both unemployment and recession were due to a lack of effective demand. Keynesian economics advocates an expansionary fiscal and monetary policy as a solution. Keynes believed that private investment depended more on uncertainty, such as expectations of prospective yields, than on interest rates; thus, increasing government spending despite a deficit was more effective in increasing effective demand (Keynes, 1936).³ Although the adverse effects of an expansionary fiscal policy during a recession, such as a crowding-out effect, have been pointed out, economic stabilization is generally considered to be one of the main functions of public finance (Musgrave, 1959; Musgrave & Musgrave, 1994).

Before the Great Depression, the classical school dominated the market, but the United States has since realized that the invisible hand is not enough to bring the economy back to its equilibrium. From 1929 to 1933, the GDP decreased by more than one fourth, and unemployment rate increased from 2.9% to 20.9% (Crafts & Fearon, 2010). With deflation, debt repayment became more difficult. The New Deal programs of the Roosevelt administration show Keynes theory in practice and highlight the importance of economic stabilization policy.⁴

Theoretically, economic stabilization can be achieved by tax smoothing, which suggests that it is optimal for a government to keep the tax rate constant over time. This implies that the government can minimize distortions from taxation in the economy by running a budget surplus when spending is temporarily high and a budget deficit when spending is temporarily low. That is, budget surpluses and deficits can be a buffer, given a certain level of spending, and the tax smoothing approach helps governments balance their budget over the business cycle (Chen, 2003). In terms of fiscal management, this principle is realized through countercyclical fiscal policy, more specifically through fiscal saving mechanisms such as budget stabilization funds, commonly called “rainy day funds” (Hou, 2003; Hou & Moynihan, 2008; Knight & Levinson, 1999; Sobel & Holcombe, 1996; Wagner, 2003).

Conventional wisdom has suggested that achieving economic stabilization through countercyclical fiscal policy is the responsibility of the national government, not necessarily the subnational governments. The rationale is that demand and labor shifts across subnational governments occur, causing unnecessary distortions in resource allocation and labor supply in the market. Citing empirical evidence against the market distortion claim, however, Gramlich (1987) argues that countercyclical fiscal policy need not be limited to the national government. The claim that countercyclical fiscal policy is the sole responsibility of the national government stems from the idea that this policy adoption would lead to a shift in production factors; however, this may not be the case: such shifts do not occur immediately and are limited in scale and scope (Gramlich, 1987). This evidence suggests that subnational governments do not need to shy away from the stabilization role. Public finance scholars have shifted their view on countercyclical fiscal policy toward the subnational level as efforts toward budget stabilization rather than economic stabilization (Hou & Smith, 2010; Wang & Hou, 2012). Empirical studies on this view have paid particular attention to whether local governments accumulate fiscal savings to play the countercyclical fiscal role and whether those savings do indeed help alleviate fiscal stress and prevent abrupt fiscal adjustments during recessions.⁵

The importance of the budget stabilization role of local governments has grown with the increasing impacts of economic downturns on local finances (Hendrick, 2006; Marlowe, 2005). During the early 1980s recession, in aggregate, the fiscal position of local governments deteriorated, making some jurisdictions that had not invested in fiscal capacity enact expenditure and service reductions and revenue increases (Bahl, 1982). Although the local sector was not as fiscally stressed as many believe, this was possible in large part by deferring necessary expenditures and reducing public services (Bahl, 1982; Wolman, 1983). As structural problems have accumulated, however, the fiscal impacts of recessions have grown in scope and strength (Knight et al., 2003). Unlike the previous two recessions, more significant discretionary fiscal actions, such as raising tax rates, expanding tax bases, and raising user fees, were taken during the early 1990s recession (Mattoon & Testa, 1992). Even though the early 2000s recession was shallow by historical standards, its aftermath was more severe than that of the 1990s recession. The deficit of the early 2000s was much bigger than that of the early 1990s, and the effects of the economic downturn lasted longer, cutting into core local functions and services (McGuire & Steuerle, 2003).

The primary sources of the increasing local fiscal pressures come from the state level. Experiencing a wave of fiscal crises, state governments have reduced financial aid to local governments, while at the same time shifting expenditure responsibilities to them (Jimenez, 2009; McFarland & Pagano, 2015; Miller & Hokenstad, 2014). State grants-in-aid for local governments, the most important revenue source for local governments (Congressional Budget Office, 2013), have steadily declined since the 1960s, as states have tried to balance their own budgets (Conant, 2003; Jimenez, 2009; Reschovsky, 2004). State aid to local governments hit the lowest level in 2011 when the federal stimulus funds that had been injected to overcome the Great Recession ran out (Gordon, 2018). Cutting financial aid means shifting expenditure responsibilities to local governments. Adjusting to fiscal hardships, states tend to reduce their share of developmental, public safety, and allocational expenditures, in particular (Jimenez, 2009).

Reflecting the fiscal challenges that local governments are facing, there have been a few important cases of local bankruptcy over the past decade.⁶ The fact that the level of local fiscal distress is different this time around is evidenced partly by how the number of bankruptcy filings has changed over time. Mattoon (2011) reports 600 municipal bankruptcy filings (less than 10 cases per decade) from 1934 (when federal legislation permitting local bankruptcies was enacted) to 2010. Governing Magazine reports that 61 municipalities have filed for bankruptcy since 2010 (Governing, 2019). Although local bankruptcies and credit defaults remain rare historically,⁷ many local governments are in serious fiscal distress (Spiotto et al., 2012).⁸

The housing crisis during the 2007 to 2009 recession negatively affected local tax revenues through a number of channels. Although property tax revenues did not drop immediately following the housing market contraction due to lags between market and assessed property values, housing price declines and foreclosures had negative impacts on real estate transfer taxes and sales tax revenues, the second largest source of local own-source revenues, either directly (through decreases in sales of construction materials) or indirectly (through decreases in consumption). Moreover, revenues from personal income tax declined due to reductions in construction-related employment (Lutz et al., 2011).⁹ To be sure, property tax is a relatively stable source of revenue for local governments, compared to sales and other taxes. However, a historical review of property tax’s cyclical variability shows that its reputation as a stable revenue source is not as solid as it has been in the past (Alm et al., 2011; Chernick et al., 2011; Paulais, 2009).

The Institute for Local Government (2013), the nonprofit research and education affiliate of the League of California Cities and the California State Association of Counties, provides a basic overview of general components of county and city revenues. According to this report, compared to cities, counties rely heavily on intergovernmental revenues. As of 2013, the proportions of funding from the federal and state government were 57.22% for counties and 38.94% for cities. In contrast, the proportions of property and sales and use taxes, which are generally not limited to covering the cost of federal and state programs for health and human services, were 24.65% for counties and 32.69% for cities. This suggests that with federal and state aid to local governments becoming increasingly uncertain and volatile, county governments are more likely to be vulnerable to unexpected shocks in the economy and changes in the county fiscal structure, such as new state mandates and voter-imposed tax cuts (Chapman, 2003).

Although the fiscal environments of local governments have steadily deteriorated, our expectation of local fiscal performance during economic downturns has risen. There has been a growing awareness that fiscal crises cannot be attributed only to cyclical factors (Knight et al., 2003) and that management matters in preventing fiscal crises (Jimenez, 2011, 2013; O’Toole & Meier, 1999). Many studies have shown that countercyclical fiscal capacity does help local governments stabilize current expenditures (Hendrick, 2006; Marlowe, 2005) and suggested that economic downturns inevitably result in cuts in expenditures, disruptions in service delivery, short-term borrowings to finance long-term debts and operational activities, and inter-fund transfers to manage the general fund (Felix, 2014). This awareness has made it difficult for local governments to attribute fiscal crises only to external causes such as cyclical factors (Knight et al., 2003). As a result, this is a motivation for a local government to build adequate capacity to cope with revenue declines, higher operating costs, and decreases in state financial aid during downturn periods in light of their unique fiscal environment, which includes the cyclical variabilities of revenue and demand for services, capital level, and operation and maintenance costs for infrastructure.

Local governments do not maintain slack resources just to minimize the effects of economic downturns. They hold slack to manage cash flow, to borrow on preferential terms, to take advantage of investment opportunities, and to finance political premises (Arapis & Reitano, 2018; Gianakis & Snow, 2007; Guo & Wang, 2017; Hendrick, 2006). These activities are in fact more important in tough times from the perspective of Keynesian economics and countercyclical fiscal policy. They are possible with prudent and strategic fiscal management, and more volatile localities are more likely to build a greater capacity to invest in infrastructure and development activities over the business cycle (Jimenez, 2013). Volatility influences the fiscal saving behavior of counties, but there are differences in how volatility in specific revenue sources affects the size of savings (Stewart et al., 2013).

A local government and its contiguous governments within a state exert mutual influence on each other under the same fiscal rules due to their proximity. In a decentralized system, local governments independently decide how much to collect (taxation) and spend (expenditure). Scholars have revealed that spatial dependence influences local fiscal decisions through comparisons with neighbors, leading to such forms of competition such as tax competition, yardstick competition, and the Leviathan hypothesis (Besley & Case, 1995; Brueckner, 2003; Burge & Piper, 2012; Shleifer, 1985; Tiebout, 1956; Wildasin, 2003; Wilson, 1986). As a result, if a local government changes its fiscal policy packages, those of their neighbors will also be affected. In the same vein, the levels of fiscal slack in both a local government and its neighbors follow the same trend as a form of spillover effects. Therefore, this study aims to account for the neighboring effects on fiscal slacks.

Fiscal Rules and Conditions in California County Governments

California has 58 counties and 482 municipalities, including the consolidated city-county government of San Francisco; in addition, there are over 400 redevelopment agencies, nearly 3,400 special districts, and more than 1,000 K-12 and community college districts (League of California Cities, 2016). In the FY 2014 to 2015, total state and local government revenues and expenditures in California were over 510 billion dollars and 515 billion dollars, respectively. The shares of local governments in the total amounts were 57.35% for revenue (approximately 293 billion dollars) and 56.31% for expenditure (approximately 290 billion dollars). More specifically, the California State Controller’s Office (CSCO; 2017) reported 66.51 billion dollars for revenue and 64.16 billion dollars for expenditure in California county governments.

Among the 58 counties, nine counties have more than 1 million residents, while 23 counties have fewer than 100,000 residents. California county governments show a wide range in population and there are significant differences in jurisdiction size.¹⁰ The Statute of California requires the state to be divided into counties and provides the powers of California counties. Except for San Francisco, counties are governed by an elected board of supervisors, and these supervisors have both legislative and executive powers. They carry out a wide range of functions that provide health and human services as state agents do (i.e., foster care, child welfare services, public health and mental health services, and substance use disorder treatment), and municipal-type services in unincorporated areas, including policing, fire protection, libraries, planning, and road repair. Furthermore, counties in unincorporated areas provide some services that municipalities provide (i.e., policing and fire protection).

The counties can levy various taxes and charges to fund public services. In particular, property taxes exhibited quite stable movement over the 1980s, with an average growth rate of approximately 10% (Legislative Analyst’s Office, 2014). However, the annual growth rate dropped rapidly to nearly 0% in the mid-1990s and then bounced back to the 1980s level (i.e., 10%). In addition, the volatility of California property taxes increased even more in the first decade of the 2000s (U.S. Bureau of Census, 2016). Figure 1 illustrates how the annual growth rate spike mid-decade when the housing market was booming and then dropped steeply in 2010 in the aftermath of the 2008 financial crisis.¹¹

Figure 1.

Annual growth rate of local property taxes in California.

In the context of California, a series of ballot initiatives including Proposition 13, legislated in 1978, have put additional fiscal constraints on local governments. Savage (1992) provides a brief review of the ballot initiatives that California’s electorate has approved. Proposition 13 limited the countywide property tax rate to 1% of a property’s assessed value, effectively cutting property taxes in half. The initiative also limited the annual increase in the property tax base (monetary values that are placed on taxable property) to 2%. The initiative stipulates that governments must obtain a two-thirds majority of taxpayers to further raise tax rates. California counties may increase the 1% tax to pay for voter-approved debt but not to increase revenues for services and general funds. In addition, Proposition 4, Proposition 98, Proposition 99, and Proposition 111 have indirectly impacted local finances by limiting the state’s ability to increase expenditures, retain revenues collected in excess of the spending limit, and issue bonds.

These propositions have imposed many forms of constraints on fiscal management under the state constitution and other statutes. At the state level, all taxes require the consent of more than two thirds of the state legislature but do not require voter consent. In addition, user fees and charges do not require voters’ approval, but they are subject to approval by a majority of the legislature. Constraints on local governments, such as counties, are more restrictive than constraints on the state. At the local level, general-purpose taxes require the consent of more than two thirds of the legislature and a majority of voters, while a special-purpose tax requires the consent of two thirds of both the legislature and voters. In terms of non-tax revenues, special assessments require the consent of a majority of both the legislature and the voters, while fees for property-related services require the consent of a majority of the legislature. These fees, excluding water, sewer, or refuse collection, electric and other utility services, also require separate voter consent.

In addition, the fiscal autonomy of counties depends on the California state constitution and legislature. Proposition 58 in 2004 mandated that state and local governments balance their own budgets. This act defined BBRs (Balanced-Budget Requirements), mid-year budget adjustments, reserve requirements, and debt-related provisions following the fiscal difficulties of the late 1990s. The Government Finance Officers Association (GFOA) emphasized the need for a general fund balance in local governments and stated that California county governments are required to have their budgets balanced. The GFOA has highlighted the fund balance in local government and recommended a general fund balance as a formal policy in 2011; furthermore, the GFOA (2011) pointed out the importance of a compact plan to replenish a fund balance to an appropriate level.

Data, Variables, and Empirical Model

Data

To empirically examine fund balance in local governments, the analysis relies on a panel data set containing the annual audit reports of 57 counties in California over the period covering 2004 to 2014.¹² Since the data set includes the recession period of 2007 to 2009, this study can control for the lagging effects of the recession, and this panel data set can capture the effects on fund balance before, during, and after the recession. Prior to explaining the variables for our empirical analysis, Table 1 summarizes the data for the descriptive statistics of the variables.

Table 1.

Descriptive Statistics (n = 627; 57 Counties in 2004–2014).

Variables	M	SD	Minimum	Maximum
General unassigned fund balance to total revenue	0.104	0.095	−0.034	0.709
Tax revenue diversification	0.362	0.137	0.076	0.783
Tax revenue volatility	0.062	0.059	0.000	0.341
Transfers to total revenue	0.564	0.098	0.321	0.799
Debt to total income	0.011	0.010	0.000	0.061
Total expenditure (per capita)	1,883.83	1,657.13	920.16	15,650.50
Total expenditure (per capita; logged)	7.405	0.423	6.825	9.658
Population (in 1,000 persons)	636	1,420	1	10,100
Population (logged)	11.990	1.797	7.009	16.129
Young population share	0.268	0.044	0.171	0.369
Senior population share	0.138	0.040	0.065	0.270
Per capita income (in US$1,000)	38.882	11.753	24.894	94.625
Per capita income (logged)	10.532	0.257	10.122	11.458
Unemployment rate (%)	9.798	4.075	3.400	29.083
Poverty rate (%)	15.229	4.882	5.600	29.700
Voters to Republican candidates (%)	50.430	12.780	16.701	71.313

Variables

Dependent variables

Governments manage various funds according to different accounting and financial reporting standards than those used by for-profit organizations. The different standards result from differing organizational goals, methods of revenue generation, and budgetary obligations that consider political, economic, and socio-demographic conditions to improve government accountability (Governmental Accounting Standards Board [GASB], 2006). Furthermore, the diversity in the nature of governmental operations and in some legal and fiscal constraints makes it difficult to record all governmental financial transactions and balances in a single accounting entity.

In 2011, the implementation of the Governmental Accounting Standards Board 54 (GASB 54) brought about changes in the way local governments report the categories of fund balance. Before the GASB 54, the fund balance at the local level was categorized into reserved and unassigned fund balances, and the unassigned fund balance comprised designated and undesignated balances. However, the data set is based on the consistent categories in our sample period, because the CSCO revised the financial transactions report forms to incorporate GASB 54, effective from the fiscal year of 2016 to 2017. Therefore, the sample years in this study are compatible with each other, although the GASB was implemented in 2011. This consistency enables this study to obtain standardized budget items. According to the Generally Accepted Accounting Principles, all California counties maintain a fund balance with the following five components (CSCO, 2018).¹³ County governments have their own fund balance formulation policies and requirements for minimum fund balances. To identify the factors that determine fund balances in California local governments, this study uses the dependent variable of general unassigned fund balance as a percentage of the total revenue of each county government because the unassigned fund balances are readily available resources.¹⁴

Explanatory variables

This study chose two groups of explanatory variables that affect the total unassigned fund balance (hereafter, UFB) to investigate determinants such as financial factors and environmental factors.

Financial factors

UFB is a function of the saving behavior of local governments, and financial conditions directly influence that behavior. We first consider the revenue structure of California county governments, which comprises mainly property tax, sales and use tax, transient lodging tax, utility user tax, other taxes, and grant revenue.¹⁵ Extant research indicates their importance in the structure of governments’ revenue sources, and revenue diversification stabilizes revenue structure because a more diversified revenue structure mitigates the risk of losing any single revenue source (Carroll, 2009; Sjoquist & Stoycheva, 2012). Stabilization in revenue structure helps governments decrease any unexpected revenue losses and increase their fund balance. Therefore, we expect that revenue diversification stabilizing local fiscal structure is more likely to increase the UFB level.

To measure the revenue structure, we obtained diversified levels of tax revenue using the Herfindahl–Hirschman index (HHI) because diversification affects the cyclical variability in the budgetary items of each county government. As a result, the revenue structure affects spending decisions. The HHI is defined in this study as:

T D_{i t} = \frac{1 - (\sum_{i = 1}^{5} r^{2})}{1 - (1 / 5)}

where $T D_{i t}$ is the diversified level of tax revenue in each government $i$ in a fiscal year $t$ and $r$ is the proportion of each funding source $(i)$ to total tax revenue.

UFB is regarded as an ideal vehicle to mitigate instability, and governments need to maintain the fund balance at an adequate level to mitigate any future uncertainty. However, fund revenue fluctuates over the fiscal years and local governments are more sensitive to those fluctuations due to their limited fiscal sources, resulting in an increase in fund revenue instability. Prior research has revealed that instability leads governments to increase tax burdens and/or to spend their UFB to resolve the uncertainty of revenue fluctuation (Duncombe & Hou, 2014; Stewart, 2009; Wang & Hou, 2012). In contrast, revenue stability expands the accumulated level of a UFB (Marlowe, 2005; Stewart, 2009). Revenue volatility from uncertainty results may result in unexpected changes in the UFB level. Because most local revenue depends on tax revenue, this study primarily focuses on tax revenue for volatility and expects that tax revenue volatility is more likely to increase the UFB level in local governments.

Tax revenue volatility is defined as the gap between actual revenue and predicted revenue (Carroll, 2009). To measure volatility,¹⁶ this study obtained the residuals using linear trend estimates that regress tax revenue on cross-sectional and year variances.¹⁷ The residuals represent how the actual tax revenue deviates from the predicted tax revenue growth. The residuals were converted to absolute values and taken in natural logarithm form, which measures the deviation of general fund revenue from the expected growth trend.

Although property tax and sales tax are the most important sources of tax revenue for California county governments, the revenue source that takes up the biggest share of the total revenue is intergovernmental transfers from the federal and state governments. However, local governments cannot play a role in the decision-making process for grant allocation by upper-level governments. The uncertainty of grant allocation poses a risk of higher fiscal burdens on local governments. Local governments hardly recognize the risks in their reliance on grants (Stewart, 2009) and spend grant funds as if the grants were their own-source revenue (Ebel & Yilmaz, 2003). Therefore, any unexpected changes (gains or losses) in transfers would affect the UFB level. A high reliance on grants is expected to pose a risk to the UFB level; thus, this study expects a negative effect of transfers on the UFB level. In addition to the revenue sources, expenditure has a strong effect on UFB and debts are negatively associated with the UFB level (Hendrick, 2006). Hence, two more financial variables, expenditure in natural logarithm form and the ratio of debts to per capita income, are considered.

Environmental factors

Economic and demographic factors are defined as environmental factors. First, budgeting and taxation are a political process because local representatives consider their voters’ preferences. Fund balance level is not separate from political influence.¹⁸ The preferences are measured as the ratio of voters to Republicans as an indicator of the political environment of a county (Wang & Hou, 2012), drawn from CQ Press (2017).¹⁹ In addition to the political environment, a county’s economic environment affects its fiscal decisions. The first economic indicator is per capita income, collected by the U.S. Bureau of Economic Analysis (2017), because a wealthy county may have a higher tax revenue capacity and be able to maintain a higher level of fund balance. However, income level is not sufficient to describe the variability of the business cycle, and unemployment and poverty rates, retrieved from the U.S. Bureau of Labor Statistics (2017), are considered more efficient to account for the effects of stabilizers and economic development on the business cycle. Finally, the ethnic properties of the population structure, as well as eligibility for retirement within a population, might affect the saving behavior in the UFB (Marlowe, 2005; Stewart, 2009). A county’s demographic composition, provided by the U.S. Bureau of Census (2017), is considered because counties should be accountable for their anonymous residents, and the ethnic composition of population may affect the fiscal decision-making process. An increase in population size expands the demand for public services and changes in population structure determine the types of and patterns in public services needed. Therefore, demographic indicators included in the model are the population in natural log form and the ratios of individuals younger than 20 years old and over 65 years old to the overall population.²⁰

According to Table 1, the dependent variable is shown to have large variances. In this sample, 17 observations of nine individual counties exhibit negative values in their general unassigned fund balance. Figure 2 illustrates the ratio of fund balance to total revenue in California county governments over the sample period. In the sample period, the unassigned fund balance was 10.45% of the total revenue. In the first year of the sample, the ratio of balance to total revenue was 8.42%; then, the unassigned general fund balances on average steadily increased to 12.10% prior to 2006, when the balance was at its highest. Since 2006, drops in the ratio of balance to total revenue were observed during the recession period, but California county governments have since maintained an average balance ratio of 10%. In particular, the fund balances increased steadily after the recession period. The trend in Figure 2 reveals that California county governments save in the boom years and spend in the bust years, especially in a period of economic recession. In the sample period, the average unreserved fund balance in California counties is between 8% and 13%.

Figure 2.

Unassigned fund balance as a percent of total revenue.

In addition, Figure 3 compares the average ratios of unassigned fund balance to total revenue in each county to see the maintenance level of fund balance in each county. The four boxplots reveal that average unassigned fund balance ratio to total revenue ranges from 0% to 5% in 13 counties, 5% to 10% in 24 counties, 10% to 15% in nine counties, and 15% higher in 11 counties. Some counties in the first two groups between 0% and 10% had negative levels in the fund balance ratio. Furthermore, some counties (i.e., Calaveras, Medocino, Monterey, Orange, Sutter, and Trinity) show greater variation in their fund balance ratio to total revenue.²¹

Figure 3.

Boxplots of average fund balance ratios in each county (2004–2014).

Model Specification

Any changes in fiscal conditions would have noticeable effects in the short term; in addition, the effects would last longer for economic and rational choices (Koethenbuerger, 2011). Because the sample includes the recession period, the fiscal conditions are a transfer function of the current moment and past changes are time-lag effects that occur over time rather than all at once (Judge et al., 1985). A fund balance level and its changes depend on time-lag effects, especially immediately before and after a recession period, and strategic interactions among counties. Thus, this study utilizes the determinants of fund balance in three ways that consider both time-lag effects from a distributed lag model of dynamic effects²² and spatial effects from a geospatial econometric model.

This study considers spatial dependence on fund balance across county governments. Different spatial econometric models have been developed to capture differences in spatial dependence. The two spatial models in widespread use are the spatial autoregressive model (SAR) and spatial error model (SEM)²³; however, they are not able to fully account for another possible spatial interaction because the explanatory variables of a county and its neighbors mutually influence a county’s decisions about its fund balance. An alternative spatial analysis is the SDM, which includes spatially lagged independent variables to capture the possible interactions (Elhorst, 2010). Furthermore, the SDM can reduce omitted variable bias (LeSage & Pace, 2010). The SDM is represented as:

\begin{array}{l} F B_{i t} = α + γ W F B_{i j t} + δ_{1} F F_{i t - 1} + λ_{1} E F_{i t} + \\ δ_{2} W F F_{i j t - 1} + λ_{2} W E F_{i j t} + μ + ε_{i t} \end{array}

(1)

ε_{i t} = ρ W ε_{i t} + τ_{i t}

(2)

where the dependent variable $(F B_{i t})$ indicates the level of the general fund balance of a county $i$ in a fiscal year $t$ . The variable $(F F)$ denotes the financial factor, and we lagged these variables to rule out reverse causality. The variable $(E F)$ recognizes environmental factors. The term $(W)$ is a row-standardized spatial weighting-matrix to obtain the average values of the fund balance of a county’s neighbors $j$ . The error term (µ) represents cross-section and time-invariant unobserved effects, and the random error term $(ε_{i t})$ finalizes the regression models.²⁴ Prior to conducting the SDM model, we considered a fixed-effect regression model and a system-general method of moments (system-GMM) model²⁵ that include all the geospatially lagged explanatory variables. The models include a set of dummy variables indicating fiscal years to specify the fixed effects over time, but the dummy variables are not spatially lagged due to multicollinearity.

Model Estimation

To examine the determinants of general UFBs in California county governments, we performed three different empirical estimations to check robustness between the empirical models: fixed effects, system-GMM, and SDM. Prior to empirically examining the determinants, several tests for their validity were conducted to determine the most appropriate specifications.

For the two fixed-effects regression models, Hausman’s specification test and the modified Wald test for groupwise heteroscedasticity were conducted. The two tests revealed that a fixed-effects estimator is more appropriate for the data and that our data has heteroscedasticity. The test results also suggest that a fixed-effects regression model with robust standard errors is the most appropriate.

For the system-GMM model, Roodman (2009) noted that GMM estimation requires three tests for validity. First, the Arellano–Bond test for autocorrelation of the error terms revealed no second-order autocorrelation AR(2) in the sample, p > AR(2) = 0.428.²⁶ Second, the Hansen-J test for overidentification failed to reject the null hypothesis that there are no overidentified instruments (p > J = 0.768). Finally, the test of exogeneity of instruments failed to reject the null hypothesis that instruments are exogenous in the two models (p = .572). The three tests reveal that the sample satisfies the system-GMM estimator requirements.

For the most appropriate spatial model, we conducted a two-step test for the selection (Belotti et al., 2016; Elhorst, 2010). The first step was to select SAR or SEM as long as the Lagrange multiplier (LM) test indicated the existence of spatial dependence. In the second step, after estimating the SDM, the likelihood-ratio (LR) test indicated the more appropriate choice between SDM and SAR/SEM. For our sample, the analysis of the SDM yielded a more consistent estimation by adding the average values of the neighbors for the independent variables to the model because the SDM model cannot be simplified to either the SAR or the SEM model. According to the two tests, the SDM approach is more appropriate because SDM is unbiased regardless of the data generation process (LeSage & Pace, 2014). Furthermore, the Hausman test indicated that using a fixed-effects SDM is more efficient than a random-effects SDM (Belotti et al., 2016).²⁷

Regression Findings

Table 2 compares the estimation results that regress the general-UFB on explanatory variables using the fixed-effects model (FE), the system-GMM model (GMM), and the SDM.^28,29 Among the four variables for a county’s financial condition, tax revenue volatility has a consistently significant and positive effect on the general-UFB, in addition to the significant effects of neighboring county governments. Positive effects of tax revenue volatility mean that a county government tends to keep a higher level of UFB when it faces a volatile revenue structure. More specifically, a 1% higher degree of volatility in a county’s tax revenue will increase the fund balance by 7.5 percentage points in the fixed-effects result, 7.1 percentage points in the GMM result, and 8.2 percentage points in the SDM result. Although the tax revenue volatility in the neighbors has no significant effect on the fund balance of a county, the persistent effects of volatility become approximately 40% greater on the fund balance.

Table 2.

Model Comparisons of the Determinants on General Unassigned Funds in California (FY 2004–2014).

Model^a	FE		GMM		SDM
Model^a	Coefficient	Coefficient × W	Coefficient	Coefficient × W	Coefficient	Coefficient × W
Spatially lagged dependent variable		−0.544**		−0.459**		−0.172*
Spatially lagged dependent variable		(0.220)		(0.178)		(0.103)
Tax revenue volatility (t − 1)	0.075**	−0.022	0.032	0.071	0.086**	−0.054
Tax revenue volatility (t − 1)	(0.038)	(0.121)	(0.034)	(0.074)	(0.038)	(0.074)
Tax revenue diversification (t − 1)	0.068	−0.018	−0.018	−0.042	0.066**	−0.034
Tax revenue diversification (t − 1)	(0.042)	(0.072)	(0.036)	(0.071)	(0.034)	(0.062)
Transfers/total revenue (t − 1)	−0.086	−0.515***	−0.139***	−0.232**	−0.095*	−0.473***
Transfers/total revenue (t − 1)	(0.110)	(0.192)	(0.049)	(0.111)	(0.051)	(0.121)
Debt/income (t − 1)	−0.272	1.394	−0.648	1.004	−0.051	0.384
Debt/income (t − 1)	(0.753)	(1.385)	(0.580)	(1.229)	(0.435)	(0.957)
Expenditure (per capita; logged)	−0.125***	−0.016	−0.159***	0.083	−0.129***	−0.046
Expenditure (per capita; logged)	(0.044)	(0.084)	(0.033)	(0.077)	(0.027)	(0.049)
Population (logged)	−0.025*	0.465*	−0.143	0.794**	−0.048	0.334*
Population (logged)	(0.014)	(0.267)	(0.212)	(0.316)	(0.120)	(0.180)
Young population share	−0.272	−1.946	0.426	−0.258	0.387	−1.207
Young population share	(0.285)	(1.972)	(0.715)	(1.313)	(0.502)	(0.966)
Senior population share	−0.330	−1.869**	1.280*	−0.527	0.010	−1.459**
Senior population share	(0.285)	(0.829)	(0.749)	(1.039)	(0.328)	(0.707)
Per capita income (logged)	−0.069	0.106	−0.197***	0.185	−0.104***	0.205**
Per capita income (logged)	(0.079)	(0.149)	(0.064)	(0.120)	(0.039)	(0.088)
Unemployment rate	−0.006*	0.008	−0.009**	0.016**	−0.001	0.003
Unemployment rate	(0.003)	(0.006)	(0.004)	(0.008)	(0.002)	(0.005)
Poverty rate	0.002	−0.004	0.002	−0.005*	0.001	−0.003
Poverty rate	(0.002)	(0.004)	(0.001)	(0.003)	(0.001)	(0.003)
Voters to Republican candidates	0.000	0.002	−0.001	0.003**	0.001	0.002
Voters to Republican candidates	(0.001)	(0.002)	(0.001)	(0.002)	(0.001)	(0.001)
Constant	1.110
Constant	(2.180)
Observations	627		570		570
Number of counties	57		57		57
R ²	.337				.382
F-statistic	4.45***		3.78***		1,058.224

Note. FE = fixed-effects; GMM = generalized method of moments; SDM = spatial Durbin model.

The value (1,058.224) in the SDM regression model indicates the log-likelihood, instead of the F-statistic value. Year dummies were included for empirical analysis but are not shown here. Robust standard errors are in parentheses.

Statistical significance levels are *p < .1. **p < .05. ***p < .01.

Although the other financial variables exhibit significant effects on the balance, they are not consistent in the three different models. The degree of tax revenue diversification leads a county government to keep a higher balance in the GMM and SDM results; however, their statistical significance level is at 90%. Moreover, a negative neighboring effect of diversification is observed in the long-run perspective of the GMM results.

A county’s dependence on transfers from the upper-level governments has a negative effect on the fund balance from the dynamic estimates of the system-GMM. Consistent with the flypaper effect (Inman, 2008), transfers to a recipient local government increase the level of local public spending more than an increase in the recipient’s income of an equivalent size. The counties’ dependence on transfers for their revenue are shown not to have significant effects on fund balance in the short-run perspective. The negative effects of transfers are only significant in the system-GMM model, which implies persistent effects from previous years, and the negative effects of transfers become greater in the long-run perspective of the dynamic estimates. This implication is consistent with the negative effects of total expenditure in a county government.³⁰

Turning our attention to spatial dependence, a strategic interaction between a county and its neighboring counties is observed. The spatial lag of the dependent variable reveals a significant and negative effect on the general-UFB of a county, which implies that a 1% increase in the fund balance of the neighboring counties will reduce the fund balance of the county. Furthermore, the comparisons of the fixed-effects and the system-GMM results reveal that the interdependence of general-UFB among county governments becomes greater in a negative way from a long-term perspective. The negative effects in the long term are 50% greater than in the short term. In addition, there are two interesting financial variables. First, the degree of diversification of tax revenue in a county’s neighboring counties has a negative effect on the balance only in the system-GMM model. In addition, the dependence on transfers exhibits a consistently negative effect in the all-spatial models. The neighboring counties that depend more on transfers for their total revenue increase their expenditure, similar to the negative effects of a county’s dependence on transfers. We can also infer that a county would expand its expenditure to follow the expansions of its neighbors as a combination of the flypaper effect and competition among governments. None of the control variables of socio-demographic conditions in a county government are statistically significant.

Discussion and Conclusion

The provision of fiscal stability is central to the purpose of government, with the management of risk and uncertainty as its basic function. Therefore, the goal of a government’s budget process is stability (Dothan et al., 2013). Local governments, closest to the citizens, are more responsible for providing essential public goods and services to residents in their localities. The UFB is usually considered a kind of savings account to help local governments cope with unexpected contingencies or revenue shortfalls due to an economic recession. Given the role of local governments to ensure the continuity of essential public service delivery, it is important to explore the determinants of their UFB levels. This research tackles this question by considering possible spatial autocorrelations. Therefore, this study not only enhances the understanding of factors that affect the UFB levels but also improves this line of research methodologically.

Regarding the potential factors that have an impact on UFB levels, previous studies have rarely considered the volatility of the revenue stream. This study deems revenue volatility an important explanatory variable if a UFB is considered to be a potential savings vehicle for a local government. The estimation confirms the expectation that a higher degree of volatility in the revenue stream tends to be associated with a higher level of UFB. The study, however, did not find consistent evidence of an association between revenue diversification and fiscal slack. This is an interesting result because it has been widely believed that revenue diversification contributes to revenue stability (Jordan & Wagner, 2008; Yan, 2011). If this is true, revenue diversification may indirectly affect the level of fiscal slack. Our results indicate that revenue diversification does not necessarily affect fiscal slack accumulation, despite its possible impacts on revenue stability.

The spatial model estimation observed spillover effects as a strategic interaction among county governments in California. This is an interesting finding consistent with numerous studies reporting spillover effects on fiscal behaviors. Governments do not operate in an isolated environment; they, especially localities, operate surrounded by governments that perform similar functions and deliver a similar package of public services due to the shorter distances to borders. With the improved mobility of local residents and the increasing exposure of those residents to information on not only their own but also neighboring governments, the idea of “voting with their feet” (Tiebout, 1956) makes more sense than ever. This environment has led local governments to operate in a competitive mode and to be constantly attentive to how their neighbors are doing their work. The growing evidence for spillover effects at the local level probably reflects this trend. The empirical findings reveal potential spillover effects on the local general unassigned fund balance, which crowds out the spending of a county to its neighboring counties, which leads the UFB to hold at a higher level.

These findings have important implications for local government practitioners responsible for fiscal performance and academics interested in local fiscal savings behavior. First, they suggest that local government practitioners should consider tax revenue volatility a serious threat to the stable delivery of essential local services. Pursuing stability, they need to pay attention to how stable or volatile their tax revenue is over the business cycle, and how robust or vulnerable their fiscal system is to unexpected revenue shocks. How much fiscal slack is needed depends on how volatile the government’s fiscal environment is (Joyce, 2001). In the interest of ensuring fiscal stability, therefore, local public administrators should look to their own fiscal structure and environment and to the potential factors that affect revenue stability. Rules for the appropriate level of fiscal reserves, like an unassigned fund balance of no less than 2 months (about 15%) of general fund operating expenditures or revenues as recommended by GFOA (or the 5% rule for upper-level governments), cannot be generalized for all levels of government. It can be claimed that a rule might be too strict for some governments but too lenient for others. What is important is to identify and maintain the optimal level of fiscal slack in accordance with each government’s fiscal conditions and unique fiscal environment.

Second, this study contributes to the fiscal savings literature by broadening the methodological spectrum. The SDM model employed in this study enabled us to control for any potential endogeneity from spatial autocorrelation and estimate the interdependence with the least omitted variable bias because all the explanatory variables are spatially lagged. The SDM results revealed strategic interactions among county governments when their UFB levels were determined. As a county expands its spending for public service, its UFB level decreases, which leads the neighbors to hold their UFB levels at a higher level. Furthermore, geospatial dependence parcels out the demographic and political factors in determining the UFB level that the extant studies above have pointed out. These findings, therefore, highlight the importance of geospatial analysis in the empirical examination of the determinants of local fiscal slack levels. They show that externalities of fiscal savings occur at the county level as well, suggesting that not controlling for spillover effects among localities could lead to bias in the empirical results.

In spite of the contributions above, this study is limited to one state, and hence one should be cautious when trying to generalize its findings.³¹ This study considers spillovers in the UFB between counties. The spillovers in a federal system, furthermore, might be observed across counties and sub-counties. These two points can raise a question for future research. As mentioned earlier, local governments do not accumulate slack resources just to weather economic downturns. They also do so for cash flow management, higher credit ratings, and investments in infrastructure and regional development with future returns. The literature suggests that the optimal level of fiscal slack depends on each government’s unique environment. In addition to revenue volatility, increased demand for services, high capital spending, high maintenance costs, and old infrastructure may also motivate higher levels of fiscal slack. Although revenue volatility can be said to be the underlying factor affecting these variables, the question of how each of them plays a role in local savings requires a separate analysis. Also, whether spillover effects of fiscal slack are the same between counties and municipalities would be an interesting question to address. Physical distances (center-to-center) between counties differ from those between municipalities, as they differ in size. Geospatial analysis is based on geographic information, and therefore, such differences in geographic characteristics might produce different results. Comparisons among local governments in different states would also be an interesting research theme. State-local fiscal relations vary across states. Some states are active in preventing local fiscal crises, while others are not. The possibility of state bailouts differs across states, leading to different levels of budget constraints for local governments. Future research is warranted to investigate how these differences in state oversight and financial aid affect local governments’ savings behavior.

Footnotes

Appendix

Table A2.

Spatial Regression Result of the Determinants on General Unassigned Funds in California (FY 2004–2014).

Model^a	FE		GMM		SDM
Model^a	Coefficient	Coefficient × W	Coefficient	Coefficient × W	Coefficient	Coefficient × W
Spatially lagged dependent variable		−0.508**		−0.532***		−0.176*
Spatially lagged dependent variable		(0.207)		(0.193)		(0.095)
Tax revenue volatility (t − 1)	0.077*	−0.052	0.030	0.070	0.088**	−0.073
Tax revenue volatility (t − 1)	(0.040)	(0.119)	(0.034)	(0.075)	(0.038)	(0.074)
Tax revenue diversification (t − 1)	0.066*	−0.035	−0.024	−0.047	0.074**	−0.029
Tax revenue diversification (t − 1)	(0.039)	(0.071)	(0.036)	(0.073)	(0.033)	(0.062)
Transfers/total revenue (t − 1)	−0.059	−0.533***	−0.128**	−0.237**	−0.039	−0.397***
Transfers/total revenue (t − 1)	(0.110)	(0.174)	(0.050)	(0.112)	(0.052)	(0.115)
Debt/income (t − 1)	−0.350	0.746	−0.445	0.857	−0.063	0.198
Debt/income (t − 1)	(0.762)	(1.428)	(0.596)	(1.275)	(0.455)	(1.027)
Ratio of debt service to total expenditure	−0.084	0.304	−0.098	0.161	−0.074	0.104
Ratio of debt service to total expenditure	(0.182)	(0.482)	(0.080)	(0.310)	(0.160)	(0.338)
Ratio of general government to total expenditure	−0.129	−0.248	−0.169***	0.081	−0.125**	−0.264**
Ratio of general government to total expenditure	(0.099)	(0.218)	(0.044)	(0.129)	(0.052)	(0.129)
Ratio of capital outlays to total expenditure	−0.321**	−0.273	−0.153**	−0.191	−0.270***	−0.254
Ratio of capital outlays to total expenditure	(0.137)	(0.254)	(0.075)	(0.191)	(0.104)	(0.223)
Population (logged)	−0.018	−0.005	−0.086	0.764**	−0.011	0.264
Population (logged)	(0.011)	(0.020)	(0.216)	(0.325)	(0.119)	(0.181)
Young population share	−0.338	−0.585	0.601	−0.370	0.162	−0.958
Young population share	(0.262)	(0.857)	(0.734)	(1.370)	(0.503)	(0.996)
Senior population share	−0.285	−1.684**	1.361*	−0.516	0.054	−1.508**
Senior population share	(0.266)	(0.737)	(0.762)	(1.093)	(0.332)	(0.743)
Per capita income (logged)	−0.077	0.298*	−0.207***	0.217*	−0.111***	0.154*
Per capita income (logged)	(0.068)	(0.174)	(0.065)	(0.121)	(0.040)	(0.089)
Unemployment rate	−0.006*	0.007	−0.008**	0.015*	−0.001	0.003
Unemployment rate	(0.003)	(0.007)	(0.004)	(0.008)	(0.002)	(0.005)
Poverty rate	0.002	−0.003	0.002	−0.006**	0.001	−0.004
Poverty rate	(0.002)	(0.004)	(0.001)	(0.003)	(0.001)	(0.003)
Voters to Republican candidates	−0.000	0.002	−0.000	0.004**	0.001	0.002
Voters to Republican candidates	(0.001)	(0.002)	(0.001)	(0.002)	(0.001)	(0.001)
Constant	0.582
Constant	(1.534)
Observations	627		570		570
R ²	.357				.388
Number of Counties	57		57		57

Note. FE = fixed-effects; GMM = generalized method of moments; SDM = spatial Durbin model.

FE and GMM are same with Table 2. Year dummies were included for empirical analysis but are not shown here. Robust standard errors are in parentheses.

Statistical significance levels are *p < .1. **p < .05. ***p < .01.

Declaration of Conflicting Interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author(s) disclosed receipt of the following financial support for the research and/or authorship of this article: Authors received funding support from Hankuk University of Foreign Studies Research Fund for this article.

ORCID iD

Sunjoo Kwak

Notes

Author Biographies

Jongmin Shon is an assistant professor in the School of Public Affairs and Administration, Rutgers University – Newark. His research interests include financial management and fiscal policy issues in public sector.

Sunjoo Kwak is an associate professor in the Department of Public Administration at the Hankuk University of Foreign Studies in Seoul, South Korea. His research interests include public budgeting and finance, fiscal management, and political economy.

References

Adams

Clark

Ezrow

Glasgow

(2004). Understanding change and stability in party ideologies: Do parties respond to public opinion or to past election results? British Journal of Political Science, 34(4), 589–610.

Alm

Buschman

R. D.

Sjoquist

D. L.

(2011). Rethinking local government reliance on the property tax. Regional Science & Urban Economics, 41(4), 320–331.

Arapis

Reitano

(2018). A glimmer of optimism in government savings accumulation? An empirical examination of municipal unassigned fund balance in Florida. Public Finance Review, 46(3), 389–420.

Arapis

Reitano

Bruck

(2017). The fiscal savings behavior of Pennsylvania school districts through the great recession. Public Budgeting & Finance, 37(3), 47–70.

Arellano

Bond

(1991). Some tests of specification for panel data: Monte Carlo evidence and an application to employment equations. The Review of Economic Studies, 58(2), 277–297.

Arellano

Bover

(1995). Another look at the instrumental variable estimation of error-components models. Journal of Econometrics, 68(1), 29–51.

Bahl

(1982). The fiscal health of state and local governments: 1982 and beyond. Public Budgeting & Finance, 2(4), 5–22.

Baltagi

B. H.

(2008). Econometric analysis of panel data. John Wiley.

Baum

C. F.

Schaffer

M. E.

Stillman

(2007). Enhanced routines for instrumental variables/GMM estimation and testing. Stata Journal, 7(4), 465–506.

10.

Belotti

Hughes

Piano Mortari

(2016). Spatial panel data models using Stata. CEIS Research Paper, 14(5), 1–38.

11.

Besley

T. J.

Case

(1995). Incumbent behavior: Vote-seeking, tax-setting, and yardstick competition. The American Economic Review, 85(1), 25–45.

12.

Bork

Møller

S. V.

(2015). Forecasting house prices in the 50 states using Dynamic Model Averaging and Dynamic Model Selection. International Journal of Forecasting, 31(1), 63–78.

13.

Brueckner

J. K.

(2003). Strategic interaction among governments: An overview of empirical studies. International Regional Science Review, 26(2), 175–188.

14.

Burge

G. S.

Piper

(2012). Strategic fiscal interdependence: County and municipal adoptions of local option sales taxes. National Tax Journal, 65(2), 387–415.

15.

California State Controller’s Office. (2017). Government financial reports. https://bythenumbers.sco.ca.gov

16.

California State Controller’s Office. (2018). Accounting standards and procedures for counties. Author. https://www.sco.ca.gov/Files-ARD-Local/ASP_Manual.pdf

17.

Carlesso

(2018, March). City council greenlights state bailout for Hartford. The Hartford Courant. https://www.courant.com/community/hartford/hc-news-hartford-bailout-vote-20180326-story.html

18.

Carroll

D. A.

(2009). Diversifying municipal government revenue structures: Fiscal illusion or instability? Public Budgeting & Finance, 29(1), 27–48.

19.

Chapman

J. I.

(2003). Local government autonomy and fiscal stress: The case of California counties. State and Local Government Review, 35(1), 15–25.

20.

Chen

D. H. C.

(2003). Intertemporal excess burden, bequest motives, and the budget deficit (Vol. 3086). World Bank Publications.

21.

Chernick

Langley

Reschovsky

(2011). The impact of the Great Recession and the housing crisis on the financing of America’s largest cities. Regional Science & Urban Economics, 41(4), 372–381.

22.

Conant

(2003). State budgeting in FY 2002: Dealing with the bust phase of the boom and bust cycle, again! Public Budgeting & Finance, 23(2), 1–4.

23.

Congressional Budget Office. (2013). Federal grants to state and local governments. Congress of the United States. https://www.cbo.gov/sites/default/files/113th-congress-2013-2014/reports/43967federalgrants.pdf

24.

CQ Press. (2017). Voting and elections collection. http://www.cqpress.com/product/CQ-Voting-and-Elections-Collection.html

25.

Crafts

Fearon

(2010). Lessons from the 1930s Great Depression. Oxford Review of Economic Policy, 26(3), 285–317.

26.

Dothan

M. U.

Hand

M. L.

Pierson

Thompson

(2013). A positive model of expenditure growth. https://papers.ssrn.com/sol3/papers.cfm?abstract_id=1077558

27.

Douglas

J. W.

Gaddie

R. K.

(2002). State rainy day funds and fiscal crises: Rainy day funds and the 1990-1991 recession revisited. Public Budgeting & Finance, 22(1), 19–30.

28.

Duncombe

Hou

(2014). The savings behavior of special purpose governments: A panel study of New York school districts. Public Budgeting & Finance, 34(3), 1–23.

29.

Ebel

R. D.

Yilmaz

(2003). On the measurement and impact of fiscal decentralization. In Martinez-Vazquez

Alm

(Eds.), Public finance in developing and transitional countries (pp. 101–119). Edward Elgar.

30.

Elhorst

J. P.

(2010). Applied spatial econometrics: Raising the bar. Spatial Economic Analysis, 5(1), 9–28.

31.

Felix

(2014). The use of inter-fund transfers to manage the “bottom line” in the municipal context. Journal of Governmental & Nonprofit Accounting, 4(1), 17–31.

32.

Gianakis

G. J.

Snow

(2007). The implementation and utilization of stabilization funds by local governments in Massachusetts. Public Budgeting & Finance, 27(1), 86–103.

33.

Gordon

(2018). Predicting municipal fiscal distress: Aspiration or reality? https://www.lincolninst.edu/sites/default/files/pubfiles/gordon_wp18tg1_0.pdf

34.

Gordon

Center

U.-B. T. P

. (2018). Predicting municipal fiscal distress: Aspiration or reality? Lincoln Institute of Land Policy.

35.

Governing. (2019). Bankrupt cities, municipalities list and map. https://www.governing.com/gov-data/municipal-cities-counties-bankruptcies-and-defaults.html

36.

Governmental Accounting Standards Board. (2006). Why governmental accounting and financial reporting is—and should be—different. https://www.gasb.org/cs/ContentServer?c=Document_C&cid=1176169371273&d=&pagename=GASB%2FDocument_C%2FDocumentPage

37.

Government Finance Officers Association. (2011). Best practice: Replenishing fund balance in the general fund (2011) (budget and CAAFR) (new). http://gfoa.org/sites/default/files/GFOA_BestPracticeReplenishingFundBalance_0.pdf

38.

Gramlich

E. M.

(1987). Subnational fiscal policy. In Quigle

J. M.

(Ed.), Perspectives on local public finance and public policy (pp. 3–27). JAI Press.

39.

Guo

H. D.

Wang

(2017). A spatial analysis of Florida county governments’ unreserved general fund balances. Public Budgeting & Finance, 37, 71–88

40.

Hendrick

(2006). The role of slack in local government finances. Public Budgeting & Finance, 26(1), 14–46.

41.

Hou

(2003). What stabilizes state general fund expenditures in downturn years—Budget stabilization fund or general fund unreserved undesignated balance? Public Budgeting & Finance, 23(3), 64–91.

42.

Hou

(2006). Budgeting for fiscal stability over the business cycle: A countercyclical fiscal policy and the multiyear perspective on budgeting. Public Administration Review, 66(5), 730–741.

43.

Hou

Duncombe

(2008). State saving behavior: Effects of two fiscal and budgetary institutions. Public Budgeting & Finance, 28(3), 48–67.

44.

Hou

Moynihan

D. P.

(2008). The case for countercyclical fiscal capacity. Journal of Public Administration Research and Theory, 18(1), 139–159.

45.

Hou

Smith

D. L.

(2006). A framework for understanding state balanced budget requirement systems: Reexamining distinctive features and an operational definition. Public Budgeting & Finance, 26(3), 22–45.

46.

Hou

Smith

D. L.

(2010). Do state balanced budget requirements matter? Testing two explanatory frameworks. Public Choice, 145, 57–79.

47.

Inman

R. P.

(2008). Flypaper effect. http://www.nber.org/papers/w14579

48.

The Institute for Local Government. (2013). Understanding the basics of county and city revenues. https://www.hpca.gov/DocumentCenter/View/4121/Understanding-the-Basics-of-County-and-City-Revenues?bidId=

49.

International City/County Management Association. (2013). Facts you should know: State and municipal bankruptcy, municipal bonds, state and local pensions [2013 fact sheet]. https://icma.org/sites/default/files/304697_2013%20Facts%20You%20Should%20Know.pdf

50.

Jimenez

B. S.

(2009). Fiscal stress and the allocation of expenditure responsibilities between state and local governments: An exploratory study. State and Local Government Review, 41(2), 81–94.

51.

Jimenez

B. S.

(2011). Management quality and state bond ratings: Exploring the links between public management and fiscal outcomes. International Journal of Public Administration, 34(12), 783–799.

52.

Jimenez

B. S.

(2013). Strategic planning and the fiscal performance of city governments during the Great Recession. The American Review of Public Administration, 43(5), 581–601.

53.

Jordan

M. M.

Wagner

G. A.

(2008). Revenue diversification in Arkansas cities: The budgetary and tax effort impacts. Public Budgeting & Finance, 28(3), 68–82.

54.

Joyce

P. G.

(2001). What’s so magical about five percent? A nationwide look at factors that influence the optimal size of state rainy day funds. Public Budgeting & Finance, 21(2), 62–87.

55.

Judge

G. G.

Hill

R. C.

Griffiths

Lutkepohl

Lee

T.-C.

(1985). Introduction to the theory and practice of econometrics. John Wiley.

56.

Keynes

J. M.

(1936). The general theory of money, interest and employment. Harcourt, Brace & World.

57.

Knight

Kusko

Rubin

(2003). Problems and prospects for state and local governments. State Tax Notes, 29(6), 427–439.

58.

Knight

Levinson

(1999). Rainy day funds and state government savings. National Tax Journal, 52(3), 459–472.

59.

Koethenbuerger

(2011). How do local governments decide on public policy in fiscal federalism? Tax vs. expenditure optimization. Journal of Public Economics, 95(11–12), 1516–1522.

60.

League of California Cities. (2016). League of California cities 2016 annual reports. http://www.cacities.org/Resources-Documents/About-Us/Annual-Reports/League-of-California-Cities-2016-Annual-Report

61.

Lee

L.-f.

(2015). Identification of spatial Durbin panel models. Journal of Applied Econometrics, 31, 133–162.

62.

Legislative Analyst’s Office. (2014). Voting requirements to increase taxes and fees. https://lao.ca.gov/LAOEconTax/Article/Detail/20

63.

LeSage

J. P.

Pace

R. K.

(2010). Spatial econometric models. In Fischer

M. M.

Getis

(Eds.), Handbook of applied spatial analysis: Software tools, methods and applications (pp. 355–376). Springer.

64.

LeSage

J. P.

Pace

R. K.

(2014). Interpreting spatial econometric models. In Fischer

M. M.

Nijkamp

(Eds.), Handbook of regional science (pp. 1535–1552). Springer.

65.

Leuchtenburg

W. E.

(1963). Franklin D. Roosevelt and the new deal, 1932-1940. Harper and Row.

66.

Levitt

S. D.

(1996). How do senators vote? Disentangling the role of voter preferences, party affiliation, and senator ideology. The American Economic Review, 44(1), 425–441.

67.

Lutz

Molloy

Shan

(2011). The housing crisis and state and local government tax revenue: Five channels. Regional Science & Urban Economics, 41(4), 306–319.

68.

Marlowe

(2005). Fiscal slack and counter-cyclical expenditure stabilization: A first look at the local level. Public Budgeting & Finance, 25(3), 48–72.

69.

Mattoon

R. H.

(2011). Local governments on the brink (Special Issue). https://ideas.repec.org/a/fip/fedhle/y2011imayn286a.html

70.

Mattoon

R. H.

Testa

W. A.

(1992). State and local governments’ reaction to recession. Economic Perspectives, 16(2), 19–27.

71.

McFarland

Pagano

M. A.

(2015). City fiscal conditions 2015. https://www.nlc.org/sites/default/files/2017-05/CSAR%20City%20Fiscal%20Conditions%202015%20FINAL.pdf

72.

McGuire

T. J.

Steuerle

C. E.

(2003). A summary of what we know—and don’t—about state fiscal crises. State Tax Notes, 29(5), 357–361.

73.

Miller

D. B.

Hokenstad

(2014). Rolling downhill: Effects of austerity on local government social services in the United States. Journal of Sociology & Social Welfare, 41, 93.

74.

Moulick

A. G.

Taylor

L. L.

(2017). Fiscal slack, budget shocks, and performance in public organizations: Evidence from public schools. Public Management Review, 19(7), 990–1005.

75.

Musgrave

R. A.

(1959). The theory of public finance: A study in public economy. McGraw-Hill.

76.

Musgrave

R. A.

Musgrave

P. B.

(1994). Public finance in theory and practice. McGraw-Hill.

77.

O’Toole

L. J.

Jr. Meier

K. J.

(1999). Modeling the impact of public management: Implications of structural context. Journal of Public Administration Research and Theory, 9(4), 505–526.

78.

Paulais

(2009). Local governments and the financial crisis: An analysis. http://documents.worldbank.org/curated/en/762521468336022194/Local-governments-and-the-financial-crisis-an-analysis

79.

Pew Charitable Trusts. (2013). The state role in local government financial distress. https://www.pewtrusts.org/~/media/assets/2013/07/23/pew_state_role_in_local_government_financial_distress.pdf

80.

Pew Charitable Trusts. (2017). State strategies to detect local fiscal distress. https://www.pewtrusts.org/en/research-and-analysis/fact-sheets/2017/08/state-strategies-to-detect-local-fiscal-distress

81.

Reschovsky

(2004). The impact of state government fiscal crises on local governments and schools. State and Local Government Review, 36(2), 86–102.

82.

Roodman

(2009). How to do xtabond2: An introduction to difference and system GMM in Stata. The Stata Journal, 9(1), 86–136.

83.

Samuelson

P. A.

Scott

(1967). Economics: An introductory analysis (Vol. 715). McGraw-Hill.

84.

Savage

J. D.

(1992). California’s structural deficit crisis. Public Budgeting & Finance, 12(2), 82–97.

85.

Shleifer

(1985). A theory of yardstick competition. The RAND Journal of Economics, 16, 319–327.

86.

Sjoquist

D. L.

Stoycheva

(2012). Local revenue diversification: User charges, sales taxes, and income taxes. In Ebel

R. D.

Petersen

J. E.

(Eds.), The Oxford handbook of state and local government finance (pp. 429–462). Oxford University Press.

87.

Sobel

R. S.

Holcombe

R. G.

(1996). The impact of state rainy day funds in easing state fiscal crises during the 1990–1991 recession. Public Budgeting & Finance, 16(3), 28–48.

88.

Spiotto

J. E.

Acker

A. E.

Appleby

L. E.

(2012). Municipalities in distress? How states and investors deal with local government financial emergencies. Chapman & Cutler.

89.

Stewart

L. S. M.

(2009). Examining factors that impact Mississippi counties’ unreserved fund balance during relative resource abundance and relative resource scarcity. Public Budgeting & Finance, 29(4), 45–73.

90.

Stewart

L. S. M.

(2011). Governmental influence on unreserved fund balances for Mississippi counties. Journal of Public Budgeting, Accounting & Financial Management, 23(4), 478–506.

91.

Stewart

L. S. M.

Phillips

J. N.

Modlin

(2013). An evaluation of the connection between economic volatility and governmental savings levels for Illinois, North Carolina, and Mississippi counties. Public Finance & Management, 13(4), 361–382.

92.

Stock

J. H.

Watson

M. W.

(1999). Business cycle fluctuations in US macroeconomic time series. In Taylor

J. B.

Woodford

(Eds.), Handbook of macroeconomics (Vol. 1, pp. 3–64). Elsevier.

93.

Tiebout

C. M.

(1956). A pure theory of local expenditures. The Journal of Political Economy, 64(5), 416–424.

94.

U.S. Bureau of Census. (2016). Annual Survey of State and Local Government Finances. https://www.census.gov/programs-surveys/gov-finances.html

95.

U.S. Bureau of Census. (2017). Population and housing unit estimates. https://www.census.gov/programs-surveys/popest/data/data-sets.html

96.

U.S. Bureau of Economic Analysis. (2017). Regional economic accounts. http://www.bea.gov/regional/index.htm

97.

U.S. Bureau of Labor Statistics. (2017). Unemployment rates for states. http://www.bls.gov/lau/

98.

Wagner

G. A.

(2003). Are state budget stabilization funds only the illusion of savings? Evidence from stationary panel data. The Quarterly Review of Economics and Finance, 43(2), 213–238.

99.

Wagner

G. A.

Sobel

R. S.

(2006). State budget stabilization fund adoption: Preparing for the next recession or circumventing fiscal constraints? Public Choice, 126, 177–199.

100.

Wang

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.

101.

Whatley

(2011). Facts you should know: State and local bankruptcy municipal bonds state and local pensions. http://knowledgecenter.csg.org/kc/content/facts-you-should-know-state-and-local-bankruptcy-municipal-bonds-state-and-local-pensions

102.

Wildasin

D. E.

(2003). Fiscal competition: An introduction. Journal of Public Economic Theory, 5(2), 169–176.

103.

Wilson

J. D.

(1986). A theory of interregional tax competition. Journal of Urban Economics, 19(3), 296–315.

104.

Wolkoff

M. J.

(1987). An evaluation of municipal rainy day funds. Public Budgeting & Finance, 7(2), 52–63.

105.

Wolman

(1983). Understanding local government responses to fiscal pressure: A cross national analysis. Journal of Public Policy, 3(3), 245–263.

106.

Yan

(2011). The interactive effect of revenue diversification and economic base on US local government revenue stability. Public Money & Management, 31(6), 419–426.