610 forum

Municipal Fiscal Policy Space and Fiscal Structure: Tools for Managing

Spending Volatility

REBECCA HENDRICK AND JARED CRAWFORD

In this research, we examine the impact of municipal governments’ internal fiscal structure and external fiscal policy space on operational spending volatility and fund balances. We test our model using a system of two equations that is estimated with panel data consisting of 265 municipalities in the Chicago metropolitan region from 1997 to 2009. We find that these governments are managing spending volatility using tools available in both dimensions and that the fund balance is a critical tool for this purpose. We also find that the impact of a stringent versus relaxed set of options (home rule) for managing this financial problem is likely to be more complex than what is represented in prior research.

INTRODUCTION

State and local governments in the United States (US) have experienced a financial roller coaster

in the last 15 years beginning with a period of fiscal expansion in the late 1990s and then two

recessions in 2001 and 2008. Income and sales taxes in many state and local governments

swelled in the late 1990s due to the strong bull market of that period that increased investment

income and consumer spending (Jenny 2003; Weiner 2004, 520). Tax revenues for both levels of

government then began to decline in 2000 from market events that precipitated what was

described as the “worst fiscal crisis since the Depression” for state government (Krane 2004, 28).

Many state and local governments had recovered from this recession by the end of 2003 or 2004

(Pew Center on the States 2010) only to be hit by the Great Recession in early 2008 that was to be

deeper and longer than the earlier recession. Unlike prior recessions, however, the Great

Rebecca Hendrick is at the Public Administration (M/C 278), University of Illinois at Chicago, 412 S. Peoria St.,

Chicago, IL 60607. She can be reached at [email protected]. Jared Crawford is at the Public Administration (M/C

278), University of Illinois at Chicago, 412 S. Peoria St., Chicago, IL 60607. He can be reached at jcrawf6@uic.

edu.

© 2014 Public Financial Publications, Inc.

24 Public Budgeting & Finance / Fall 2014

Recession significantly reduced property values, which is the primary source of revenue for most

local governments.

The volatility of state and local government financial condition during this time period offers

a good opportunity to examine its effects on government spending and conditions that impact

this relationship. Volatility in conditions that affect government spending and revenue is not

considered desirable because it is difficult to predict these events and budget and strategize for

the future (Jordan 2003; Gamage 2010). Volatility in revenue sources may also require

governments and public organizations to cycle continuously between growth and cut-back in

their spending, especially if they have no surplus funds or ability to borrow. Alternatively,

governments may raise or lower tax rates and charges to compensate for such volatility, but these

changes are very disruptive to constituents and operations and are likely to be more costly to

manage than slow growth and decline. These types of financial problems have engendered a

large body of research and some standards of policy and practice on topics such as how

governments should manage volatility and fiscal stress using internal resources (e.g. rainy day

funds), how to lessen the effects of such events on government revenue streams (e.g. revenue

diversification), and the impact of fiscal policy constraints on the ability of government to

respond to such events (e.g. tax and expenditure limitations [TELs]).

The research presented here develops a conceptual model and tests an empirical model of

financial problem solving in local government that merges these three factors to explain

operational spending volatility in 265 municipalities in the Chicago metropolitan region from

1997/1998 to 2009. This model focuses on two broad theoretical components of this system that

determine the tools that local governments are likely to employ when solving financial problems

such as fiscal stress and volatility. These elements are the government’s fiscal policy space and

fiscal structure. The fiscal policy space (FPS) is the set of exogenous parameters within which

city officials operate, decide, and create fiscal policy (Pagano and Hoene 2010), and it delineates

a fundamental portion of the range of options available to them for solving financial problems. A

government’s fiscal structure is produced by the fiscal choices that officials make over time

within their FPS and denotes its key financial characteristics, such as diversification of revenue,

tax burden, and level of fund balance. Compared to the FPS, fiscal structure is relatively

endogenous, although some of its features are more stable and less changeable by government

officials than others.

The empirical model tested here incorporates different attributes of the FPS and fiscal

structure into a system of two equations that explains spending volatility and fund balance in

these municipalities over three distinct business cycles. The model specifies fund balance as

endogenous in the spending volatility equation based on the expectation that government

officials actively manipulate this fiscal structural feature and tool more than others to manage

spending volatility, and they manipulate it in conjunction with these other features as part of a

broader strategy for managing uncertainty and the environment. As conceptualized here, the

fund balance is a central feature within this process and it will be important to determine how the

FPS and other structural features impact it directly to gain a more comprehensive understanding

of how local governments manage spending volatility. Thus, the model presents fund balance as

a separate dependent variable in the system and not simply as an endogenous variable to be

Hendrick and Crawford / Municipal Fiscal Policy Space and Fiscal Structure 25

instrumented in the spending volatility equation. Additionally, the model does not consider

spending volatility to have an effect on fiscal structure in the current time period, including the

fund balance, so the model is recursive rather than simultaneous. In this case, spending volatility

is conceptualized as being directly affected by the attributes of the government’s FPS, structural

features, and external financial conditions connected with the business cycle, and it is indirectly

affected by these factors via their impact on the fund balance.

The model also examines the effect of one particular attribute of the FPS that is usually

associated with state government limitations on local government finances. Generally speaking,

that attribute might be characterized as strict or relaxed state government restrictions on local

government fiscal policies and the tools they use to solve financial problems. Most often these

constraints take the form of TELs that restrict local governments’ ability to increase taxes,

usually property taxes, but states also determine which sources of revenue are available to

government and make many other demands on their finances and practices. Testing the model

using municipalities in Illinois provides an opportunity to examine a broader set of financial

constraints on local governments than TELS alone and determine the combined impact of these

constraints on spending volatility and the use of the fund balance.

In Illinois, home rule municipalities have much more discretion than nonhome rule

municipalities over property taxes, sales taxes, debt, and other financial areas including contracts

and impact fees. Home rule municipalities can levy more types of taxes than nonhome rule

municipalities and are not subject to property tax limitations. About half the municipalities in the

region are home rule and many of these acquired home rule status during the period of the study.

Thus, the model incorporates the effects of a strict versus relaxed set of constraints on the financial

options available to local governments (the fiscal tool box) for managing spending volatility.

This framework for understanding how local governments manage revenue and spending

volatility, and financial problems more generally, is much more comprehensive than what has

been presented in previous studies that usually focus on one feature of government fiscal

structure, such as the fund balance or revenue diversification. Additionally, very little of this

research examines financial problem solving at the local level. Consistent with prior research at

both the state and local level, the panel data examined here shows that revenue diversification

and fund balance reduce spending volatility in these governments. The results also show that

governments maintain higher fund balances when other structural and environmental features

limit or enhance their revenue volatility and flexibility in managing spending volatility.

Furthermore, the effects of some structural features on spending volatility and fund balance are

weaker for home rule governments, which suggest that flexibility in fiscal options reduces the

importance of specific tools for managing financial problems. The next section of this study

presents the theoretical background for the conceptual model.

FISCAL OPTIONS, FISCAL STRUCTURE, AND FINANCIAL PROBLEMS

How local government officials cope with fiscal stress, financial volatility, and other problems

that threaten the achievement of their policy objectives and the fiscal health of their government

26 Public Budgeting & Finance / Fall 2014

depends on the options or tools available to them. Some of these tools are part of the internal

financial structure of government that officials have created over time, and other tools, such as

the FPS, are part of the external environment over which officials have less control. Pagano and

Hoene (2010) recognize five attributes that determine the FPS of U.S. cities. Of these attributes,

the state/intergovernmental context is particularly important because of its impact on a local

government’s economic base and spending needs and demands, which are two other attributes of

the FPS.

The economic base determines the wealth of the revenue bases from which local governments

draw revenues to finance spending. Spending needs and demands determine how much they must

spend to ensure that services are adequate and citizens satisfied. State governments, however,

control local governments’ taxing authority, debt limits, services they are responsible for, and

many other areas of fiscal policy. Some state governments even dictate local governments’

financial management practices, such as in the areas of budgeting and accounting. Thus, state

government context controls local governments’ access to revenue bases irrespective of the

wealth of the revenue base, and it controls the scope and level of services local governments must

provide irrespective of the needs and demands for such services. Although a jurisdiction may have

a lot of commercial activity, many states do not allow their local governments to levy a sales tax,

which eliminates this option from their fiscal toolbox. Similarly, the need for public service

expenditures may be high in a jurisdiction, but only some local governments may be required to

provide these services and their spending for these services may be limited by state government.

According to the FPS framework, the state context establishes much of the size and content of

the fiscal toolbox of local governments, but their internal fiscal structure provides other tools for

solving financial problems. Although state government may limit the ability of local government

to increase taxes and spending, within these parameters local governments decide the level of

taxation and debt, which sources of revenues to rely on, how much to spend for different services,

and how much revenue to reserve each year. All other things being equal, revenue in

municipalities will be more diversified in states that allow these governments to levy a sales or

income tax than in states that do not allow them to levy such taxes. Likewise, jurisdictions with

high property values can achieve the same level of services at a lower tax rate compared to

jurisdictions with low property values. Some of these structural features, such as reserves and

revenue diversification, are established to help governments cope with problems such as

volatility and fiscal stress. Other structural features, such as debt levels and earmarking of

revenues for particular purposes, create future obligations for government and limit options

within their fiscal toolbox.

The decisions local governments make to cope with financial problems can be conceptualized

as adaptation as described in the seminal works of Simon (1973), Thompson (1967), and Cyert

and March (1963). With respect to financial decisions, governments are viewed as adapting their

financial structure and practices to changes in the environment, including changes to the

attributes of their FPS, but their options or tools are also constrained within the parameters of

their FPS and by the availability of tools within their fiscal toolbox. A good example of this

process is evident from local governments’ responses to TELS that came into existence in most

states beginning in the 1970s.

Hendrick and Crawford / Municipal Fiscal Policy Space and Fiscal Structure 27

Most TELS target the property tax by limiting aggregate tax levies, rates, or assessed

values and/or the increases to these property tax components. Research shows that local

governments have adapted to these changes by becoming more reliant on elastic and volatile

revenue sources over which they have control rather than reducing spending (Mullins and

Wallin 2004; Brunori et al. 2008). Many governments also have instituted practices such as more

aggressive earmarking and property assessments to circumvent these constraints (Martell and

Teske 2007; Springer et al. 2009). TELs seem to curb the growth of local governments only in

states where the constraints are particularly strict or comprehensive or where there are additional

limitations on spending and revenues (Shadbegian 1999; Brown 2000; Cornia and Walters

2006). Research also shows that local governments with particularly stringent TELs are not able

to rebound from recessions as well and have more revenue volatility than governments with less

strict and binding constraints (Levinson 1998; Martell and Teske 2007).

One lesson that can be drawn from research on TELS is that governments with a liberal or

relaxed FPS will have greater flexibility to adapt to changes in their environment, either positive

or negative, and maintain spending priorities and growth than governments with a more stringent

TEL and limited FPS. We also expect that governments with fewer constraints on their FPS to

have greater capacity to manage revenue and spending volatility and fiscal stress in the short

term. All governments, however, have the ability to structure their fiscal features in ways that

help them accommodate both expected and unexpected future changes in their environment

given the limitations imposed by their FPS. For instance, governments with less flexibility in

their FPS or more volatile revenues may maintain a higher fund balance, which allows them to

maintain spending levels when recessions occur, and they may rely on their fund balance more

than governments with a lot of flexibility and fiscal tools.

From an accounting perspective, a fund balance is the difference between all assets and

liabilities (net assets) in an account or fund (e.g. general fund). The fund balance also represents

the accumulation of monetary surpluses (revenues minus expenditures) over time, and is easily

accessible by government to meet the spending obligations of that fund or even other funds

through interfund borrowing. Most state and local governments maintain fund balances for the

purpose of managing cash flow during the fiscal year and as reserved resources for specific

purposes in the future. Many state governments also have created separate rainy day funds as a

source of surplus funds to manage volatility and fiscal stress, but local governments tend to

supplement and use their fund balances for this purpose (Wolkoff 1987). Prior research shows

that fund balances are often equal to or greater than 50 percent of expenditures, which is far

greater than what is usually needed to manage cash flow (Marlowe 2005; Hendrick 2011).

Most of the research on the effect of surplus funds on the volatility of government spending

and responses to recessions is at the state level and shows that, on balance, these funds stabilize

state spending and give these governments greater capacity to manage fiscal shocks. Hou (2006),

for instance, finds that states that established rainy day funds before the 1980, 1982, and 1991

recessions fared better than those that did not. He also finds that rainy day funds have a slight

counter-cyclical effect on state spending, but fund balances have no such effect

(Hou 2003, 2005). Douglas and Gaddie (2002) determined that large rainy day funds helped

to alleviate state fiscal stress in the 1991 recession. Somewhat later, Hou and Moynihan (2008)

28 Public Budgeting & Finance / Fall 2014

showed that states were able to use both types of surplus funds to enable spending and avoid tax

cuts during the 1991 and 2001 recessions. Research also shows, however, that these funds can be

destabilizing if not properly implemented (Pollock and Suyderhoud 1986) and that governments

often do not deposit enough in them to be effective (Sobel and Holcombe 1996; Navin and

Navin 1997; Lav and Berube 1999).

At the local level, Marlowe (2005) finds that fund balances have a marginal but positive effect

on the direction and level of expenditures relative to expected growth from 1990 to 2000 and

Hendrick (2006) shows that the fund balance and other surplus resources helped municipalities

in the Chicago metropolitan region improve their current financial condition in the late 1990s and

early 2000s. This research also shows that these governments accumulated higher fund balances

when faced with certain conditions that increased their exposure to risky and volatile events and

when they lacked slack resources in other areas of their fiscal structure, such as capital spending,

debt service, and enterprise funds. Beyond these two studies, however, there is very little

research on how local governments use their fiscal structure to manage volatility and other

financial problems.

Fund balances and other surplus funds represent only one type of slack resources within a

government’s fiscal structure. When defined broadly as the pool of resources in excess of

the minimum necessary to produce a given level of output (Moe 1997), fiscal slack can also be

nonmonetary resources such as excess employees. On the spending side, fiscal slack can be

discretionary spending such as capital maintenance and travel that can be reduced easily during

recessions and in response to other fiscal shocks (Hendrick 2004). The practice of spending

surplus own-source revenues on capital projects during fiscal good times and reducing that

spending during fiscal bad times is well documented (Mattson 1994; Pagano 2002; Wang and

Hou 2009).

Government fiscal structure can also create financial risks for the government by increasing

its exposure to the negative effects of fiscal stress and volatility. For example, structural features

such as reliance on elastic revenue or state-shared revenues over which local governments have

no control make these governments more vulnerable to recessions and to decisions made by

state government to reduce the revenues they provide to local governments. Maintaining higher

levels of slack resources can counteract these effects, which demonstrates the tradeoff between

these features. Specifically, higher levels of risk in a government’s fiscal structure can be

balanced with higher levels of slack to improve adaptability, while lower levels of risk require

less slack to maintain financial condition and stabilize services delivery during volatile periods.

In effect, slack and risk represent opposite sides of the same coin in terms of their effects on

government.

Diversifying sources of revenue is widely seen as a way of reducing the effects of volatility on

government and making them less vulnerable to recessions and other negative shocks

(Shannon 1987; Pagano and Hoene 2004). This mechanism is similar to how a diversified

investment portfolio reduces the risks associated with investments. There is some evidence that

revenue diversification insulates states from economic downturns (Carroll 2005), reduces the

volatility of revenue in municipal governments (Carroll 2009), and leads to greater financial

stability in nonprofit organizations (Carroll and Stater 2009). But there is also evidence that the

Hendrick and Crawford / Municipal Fiscal Policy Space and Fiscal Structure 29

effectiveness of a more diversified financial structure depends on the way different sources of

revenue perform relative to each other (Thompson and Gates 2007). Specifically, governments

that draw more revenue from elastic and unstable sources, such as corporate income and sales

taxes, will have more volatile revenue and will be impacted more by recessions and fiscal shocks

than governments with revenues from different stable sources (Edgerton, Haughwout, and Rosen

2004; Bruce, Fox, and Tuttle 2006). There is also evidence that local governments with TELs and

home rule status have more diversified revenue structures than those without TELs and home

rule (Carroll and Johnson 2010).

MODEL, HYPOTHESES, AND VARIABLES

The prior description of how governments cope with volatility, fiscal stress, and financial

problems specifies a process in which many of the events that create these problems, such as

recessions, are exogenous to the government. Officials solve these problems by using the tools

available in the government’s fiscal toolbox that are a product of its FPS, which is also

exogenous, and fiscal structure. The choice of a tool from the fiscal toolbox will alter the

government’s current fiscal structure and may even eliminate the availability of the tool or

weaken its effect in the future, such as when a government depletes its fund balance during a

long recession. Thus, fiscal structure is relatively endogenous compared to the FPS because it

is the product of officials’ prior decisions about which tools to use to solve financial problems,

although some structural features may have more exogeneity than others. This is especially

true for features that are affected significantly by exogenous events, such as revenue

diversification and reliance on state aid that are determined, in part, by state laws and

decisions about revenues and aid to local governments. Structural features that are not likely

to change dramatically over time and that are more institutional, such as relative spending

levels for particular services, also have exogenous characteristics as do prior decisions about

debt and other financial obligations that the government cannot easily eliminate in the current

time period.

As conceptualized here, the fund balance is the most important structural feature for

managing spending volatility and it is the one over which government officials have the most

control. Thus, spending volatility and fund balance are likely to be affected by fiscal shocks in

similar ways that bias the estimate of the effect of the latter on the former and require

instrumental variables (IV) to estimate this effect. But, according to the evidence presented here,

it is also important to understand how officials manage the government’s fund balance to provide

a more comprehensive understanding of how spending volatility is managed within the system.

In this case, spending volatility is not a tool for managing the fund balance and it does not

represent a constraint on the fiscal tool box, so the former is not expected to affect the latter in the

current fiscal year. The model, therefore, is recursive and both equations should be estimated

according to what is specified theoretically rather than simply using expected values from the

fund balance equation (with instruments) to estimate the endogenous effects of this variable in

the spending volatility equation.

30 Public Budgeting & Finance / Fall 2014

Hypotheses and Relationships

Applying this perspective to the financial problem of volatility in a local government’s fiscal

environment allows us to specify a conceptual model of the process and the expected

relationships in more detail. The model assumes that the volatility of a government’s revenue

bases affect the volatility of revenues collected and, in turn, the volatility of spending unless

the government alters rates and charges on the revenue bases (Carroll and Goodman 2012). The

model also recognizes that changes in spending needs and demands will create changes in actual

spending, but assumes that these effects will be less than the effects of volatility in the revenue

bases because governments of the size examined here tend to be driven more by revenue

conditions than spending conditions (Hendrick 2011).

Local government FPS, especially restricted versus relaxed policy options, will impact

whether and how the government can modify revenues collected to compensate for changes to

revenue bases and spending needs. The model further assumes, however, that governments are

not likely to change tax rates or charges to manage these types of events. Rather, governments

will use such tools to resolve more permanent financial problems that require additional revenue

from the revenue bases. This assumption is consistent with a large body of research that shows

governments have a hierarchy of responses to fiscal stress and that tax rates are less likely to be

used to solve financial problems in the near-term (see Levine 1980; Levine, Rubin, and

Wolohojian 1981).

Based on this conceptual model and set of assumptions, this research presents the following

hypotheses about the effects of FPS and fiscal structure on spending volatility:

H1: Volatility in revenue bases will increase volatility of spending in local governments.

H2: Governments with a restrictive FPS will have higher spending volatility than governments

with a nonrestrictive FPS controlling for the volatility of the revenue bases.

H3: Irrespective of volatility in the revenue bases, local governments can mitigate spending

volatility using fund balances over which they have easy access.

H4: Total spending volatility may increase, however, if governments’ spending flexibility is

constrained due to features such as high levels of debt service and reliance on grants that

limit the ability to transfer resources in some areas to offset changing needs and demands in

other areas or reductions on the revenue side.

H5: Local governments can rely on revenue diversification to minimize the effects of volatility

in individual revenue bases on total revenues collected.

Revenue diversification represents an ex ante tool for managing spending volatility because it

directly affects the volatility of revenue collected. The fund balance, on the other hand,

represents an ex post tool that officials use to compensate for the volatility of revenue collected

Hendrick and Crawford / Municipal Fiscal Policy Space and Fiscal Structure 31

after the fact. Relative to revenue diversification and other structural features that are less

variable, officials are considered to actively manage spending volatility in the current time

period with the fund balance in the same time period. As a result, spending volatility and the

errors estimated from this equation are likely to be affected indirectly by factors that are expected

to impact fund balance directly making this variable endogenous within the system.

To determine the direct effects of other structural features and exogenous variables in the

system on fund balance and their indirect effects on spending volatility, a second equation is

estimated with fund balance as the dependent variable. To reduce the endogeneity of other

structural features in the two equations, these features are measured in the prior time period and

represent the range of options available to local government in their fiscal toolbox to manage

spending volatility in the current time period. Thus, structural features measured in the prior time

period are considered to be exogenous relative to the fund balance that is measured in the current

time period.

The conceptual model of financial problem solving presented previously suggests how

particular structural features and other conditions will affect the fund balance directly as shown

in the hypotheses below:

H6: Research on state rainy day funds indicates that local government officials should build

fund balances during fiscal good times and then draw down these funds to reduce deficits

during fiscal bad times.

H7: Although government officials maintain fund balances to accommodate external and

structural conditions that could affect spending volatility, they will trade off other sources

of slack to manage fiscal stress. Thus, local governments with sources of slack other than the

fund balance will maintain a lower fund balance.

H8, H9: In contrast, governments that depend more on elastic revenue or have a more restricted

FPS will maintain a higher fund balance.

H10: More generally, the model also suggests that government structural features will have a

greater impact on spending volatility in governments with a more restricted FPS. In other

words, governments with less restrictive FPS will rely less on features, such as fund

balance and revenue diversification, to manage spending volatility.

The Empirical Model and Variables

Based on the conceptual model proposed here, this research estimates the empirical model

shown in Table 1 for 265 municipal governments in the Chicago metropolitan region using data

from 1997 to 2009. The empirical model in the table contains two equations, and Table 2 shows

descriptive statistics for all variables in both equations. The first equation explains fund balance

as a function of percent of total revenue coming from sales taxes, percent grants and other

earmarked revenues, percent operational spending for enterprises in the prior year, percent

32 Public Budgeting & Finance / Fall 2014

TABLE 1

Basic Model, Variables, and Data

OSVit¼b0þb1RVitþb2FBitþb3FSit�1þb4RWSNitþb5CSNitþb6HRitþb7Tt(þb8FS�HRit)þeit (2)

FBit ¼ a0 þ a1FSit�1þa2RWSNit þ a3HRit þ a4Tt þ eit (1)

i ¼ municipalities (n ¼ 265); t ¼ time (n ¼ 12)

Description (label) Equation Data source

Spending and revenue volatility

Operational spending a (OSV) 2 IL Comptroller

Sales transaction receipts (sales tax base) (RV) 1, 2 IL Department of Revenue

Equalized assessed value (property tax base) (RV) 2 IL Department of Revenue

Nontax revenue a , t � 1 (RV) 2 IL Comptroller

Intergovernmental revenue a (RV) 2 IL Comptroller

Fiscal structure (FS)

% Fund balance of operational spending a (FB) 1, 2 IL Comptroller

Own-source revenue diversification (sales,

property, other tax, nontax) a , t � 1

2 IL Comptroller

% Grants and other earmarked revenue of total

revenue a , t � 1

1, 2 IL Comptroller

% Sales tax of own-source revenue a , t � 1 1 IL Comptroller

% Enterprise spending of total operational

spending in governmental a and enterprise funds,

t � 1

1 IL Comptroller

% Capital spending of total spending b , t � 1 1 IL Comptroller

% Deficit or surplus of operational spending a , t � 1 1

Fiscal policy space: revenue wealth and spending needs (RWSN)

Income per capita 1 U.S. Census, 1990 (1989), 2000

(1999), ACS 2005–2009 c

Residential EAV 1 IL Department of Revenue

Miles from Chicago (time-invariant) 1, 2 U.S. Census Gazeteer files

Population (size) 1, 2 U.S. Census, 1990, 2000, 2010 c

Change in spending needs (CSN)

% Change in population, t/t � 1 2 U.S. Census, 1990, 2000, 2010c Fiscal policy space: revenue base access and other financial privileges (HR)

Home rule municipalities 1, 2 Various sources

Time period dummy variables (T)

Fiscal bad times: 2002, 2003, 2008, 2009 2

Volatility periods: (1) 2000–2004; (2) 2005–2009 1

a Includes general fund, debt service, and special revenue funds. b Includes general fund, debt service, special revenue funds, and capital fund.

c Data for years not represented are extrapolated from years for which data is available.

Hendrick and Crawford / Municipal Fiscal Policy Space and Fiscal Structure 33

capital spending in the prior year, percent deficit or surplus in the prior year, income per capita,

percent of equalized assessed value (EAV) that is residential, miles from the city of Chicago,

population of the jurisdiction, whether the government is home rule, and whether the year

constitutes fiscal bad times. The second equation explains operational spending as a function of

all measures of revenue volatility, fund balance, own-source revenue diversification, percent

grants and other earmarked revenues, miles from the city of Chicago, population, percent change

in population, whether the government is home rule, and two dummy variables that represent

observed trends in operational spending in these governments over time. Equation (2) also

TABLE 2

Descriptive Statistics of Variables

AVPR

spending

(OSV)

AVPR sales

transaction

receipts

(RV)

AVPR

equalized

assessed

value (RV)

AVPR

nontax

revenue

(RV)

AVPR

intergovernmental

revenue (RV)

% Fund balance

of operational

spending

Mean 12 12 13 22 25 82

Median 7 6 5 12 14 64

SD 22 21 93 112 55 88

Coeff. of variation 1.80 1.83 7.30 5.04 2.24 1.07

Minimum 0 0 0 0 0 �98 Maximum 541 516 4,232 5,952 1,892 1,564

Revenue

diversification

% Grants and

earmarked

revenue

% Sales tax

of own-source

revenue

% Enterprise

spending

of total

spending

% Capital

spending of

governmental

spending

% Deficit or

surplus of

operational

spending

Mean 0.86 6 23 19 10 16

Median 0.89 5 21 18 5 9

SD 0.13 5 16 12 14 65

Coeff. of variation 0.15 0.85 0.67 0.63 1.32 4.02

Minimum 0.02 0 0 0 0 �70 Maximum 1.00 75 85 64 98 2,053

Income

per capita

% Residential

EAV of total

EAV

Miles

from

city of

Chicago Population

% Change in

population,

1997–2009

Home

rule, 2003

Mean 31,981 77 30 18,113 1.7 Home rule,

N ¼ 105, 40% Median 27,065 80 28 11,365 0.6

SD 16,003 17 14 21,464 7.5

Coeff. of variation 0.50 0.22 0.45 1.19 4.43 Nonhome rule,

N ¼ 160, 60% Minimum 8,082 2 8 89 �3.8 Maximum 113,259 100 68 186,991 355

AVPR, absolute value of residual as percent of predicted.

34 Public Budgeting & Finance / Fall 2014

estimates the interactive or conditional effects of home rule on the impact of fiscal structure on

spending volatility.

Technically, the state context attribute of the FPS that restricts or expands local governments’

options for managing fiscal stress and volatility cannot be tested on municipalities in the same

state. Additionally, the tremendous differences in state context across the 50 states with respect

to the financial privileges of local governments and local FPS makes it difficult to capture these

effects (Amiel, Deller, and Stallman 2009). These differences also create problems of sample

bias and generalizeability for any study that does not examine local governments in all states. But

variations in “home rule” privileges of the municipalities examined here provide an opportunity

to examine effects that are comparable to that of a TEL in conjunction with other state contextual

features that determine a broad range of financial options available to local governments.

In addition to the property tax, all Illinois municipalities may levy taxes on cigarettes, photo

finishing, motel occupancy, automobile rental, and utilities (telephone, natural gas, and

electricity). Home rule municipalities, however, can levy many other taxes, including an

additional sales tax and a real estate transfer tax, and they are not subject to property tax

limitations. They also have additional powers regarding contracts, regulation, and economic

development that greatly broaden their financial options in these areas. In other words, home rule

governments have a much larger fiscal toolbox for solving financial problems than nonhome rule

governments. In this context, having a larger toolbox is expected to reduce spending volatility

and increase the size of the fund balance, but these effects also are likely to be more complex than

a simple dummy variable.

Home rule status is automatically granted to municipalities in Illinois with populations greater

than 25,000, but smaller municipalities can obtain home rule status and large governments can

repeal home rule through referendum, although few large governments have done so. Thus, it is

difficult to separate the effects of home rule and population in large governments and the additive

and interactive effects of home rule that are observed here will apply only to small governments.

However, the findings can be extrapolated to large governments similar to the logic of a

regression discontinuity design in which the treatment and control are assigned based on

population size. More importantly, the impact of home rule on the size of the toolbox of smaller

governments cannot be overlooked.

Seven variables are used to represent different features of the fiscal structure of the

governments observed here. All structural features except the fund balance are lagged and

treated as exogenous in the prior year because they represent the range of options available to the

governments in the current time period. As noted in Table 1, some of these variables are

measured for all governmental funds (general, special revenue, debt service, and capital funds),

but most variables exclude capital funds because they are not operational. All structural variables

except “percent enterprise funds” that account for water and sewer revenue and spending in these

governments, exclude enterprise funds because decisions about these operations are largely

separate from decisions about governmental funds. In small governments such as these

municipalities, a significant portion of nonenterprise operational spending comes from special

revenue funds, and decisions about such spending and other aspects of fiscal structure often are

made across the three operational governmental funds. Thus, governmental funds are merged to

Hendrick and Crawford / Municipal Fiscal Policy Space and Fiscal Structure 35

capture these relationships, but both equations include a measure of percent grants and other

earmarked revenue (of total revenue) to reflect the loss of discretion and flexibility in choices

about special revenue funds (see H4).

The fund balance is measured as a percent of operational spending and excludes capital funds,

which often contain a significant level of bond proceeds that are strictly earmarked for capital

projects and can inflate the fund balance greatly. In this case,capital and enterprise expenditures as a

percentage of total spending in the prior year constitute other sources of slack resources that these

governments can reassign during fiscal bad times. For instance, governments can slow capital

spending or reassign capital dedicated revenues to operations and charge more internal services to

enterprise funds during recessions. Thus, governments with more capital and enterprise spending

will reduce their fund balance (H7). The level of surplus or deficit as a percentage of operational

spending in the prior year also directly affects the fund balance because it is added to or subtracted,

respectively, from the fund balance in the current year. Alternatively, governments will increase

their fund balance if they have a high level of sales taxes as a percentage of total revenue because of

the volatility this condition creates in the total revenues of these governments (Mikesell 1984).

Revenue diversification is the last measure of fiscal structure that is included in the model, but

only in the second equation. It is measured using a reversed Herfindahl–Hirschman Index (HHI)

that is calculated for the four primary categories of own-source revenue in these governments:

property tax, sales tax, other taxes, and nontax revenue. The equation for revenue diversification

is shown below where Rj is the fraction of revenue generated by each category.

RDit ¼ 1 � P4

j¼1 R 2 j

0:75 ð1Þ

According to the equation, the sum of the squared proportions are subtracted from 0 to reflect

diversification and not concentration (HHI) and divided by the highest level of diversification to

standardize the values between 0 and 1, where 1 indicates a high degree of revenue diversification.

According to Table 1, the model includes five variables that represent jurisdictional

characteristics that can affect either or both spending volatility and fund balance. Personal

income per capita is a general measure of government fiscal capacity and revenue wealth (Berne

and Schramm 1986; Rafuse and Marks 1991) that is likely to have a positive effect on the

government’s fund balance. Percent EAV that is residential reflects land use in the jurisdiction

and the extent to which daytime service populations are greater than the population of the

jurisdiction. Jurisdictions that are residential rather than commercial or industrial tend to have

greater spending needs but more stable revenue and spending conditions overall (Hendrick

2011), which allow governments to maintain a lower fund balance.

Because all governments in the study tend to be more revenue than spending driven

(Hendrick 2011), the effect of changes in spending needs on actual spending is expected to be

less than the effect of revenue base volatility. With respect to spending needs, these

municipalities are responsible primarily for public safety, transportation, and utilities (water and

sewer) that are greatly affected by population growth, but they are not responsible for schools or

services to the poor, elderly, or unemployed. Change in population is, therefore, expected to

36 Public Budgeting & Finance / Fall 2014

capture change in most spending needs in the spending volatility equation. Additionally,

governments that are further away from the City of Chicago (miles from Chicago) are more

likely to experience development and other changes that increase both spending volatility and

fund balance. The population of the jurisdiction also is expected to affect both spending volatility

and fund balance. Studies of organizations have long recognized that large organizations have

more flexibility and slack resources (Thompson 1967). For instance, large governments with

many employees are likely to have vacant positions that can remain open during periods of fiscal

stress without detriment to core services compared to small governments in which individual

employees are likely to perform multiple and critical duties.

Defining and Measuring Volatility

Table 1 shows that the model includes four sources of revenue volatility in addition to spending

volatility. Although only home rule governments can levy a discretionary sales tax in Illinois, a

portion of state sales taxes are distributed back to all municipalities based on point of sale. Thus,

all municipal governments in the region rely on sales taxes to some extent, and these revenues

will vary with respect to sales transaction receipts (the sales tax base) within the jurisdiction. In

fact, sales tax revenue represents almost 25 percent of all own-source revenue for these

governments on average as shown in Table 2. Property taxes are the largest source of revenue for

these governments and constitute about 33 percent of total own-source revenue, although the

volatility of the property tax revenue base (EAV) is not likely to affect the volatility of property

taxes collected. In Illinois, the property taxes requested by local governments and distributed to

them by the county are determined (extended) based on the property tax levy and not the tax rate.

Intergovernmental revenue, which is comprised mostly of state grants but also includes state

income tax that is shared with municipalities, is likely to be quite volatile due to the lumpiness of

state grants to municipal governments. About 18 percent of total municipal revenues come from

intergovernmental revenue on average, and this source of revenue is assumed to be exogenous with

respect to spending. Nontax revenue is about 30 percent of own-source revenues on average and,

therefore, is a major source of revenue. It includes licenses, permits, and miscellaneous revenue and

is likely to be volatile, especially for governments undergoing a lot of development. Unfortunately,

it is not possible to observe the individual revenue bases that comprise nontax revenue, which are

exogenous, but their combined significance requires these sources to be recognized in the model.

To reduce potential endogeneity of observing taxes collected this variable is measured using

nontax revenues collected in a prior year (t � 1). According to Kennedy (1998, 169), lagged endogenous variables can be treated as exogenous because they are determined prior to the

current period’s values of the endogenous variables. Technically, lagged endogenous variables

are “predetermined variables” and, assuming the errors in the model are not correlated over time,

their use as regressors creates reduced form estimates that are asymptotically unbiased. 1

1. All fiscal structural variables except fund balance also are lagged one year to create predetermined variables. A

Wooldridge (2002, 176) test for serial correlation of errors in panel data with instrumental variables (Davidson and

MacKinnon 1993, 369–371) was not statistically significant at the 0.05 level.

Hendrick and Crawford / Municipal Fiscal Policy Space and Fiscal Structure 37

The following equations show how volatility is measured for spending and all revenue

variables in the model using a two-step method that summarizes the variation or deviation of

actual values of each of the five fiscal features from their long-term trend. This method has been

used in many prior studies of the effects of volatility in governments and nonprofit organizations

(Hou 2003; Marlowe 2005; Carroll 2009; Wang and Hou 2009; Carroll and Goodman 2011).

Yit ¼ ai þ biTt þ eit ð2Þ

Vit ¼ Yit � Ŷ it

Ŷ it

� � � �

� � � � ð3Þ

where Yit is the financial feature for which volatility is being assessed (spending, sales

transaction receipts, EAV, nontax revenue, and intergovernmental revenue) for municipality i in

year t; ai is the constant for municipality i; bi is the trend parameter for municipality i; and Vit is

the volatility measure for municipality i in year t from 1997 to 2009.

In the first step, the values of each feature are regressed against a linear time variable (T). In

step 2, the predicted values of equation (2) are subtracted from the actual values for each feature

to ascertain the deviation of the actual values from the average change in that feature over the

observed time period. The deviations or residuals are then divided by the predicted value to

standardize them, and the absolute values of these proportions are calculated to create the final

indicator of volatility as shown in equation (3).

The descriptive statistics in Table 2 show that the volatility of intergovernmental revenue is

the highest compared to the other volatility measures with an average and median Vit of 25 and

14, respectively. Nontax revenues has the second highest volatility with an average of 22 and

median of 12, but the volatility of spending, EAV, and sales transaction receipts are very similar

with means of 12 and 13 and medians that are 5–7. Examining the median value of the volatility

measure for operational spending for each fiscal year in the data (not presented here) shows a

clear downward trend in spending volatility with different aggregate values from 1997 to 1999,

2000 to 2004, and 2005 to 2009. An examination of trends in the median revenue volatility

measures (also not presented here) indicates that none of the revenue sources have this trend.

Investigation of other trends in the data suggests that the effects of two subsequent debilitating

recessions on the municipalities after a period of fiscal good times were cumulative. The good

times of the late 1990s allowed governments to spend much more than what would be expected

from their trends for the entire time period. Each recession, then, increasingly constrained their

spending towards their overall trend from this high point. These effects are represented in

equation (2) of the model using two dummy variables to denote the last two time periods and the

constant (intercept) of the equation will reflect the effects of the first time period on spending

volatility. Additionally, equation (1) of the model contains one dummy variable that denotes the

fiscal bad times for these governments (2002, 2003, 2008, and 2009) that correspond to the two

recessions. The constant of this equation, therefore, represents the effects of fiscal good times on

the fund balance.

38 Public Budgeting & Finance / Fall 2014

Estimation, Testing, and Data

The system of two equations is estimated using two basic approaches in Stata. First, both

equations are estimated simultaneously using generalized three-stage least squares (3SLS) to

account for the endogeneity of fund balance and determine the indirect effects of FPS and fiscal

structure on spending volatility that occur through fund balance as specified by the conceptual

model. Unfortunately, the Stata routine for 3SLS does not recognize panel data or estimate

standard errors that are robust under the assumption that they are correlated over time within

units (clustered errors). Because the volatility variables are calculated as a function of values

unique to each municipal government, the variance of the residuals is likely to vary significantly

by municipal government in the pooled data.

In order to correct for this condition, recognize the panel structure of the data, and obtain an

estimate of equation (1), the two equations are estimated separately using a second approach in

which equation (2) (spending volatility) is estimated using several different IV methods.

Unfortunately, these methods do not produce an estimate of equation (1) (fund balance) without

all exogenous variables in the system. Both 3SLS and IV estimation regresses all exogenous

variables in the model on the fund balance to obtain unbiased estimates of its effect on spending

volatility, but only 3SLS provides estimates of the first equation with only the explanatory

variables as specified by the conceptual model. Equation (1) is, therefore, estimated separately in

this second approach with only these variables to determine their direct effects on the fund

balance and their indirect effects on spending volatility.

This equation is estimated using ordinary least squares (OLS) for fixed effects (FE) and no

panel effects and generalized least squares (GLS) for random effects (RE). All of these routines

recognize clustered errors. The IV estimation methods for equation (2) include two-stage least

squares (2SLS) estimated with and without a panel structure (FE and RE) and generalized

method of moments (GMM) without a panel structure. All of the IV routines except 2SLS for RE

allow for clustered errors. 2 Equation (2) also is estimated with and without the interactive effects

using all estimation methods for this equation.

This research uses multiple methods of estimating the model because each of them has

particular strengths and none of them are completely appropriate or without limitation under the

circumstances presented here. With respect to the panel data structure and the choice of FE or RE

estimation, Hausman tests of fixed versus RE were calculated for equations (1) and (2) with and

without the interaction effects. The tests show that the null hypotheses can be rejected at the

0.0001 level for both cases indicating that FE estimation is appropriate for both forms of the

model. Additionally, Greene (2000, 567) asserts that RE models are most appropriate when

the data consist of a sample of the population because the coefficients represent a random sample

all effects in the population. The analysis here is conducted on the entire population of

2. All estimation is performed with Stata with the exception of IV for 2SLS and GMM (no panel data) that is

written by Baum, Schaffer, and Stillman (2007) and IV for panel data that is written by Schaffer (2010). Although

clustering errors are considered to be desirable for all panel data (Moulton 1990; Bertrand, Duflo, and

Mullainathan 2004), the IV routine for panel data in Stata does not support clustering of errors, and the routine

written by Schaffer does not allow clustering for RE.

Hendrick and Crawford / Municipal Fiscal Policy Space and Fiscal Structure 39

municipalities in the Chicago region making interpretation of the random coefficients somewhat

problematic. The problem with FE, however, is that many variables in both equations are

relatively time-invariant (e.g. revenue diversification, income per capita, residential EAV) and

become statistically insignificant in the presence of full FE. One variable, miles from Chicago,

does not vary over time and so must be excluded as an independent variable in both equations

using FE. In this case, RE is a second best option for estimating the effects of time-invariant and

unit-invariant variables (time period dummies) despite the Hausman test results (Plümper and

Troeger 2004). Although both dependent variables are explained largely by FE, these effects are

unique to each municipality and have little theoretical relevance compared to the variables in the

conceptual model.

With respect to IV estimation, the first equation of the model presents a justification for the IV

used to estimate fund balance in the first equation and produces a predicted value that is not

correlated with the errors in the second equation. The system of two equations is overidentified and

has more IV than are needed to estimate all the parameters, which is not ideal for GMM estimation.

On the other hand, GMM has advantages over 2SLS for FE, systems of equations, and in the

presence of heteroskedasticity (Wooldridge 2001). Positive heteroskedasticity is often a significant

problem with financial data due to the positive skewness of many variables, and the high number of

zero values for enterprise and especially capital spending exacerbate this problem. Although

clustered errors are robust to heteroskedasticity (Peterson 2009), not all of the estimation methods

used here cluster or calculate robust White standard errors. Thus, most of the variables in both

equations, including the dependent variables, are transformed using natural logarithms to compress

the scales in which these variables are measured (Tukey 1977). 3 In cases where both the dependent

and independent variables are logged, the coefficients are interpreted as elasticities. Additionally,

the fiscal structural variables that are multiplied by home rule in the estimation of equation (2) with

interaction terms are centered to make the coefficients more meaningful and reduce the

multicollinearity of interactive variables with their component variables.

The estimation of the endogenous variable in equation (1) using IV also requires the

instruments to be tested for validity and weakness. If the model is overidentified as this one is, we

can test whether the instruments in the first equation are uncorrelated with the errors in

equation (2) and that the instruments are correctly excluded from this equation using the Sargon

test for 2SLS and the Hansen-J test for GMM (Murray 2006; Baum, Schaffer, and

Stillman 2007). Both tests indicate that the instruments are valid at probabilities of 0.16 and

0.47, respectively, for the model without interactions and 0.12 and 0.56 for the model with

interactions. Both of these tests are robust and consistent in the presence of heteroskedasticity.

The weakness of the instruments can also be checked by examining an F-statistic on the partial

correlation of the excluded instruments and using the rule of thumb that this value should be

greater than 10. The F-statistic is 14 for the model without interactions and 11 for the model with

interactions. 4 In addition, the IV routine that is used to calculate the estimates for 2SLS and

3. A value of 1 is added to variables that have a minimum value of 0 and 1þ minimum value is added to variables that have negative values.

4. Other tests of weak instruments yield the same conclusion.

40 Public Budgeting & Finance / Fall 2014

GMM reports a test of endogeneity that examines whether the endogenous regressors can

actually be treated as exogenous (Ho). This test shows the null hypothesis can be rejected at the

0.0000 level for both versions of the model. 5

EMPIRICAL RESULTS

Table 3 shows the results for equation (1) (fund balance) estimated independently of equation (2)

(spending volatility) as specified according to the model in Table 1. The coefficients for this

model are reported in the first four columns of the table and the probabilities associated with the

standard errors are reported in the last four columns of the table. The table shows that the results

differ significantly for FE compared to the other estimation methods. As expected, the

coefficients for FE are lower and the probabilities of statistical significance are much higher,

especially for the variables that are relatively time-invariant. Also consistent with the expected

properties of the different estimation methods, the coefficients are highest and the probabilities

are lowest for the 3SLS method.

In accordance with the second set of hypotheses, the coefficients show that the fund balance of

a municipal government is higher when fiscal times are good and the jurisdiction has higher

income per capita (not statistically significant for FE), is less residential, and is further away from

the City of Chicago. In other words, these governments tend to save when times are good and

their FPS is more munificent, but also under conditions in which spending needs are likely to be

more uncertain and less stable. Also as expected, governments of small jurisdictions accumulate

greater fund balances than governments of large jurisdictions, but this result is not statistically

significant for either RE or FE. Contrary to expectations, however, home rule has a positive

effect on fund balance, but it is statistically significant only for 3SLS estimation.

With respect to the coefficients for the fiscal structural variables, all variables have the

expected effect. Fund balances are higher in governments that have more surplus in the prior

year, greater reliance on sales taxes (not statistically significant for RE or FE), and higher levels

of grants and earmarked revenue relative to total revenue (not statistically significant for RE or

FE). Fund balances are lower in governments that have a higher level of enterprise and capital

spending relative to total spending. The negative relationships between capital and enterprise

spending in the prior year and fund balance in the current year show the tradeoffs that

governments often make between the fund balance and these two sources of revenue.

Table 4 shows the results for equation (2) estimated five ways, four of which use IV. Overall,

the coefficients for this equation are less statistically significant than equation (1) and the model

is less predictive as indicated by the low R 2 in this table. Similar to the results for the first

equation, standard errors of the estimates are lowest for the 3SLS and highest for FE estimation,

but many of the FE coefficients for equation (2) are very different than coefficients for the other

estimation methods. The results show that neither change in population nor home rule are

5. VIF collinearity factors on equations (1) and (2), which are 1.32 and 1.49, respectively, show very little problem

in this regard.

Hendrick and Crawford / Municipal Fiscal Policy Space and Fiscal Structure 41

statistically significant with any method of estimation, although the home rule coefficient and

test of significance probability for FE suggests that spending volatility may be lower for

governments with these privileges (and without controlling for miles from Chicago).

With the exception of the FE coefficients, most of the other coefficients for the other

estimation methods are statistically significant at a 0.10 level (except percent grants and

earmarked revenue and miles from Chicago for GMM). As expected according to the first set of

hypotheses, these coefficients show that spending volatility is lower when governments have a

higher fund balance, more revenue diversification, and higher population. Also as expected,

spending volatility is higher when volatility is higher for the four sources of revenue, the

TABLE 3

Equation (1): Fund Balance in Chicago Municipal Governments, 1998–2009,

No Home Rule Interactions

N ¼ 3,044 Coefficients Probability of SE (r)

1 2 3 4 2 3 4

% Fund balance (ln) 3SLS

No

panel a

Panel,

RE a

Panel,

FE a

1 (Cluster) (Cluster) (Cluster)

% Deficit or surplus

(ln), t � 1 0.2768 0.2858 0.1510 0.1355 0.00 0.00 0.00 0.00

% Sales tax (ln), t � 1 0.0351 0.0350 0.0123 �0.0029 0.00 0.09 0.62 0.93 % Grants and

earmarked revenue

(ln), t � 1

0.0559 0.0552 0.0146 0.0080 0.00 0.00 0.13 0.40

% Enterprise funds

(ln), t � 1 �0.0424 �0.0454 �0.0496 �0.0441 0.00 0.01 0.00 0.05

% Capital funds (ln),

t � 1 �0.0160 �0.0172 �0.0137 �0.0139 0.00 0.03 0.01 0.01

Income per capita (ln) 0.2202 0.2142 0.1884 0.1045 0.00 0.00 0.00 0.30

Population (ln) �0.0300 �0.0289 �0.0247 �0.0397 0.00 0.04 0.13 0.56 % Residential EAV

(ln) b

�0.1078 �0.1066 �0.1102 �0.1212 0.00 0.02 0.02 0.07

Miles from Chicago

(ln)

0.0517 0.0509 0.0798 – 0.00 0.06 0.00 –

Fiscal bad times (02,

03, 08, 09)

�0.0463 �0.0482 �0.0417 �0.0369 0.00 0.00 0.00 0.00

Home rule 0.0410 0.0402 0.0192 0.0186 0.00 0.13 0.47 0.63

Constant, fiscal good

times

2.17 2.20 3.08 4.50 0.00 0.00 0.00 0.00

R 2

0.25 0.25 0.23 0.19

a Single equation estimation. b Omitted due to collinearity in fixed effects.

42 Public Budgeting & Finance / Fall 2014

T A B L E 4

E q u a ti o n (2 ):

O p e r a ti o n a l S p e n d in g in

C h ic a g o M u n ic ip a l G o v e r n m e n ts , 1 9 9 8 – 2 0 0 9 , N o H o m e R u le

In te r a c ti o n s

N ¼ 3 ,0 4 4

C o e ff ic ie n ts

P r o b a b il it y o f S E (r )

1 2

3 4

5 2

3 5

O p e r a ti o n a l sp e n d in g

v o la ti li ty

(l n )

3 S L S

IV N o

p a n e l

IV G M M

P a n e l IV

,

R E

P a n e l IV

,

F E b

1 (C

lu st e r )

(C lu st e r )

4 (C

lu st e r )b

V o la ti li ty

sa le s

tr a n sa c ti o n re c e ip ts

(l n )

0 .0 0 0 2

0 .0 0 0 2

0 .0 0 0 2

0 .0 0 0 2

0 .0 0 0 1

0 .0 1

0 .0 0

0 .0 0

0 .0 3

0 .2 4

V o la ti li ty

E A V

(l n )

0 .0 3 6 7

0 .0 3 5 9

0 .0 3 6 7

0 .0 3 7 0

0 .0 3 5 3

0 .0 4

0 .1 0

0 .0 8

0 .0 5

0 .0 9

V o l. in te rg o v t. re v e n u e

(l n )

0 .0 6 3 9

0 .0 5 9 1

0 .0 5 1 9

0 .0 4 1 4

0 .0 2 4 0

0 .0 0

0 .0 1

0 .0 1

0 .0 2

0 .2 6

V o l. n o n ta x re v e n u e

(l n ), t � 1

0 .0 8 3 1

0 .0 8 4 9

0 .0 7 8 0

0 .0 4 4 4

0 .0 0 7 8

0 .0 0

0 .0 0

0 .0 0

0 .0 2

0 .6 9

% F u n d b a la n c e (l n )

�0 .5 1 7 1

�0 .5 4 6 0

�0 .4 6 1 5

�0 .4 8 8 2

�0 .2 0 2 3

0 .0 0

0 .0 3

0 .0 6

0 .0 4

0 .6 4

R e v e n u e d iv e rs if ic a ti o n ,

t � 1

�0 .7 3 2 1

�0 .9 1 1 2

�0 .9 8 3 5

�0 .6 9 7 0

�0 .2 5 7 1

0 .0 0

0 .0 2

0 .0 1

0 .0 0

0 .4 3

% G ra n ts a n d e a rm

a rk e d

re v e n u e (l n ), t � 1

0 .0 9 0 1

0 .0 8 9 1

0 .0 6 0 9

0 .0 6 9 9

0 .0 2 4 0

0 .0 0

0 .0 5

0 .1 6

0 .0 4

0 .5 6

% C h a n g e p o p (l n )

0 .0 6 9 7

0 .0 3 6 4

0 .0 8 8 7

0 .0 4 5 5

0 .0 5 3 5

0 .1 9

0 .6 7

0 .2 8

0 .4 8

0 .5 2

P o p u la ti o n (l n )

�0 .2 2 3 9

�0 .2 1 7 1

�0 .2 2 3 0

�0 .2 3 6 4

�0 .4 9 4 9

0 .0 0

0 .0 0

0 .0 0

0 .0 0

0 .0 0

M il e s fr o m

C h ic a g o a

(l n )

0 .1 8 1 8

0 .1 9 4 8

0 .1 4 1 4

0 .1 9 9 0

– 0 .0 0

0 .0 4

0 .1 1

0 .0 2

–

2 0 0 0 – 2 0 0 4

�0 .5 0 9 3

�0 .5 5 7 1

�0 .5 5 9 8

�0 .5 5 7 2

�0 .5 0 4 6

0 .0 0

0 .0 0

0 .0 0

0 .0 0

0 .0

2 0 0 5 – 2 0 0 9

�0 .2 4 9 7

�0 .2 7 2 8

�0 .2 7 0 4

�0 .2 6 8 1

�0 .2 3 6 4

0 .0 0

0 .0 0

0 .0 0

0 .0 0

0 .0 0

H o m e ru le

0 .0 1 5 3

0 .0 1 0 5

0 .0 1 6 5

�0 .0 1 9 7

�0 .2 6 2 9

0 .7 6

0 .8 9

0 .8 3

0 .7 7

0 .1 0

C o n st a n t, 1 9 9 8 – 1 9 9 9

6 .1 3

6 .4 2

6 .2 8

6 .2 8

– 0 .0 0

0 .0 0

0 .0 0

0 .0 0

–

R 2

0 .1 2

0 .1 1

0 .1 2

0 .1 2

0 .0 5

a O m it te d d u e to

c o ll in e a ri ty

in fi x e d e ff e c ts .

b C o n st a n t is n o t re p o rt e d w it h th is ro u ti n e .

Hendrick and Crawford / Municipal Fiscal Policy Space and Fiscal Structure 43

government uses more grants and earmarked revenue, and the jurisdiction is further

from Chicago. For FE estimation, however, only four variables are statistically significant in

Table 4—EAV volatility, population, and the two dummy variables for volatility years.

As with Table 4, only the four variables indicated are statistically significant for FE estimation

in Table 5 with the negative effects of home rule shown to be much greater than for the other

estimation methods. For the other estimation methods, however, coefficients for individual

variables are similar across the models presented in the two tables, and the coefficients for the

variables that are multiplied by home rule are different compared to coefficients for the same

variables in Table 4, although some differences are not statistically significant. Additionally,

some of the variables in Table 5 are also less statistically significant than what is shown in

Table 4. For all estimation methods except FE, the probabilities for statistical significance in

Table 5 show that home rule is not likely to have an interactive effect with percent grants and

earmarked revenue, but may have an interactive effect with revenue diversification. These

probabilities also show that home rule has a statistically significant interactive effect with fund

balance at the 0.10 level for 2SLS and GMM only. These results indicate that revenue

diversification has a much smaller and even a slight positive effect on spending volatility in home

rule governments compared to nonhome rule governments where the effect is negative. For

2SLS and GMM, fund balance has a weaker negative effect on spending volatility in home rule

governments, which is consistent with H10.

DISCUSSION AND CONCLUSION

Depending on which estimation results are examined, conclusions about what factors affect the

spending volatility and fund balances of municipal governments in the Chicago metropolitan

area can vary greatly. The results for FE indicate that both dependent variables are determined by

events that are unique to each municipality and cannot be explained by the independent variables

identified here or any of the factors that have been identified in prior research (e.g. revenue

diversification). On the other hand, the results for the other estimation methods demonstrate that

how these governments manage spending volatility using the fund balance is complex and linked

with governments’ other fiscal structural features and FPS. Although the independent variables

do not explain a great deal of the variance of either spending volatility or fund balance in any of

the models estimated with methods other than FE, these results show that fiscal structure,

revenue wealth, spending needs, and revenue volatility affect the two dependent variables in the

manner expected. More importantly, the results show that some of these governments’ fiscal

structural features and FPS attributes have an indirect impact on spending volatility via their

direct impact on the fund balance. On the other hand, other structural features and FPS attributes

of these governments directly affect spending volatility.

The results for estimation methods other than FE also show that the impact of strict versus

relaxed fiscal policy options as measured by home rule status is not linear, as is often presented in

research on the impact of TELS and state-level constraints on local government fiscal policies.

Rather, this variable, which measures the level of constraints on governments’ policy options

44 Public Budgeting & Finance / Fall 2014

T A B L E 5

E q u a ti o n (2 ): S p e n d in g V o la ti li ty

in C h ic a g o M u n ic ip a l G o v e r n m e n ts , 1 9 9 8 – 2 0 0 9 , H o m e R u le

In te r a c ti o n s a n d C e n te r e d

N ¼ 3 ,0 2 2

C o e ff ic ie n ts

P r o b a b il it y o f S E (r )

1 2

3 4

5 2

3 5

O p e r a ti o n a l sp e n d in g

v o la ti li ty

(l n )

3 S L S

(H R )

IV N o

p a n e l (H

R )

IV G M M

(H R )

P a n e l IV

,

R E (H

R )

P a n e l IV

,

F E (H

R )

1 (C

lu st e r )

(C lu st e r )

4 (C

lu st e r )

V o la ti li ty

sa le s

tr a n sa c ti o n re c e ip ts

(l n )

0 .0 0 0 2

0 .0 0 0 2

0 .0 0 0 2

0 .0 0 0 2

0 .0 0 0 1

0 .0 0

0 .0 0

0 .0 0

0 .0 2

0 .2 2

V o la ti li ty

e q . a ss e ss e d

v a lu e (l n )

0 .0 3 3 9

0 .0 3 2 2

0 .0 3 2 5

0 .0 3 5 6

0 .0 3 5 4

0 .0 6

0 .1 5

0 .1 3

0 .0 6

0 .0 9

V o la ti li ty

in te rg o v t.

re v e n u e (l n )

0 .0 6 2 6

0 .0 5 7 9

0 .0 5 1 3

0 .0 4 1 8

0 .0 2 4 4

0 .0 0

0 .0 1

0 .0 2

0 .0 2

0 .2 5

V o la ti li ty

n o n ta x

re v e n u e (l n ), t � 1

0 .0 8 0 5

0 .0 8 1 2

0 .0 7 7 3

0 .0 4 2 2

0 .0 0 7 7

0 .0 0

0 .0 0

0 .0 0

0 .0 2

0 .7 0

% F u n d b a la n c e (l n )

�0 .4 7 5 1

(� 0 .7 9 )

�0 .7 9 2 4

(� 0 .0 5 )

�0 .6 5 1 5

(� 0 .0 6 )

�0 .6 0 5 2

(� 0 .1 4 )

�0 .1 8 9 8

(� 0 .1 9 )

0 .0 4

0 .0 3

0 .0 5

0 .0 5

0 .7 4

% F u n d b a la n c e

(l n ) � h m ru le

�0 .3 1 7 6

0 .7 4 0 5

0 .5 9 4 8

0 .4 6 6 2

�0 .0 0 1 2

0 .2 2

0 .0 5

0 .1 0

0 .1 4

1 .0 0

R e v e n u e

d iv e rs if ic a ti o n , t � 1

�1 .1 1 7 9

(0 .0 3 )

�1 .6 8 0 2

(0 .2 2 )

�1 .6 0 3 6

(0 .1 7 )

�1 .0 7 1 6

(0 .1 2 )

�0 .2 9 0 9

(� 0 .0 1 )

0 .0 0

0 .0 0

0 .0 0

0 .0 0

0 .4 3

R e v e n u e

d iv e rs if ic a ti o n ,

t � 1 � h m ru le

1 .1 5 2 8

1 .9 0 5 9

1 .7 7 5 6

1 .1 8 6 9

0 .2 7 9 2

0 .0 0

0 .0 0

0 .0 1

0 .0 1

0 .6 6

% G ra n ts

a n d

e a rm

a rk e d re v e n u e

(l n ), t � 1

0 .0 8 4 5

(0 .0 9 )

0 .1 1 1 5

(0 .0 8 )

0 .0 8 6 9

(0 .0 5 )

0 .0 7 0 4

(0 .0 9 )

0 .0 0 5 0

(0 .0 6 )

0 .0 5

0 .0 7

0 .1 5

0 .1 3

0 .9 3

(c o n ti n u e d )

Hendrick and Crawford / Municipal Fiscal Policy Space and Fiscal Structure 45

T A B L E 5

(C o n ti n u e d )

N ¼ 3 ,0 2 2

C o e ff ic ie n ts

P r o b a b il it y o f S E (r )

1 2

3 4

5 2

3 5

O p e r a ti o n a l sp e n d in g

v o la ti li ty

(l n )

3 S L S

(H R )

IV N o

p a n e l (H

R )

IV G M M

(H R )

P a n e l IV

,

R E

(H R )

P a n e l IV

,

F E (H

R )

1 (C

lu st e r )

(C lu st e r )

4 (C

lu st e r )

% G ra n ts

a n d

e a rm

a rk e d re v e n u e

(l n ), t � 1 � h m ru le

0 .0 1 2 6

�0 .0 2 8 7

�0 .0 3 7 4

0 .0 1 7 4

0 .0 5 3 3

0 .8 4

0 .7 4

0 .6 6

0 .8 0

0 .4 9

% C h a n g e p o p (l n )

0 .0 7 3 6

0 .0 2 8 5

0 .0 8 8 1

0 .0 3 9 1

0 .0 6 1 5

0 .1 7

0 .7 6

0 .3 1

0 .5 5

0 .4 8

P o p u la ti o n (l n )

�0 .2 3 3 2

�0 .2 2 1 9

�0 .2 2 2 2

�0 .2 4 4 1

�0 .5 1 2 7

0 .0 0

0 .0 0

0 .0 0

0 .0 0

0 .0 0

M il e s fr o m

C h ic a g o

(l n )

0 .1 6 8 0

0 .1 8 7 5

0 .1 3 7 3

0 .1 8 7 4

– 0 .0 1

0 .0 5

0 .1 3

0 .0 3

–

2 0 0 0 – 2 0 0 4

�0 .5 1 1 7

�0 .5 5 5 1

�0 .5 5 9 4

�0 .5 5 2 1

�0 .4 9 9 7

0 .0 0

0 .0 0

0 .0 0

0 .0 0

0 .0

2 0 0 5 – 2 0 0 9

�0 .2 4 7 1

�0 .2 6 7 3

�0 .2 6 7 4

�0 .2 6 2 8

�0 .2 3 3 6

0 .0 0

0 .0 0

0 .0 0

0 .0 0

0 .0 0

H o m e ru le

0 .0 3 1 7

0 .0 4 7 4

0 .0 4 5 9

0 .0 0 3 7

�0 .2 4 9 6

0 .5 3

0 .5 5

0 .5 5

0 .9 6

0 .1 2

C o n st a n t, 1 9 9 8 – 1 9 9 9

3 .0 9

7 .0 9

6 .4 8

3 .3 7

0 .0 0

0 .0 0

0 .0 0

0 .0 0

R 2

0 .1 2

0 .1 1

0 .1 2

0 .1 2

0 .0 5

46 Public Budgeting & Finance / Fall 2014

and fiscal toolbox, alters the impact of some structural features on spending volatility. Based on

theory and prior research, however, one would expect such policy constraints to have a more

significant impact on how governments respond to environmental threats and conditions. But the

lack of an observed linear effect of home rule on either the fund balance or spending volatility

and the weakness of the observed conditional relationships for spending volatility indicate that

these relationships need to be more fully investigated and better specified in the future.

The picture of financial problem solving at the local level that is presented and observed here

is much more multifarious than what has been examined previously and demonstrates the

importance of fiscal structure to a government’s fiscal toolbox. Specifically this research shows

that governments are not completely bound or constrained by their external environment in

solving financial problems. Rather, governments have resources and options for managing fiscal

stress and volatility within their internal environment and that these features can act as filters or

buffers on external events.

The results also suggest that a government’s fiscal structure represents a set of moving parts

that are likely to be used to solve multiple problems at the same time. Much of the research that

investigates the impact of structural features such as revenue diversification, fund balance, and

state contextual features such as TELS on financial problem solving by governments does not

capture this complexity. This dynamic is not easy to model, however, due to the inherent

endogeneity of the attributes of government fiscal structure and their strong correlation with

relatively time-invariant unit characteristics. Furthermore, empirical models of volatility are

hard to estimate accurately given the limitations of measuring this concept. Volatility, by its very

nature and dictionary definition, implies something that is hard to predict and explain. Thus, it is

not surprising that the explained variance for this model is low similar to prior research that also

models government volatility and fund balances.

More generally, this study suggests that research on the subject of financial problem solving

needs to be guided by a conceptual model that recognizes the reciprocal and conditional nature of

these relationships. This study has proposed a conceptual model that uses the broad and distinct

constructs of FPS and fiscal structure to explain how local governments solve financial problems.

This study also has defined and applied fiscal structure in a more comprehensive manner than has

been done in the past, but more conceptual work needs to be done to flesh out how qualities of the

external environment affect the internal environment and how both dimensions affect the

decisions local governments make in the short-run. Although only one type of financial problem

was observed here, spending volatility, this conceptual model could easily be applied to other

types of financial problems such as fiscal stress that might be easier to conceptualize and

measure.

REFERENCES

Amiel, Lindsay, Steven Deller, and Judith Stallman. 2009. “The Construction of a Tax and Expenditure

Limitation Index for the US.” Department of Agricultural and Applied Economics Staff Paper No. 536.

University of WI-Madison.

Hendrick and Crawford / Municipal Fiscal Policy Space and Fiscal Structure 47

Baum, Christopher F., Mark E. Schaffer, and Steven Stillman. 2007. IVREG2: Stata Module for Extended

Instrumental Variables: SLS, GMM and AC/HAC, LIML, and k-class Regression. Boston College

Department of Economics, Statistical Software Components, S425401. Accessed July 8, 2013. http://

ideas.repec.org/c/boc/bocode/s425401.html

Berne, Robert and Richard Schramm. 1986. The Financial Analysis of Governments. Englewood Cliffs, NJ:

Prentice Hall.

Bertrand, Marianne, Esther Duflo, and Sendhil Mullainathan. 2004. “How Much Should We Trust

Differences-in-Differences Estimates.” Quarterly Journal of Economics. 119 (1): 249–275.

Brown, Tom. 2000. “Constitutional Tax and Expenditure Limitation in Colorado: The Impact on Municipal

Government.” Public Budgeting & Finance. 20 (3): 29–50.

Bruce, Donald, William F. Fox, and Mark H. Tuttle. 2006. “Tax Base Elasticities: A Multi-State Analysis of

Long-Run and Short-Run Dynamics.” Southern Economic Journal. 73 (2): 315–341.

Brunori, David, Michael Bell, Joseph Cordes, and Bing Yuan. 2008. “Tax and Expenditure Limits and Their

Effects on Local Finances in Urban Areas.” In Urban and Regional Policy and Its Effects. Volume 1,

edited by Margery Austin Turner, Howard Wial, and Harold Wolman, 109–155. Washington, D.C.:

Brookings Institution Press.

Carroll, Deborah A. 2005. “Are State Governments Prepared for Fiscal Crises? A Look at Revenue

Diversification during the 1990s.” Public Finance Review. 33 (5): 575–602.

———. 2009. “Diversifying Municipal Government Revenue Structures: Fiscal Illusion or Instability?”

Public Budgeting & Finance. 29 (1): 27–48.

Carroll, Deborah A. and Christopher B. Goodman. 2011. “The Effects of Assessment Quality on Revenue

Volatility.” Public Budgeting & Finance. 31 (1): 76–94.

———. 2012. “Understanding and Measuring Revenue Elasticity, Volatility, and Implications for Local

Government Fiscal Health.” In Handbook of Local Government Fiscal Health, edited by H. Helisse

Levine, Jonathan B. Justice, and Eric Scorsone, 225–250. Burlington, MA: Jones and Bartlett.

Carroll, Deborah A. and Terri Johnson, 2010. “Examining Small Town Revenues: To What Extent Are They

Diversified?” Public Administration Review. 70 (2): 223–235.

Carroll, Deborah A. and Keely Jones Stater. 2009. “Revenue Diversification in Nonprofit Organizations: Does

it Lead to Financial Stability?” Journal of Public Administration Research and Theory 19 (4): 947–956.

Cornia, Gary C., and Lawrence C. Walters. 2006. “Full Disclosure: Controlling Property Tax Increases during

Periods of Increasing Housing Values.” National Tax Journal. 59 (3): 735–749.

Cyert, Richard M. and James G. March. 1963. A Behavioral Theory of the Firm. Upper Saddle River, NJ:

Prentice Hall.

Davidson, Russell and James G. MacKinnon. 1993. Estimation and Inference in Econometrics. New York,

NY: Oxford University Press.

Douglas, James W. and Ronald Keith Gaddie. 2002. “State Rainy Day Funds and Fiscal Crises: Rainy Day

Funds and the 1990–1991 Recession Revisited.” Public Budgeting & Finance. 22 (1): 19–30.

Edgerton, Jesse, Andrew F. Haughwout, and Rae Rosen. 2004. “Institutions, Tax Structure and State–Local

Fiscal Stress.” National Tax Journal 57 (1): 147–158.

Gamage, David. 2010. “Preventing State Budget Crises: Managing the Fiscal Volatility Problem.” California

Law Review. 98: 749.

Greene, William H. 2000. Econometric Analysis. Fourth Edition. Upper Saddle River, NJ: Prentice Hall.

Hendrick, Rebecca. 2004. “Assessing and Measuring the Fiscal Health of Local Governments: Focus on

Chicago Suburban Municipalities.” Urban Affairs Review. 40 (1): 78–114.

———. 2006. “Role of Slack in Managing Local Government Fiscal Conditions.” Public Budgeting &

Finance. 26 (1): 14–46.

———. 2011. Managing the Fiscal Metropolis: The Financial Polices, Practices and Health of

Municipalities. Washington, D.C.: Georgetown University Press.

48 Public Budgeting & Finance / Fall 2014

Hou, Yilin. 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.

———. 2005. “Fiscal Reserves and State Own-Source Expenditure in Downturn Years.” Public Finance

Review. 33 (1): 117–144.

———. 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.

Hou, Yilin, and Donald P. Moynihan, 2008. “The Case for Countercyclical Fiscal Capacity.” Journal of Public

Administration Research & Theory. 18 (1): 139–159.

Jenny, Nicholas W. 2003. “The Fiscal Crisis in the States: How the State Went From Boom to Bust.” Business

Economics. 38 (3): 52–57.

Jordan, Meagan M. 2003. “Punctuations and Agendas: A New Look at Local Government Budget

Expenditures.” Journal of Policy Analysis and Management. 22 (3): 345–360.

Kennedy, Peter. 1998. A Guide to Econometrics. Fourth Edition. Cambridge, MA: MIT Press.

Krane, Dale. 2004. “The State of American Federalism, 2003–2004: Polarized Politics and Federalist

Principles.” Publius. 34 (3): 1–53.

Lav, Iris J. and Alan Berube. 1999. “When It Rains It Pours: A Look at the Adequacy of State Rainy Day Funds

and Budget Reserves.” State Tax Notes, May 17: 1639–1660.

Levine, Charles. 1980. “Organizational Decline and Cutback Management.” In Managing Fiscal Stress,

edited by Charles Levine, 3–30. New York, NY: Chatham House.

Levine, H. Charles, Irene S. Rubin, and George Wolohojian. 1981. The Politics of Retrenchment: How Local

Governments Manage Fiscal Stress. Beverly Hills, CA: Sage Publications.

Levinson, Arik. 1998. “Balanced Budgets and Business Cycles: Evidence From the States.” National Tax

Journal. 51 (4): 715–732.

Marlowe, Justin. 2005. “Fiscal Slack and Counter-Cyclical Expenditure Stabilization: A First Look at the

Local Level.” Public Budgeting & Finance. 25 (3): 48–72.

Martell, Christine R. and Paul Teske. 2007. “Fiscal Management Implications of the TABOR Bind.” Public

Administration Review. 67 (4): 673–687.

Mattson, Gary A. 1994. “Retrenchment and Fiscal Policy Planning: The Political Culture of Small Southern

Towns.” Public Productivity and Management Review. 17 (3): 265–279.

Mikesell, John L. 1984. “Title: The Cyclical Sensitivity of State and Local Taxes.” Public Budgeting &

Finance. 84 (4): 32–39.

Moe, Terry. 1997. “The Positive Theory of Public Bureaucracy.” In Perspectives in Public Choice, edited by

Dennis C. Mueller, 455–480. Cambridge, MA: Cambridge University Press.

Moulton, Brent R. 1990. “An Illustration of a Pitfall in Estimating the Effects of Aggregate Variables on Micro

Units.” The Review of Economics and Statistics. 72 (2): 334–338.

Mullins, Daniel R. and Bruce A. Wallin. 2004. “Tax and Expenditure Limitations: Introduction and

Overview.” Public Budgeting & Finance. 24 (4): 2–15.

Murray, Michael P. 2006. “Avoiding Invalid Instruments and Coping With Weak Instruments.” Journal of

Economic Perspectives. 20 (4): 111–132.

Navin, John C. and Leo J. Navin. 1997. “The Optimal Size of Countercyclical Budget Stabilization Funds: A

Case Study of Ohio.” Public Budgeting & Finance. 17 (2): 114–127.

Pagano, Michael A. 2002. “Municipal Capital Spending during the Boom.” Public Budgeting & Finance. 22

(2): 1–20.

Pagano, Michael A. and Christopher Hoene. 2004. City Fiscal Conditions in 2004. Washington, D.C.:

National League of Cities.

———. 2010. “States and the Fiscal Policy Space of Cities.” In The Property Tax and Local Autonomy, edited

by Michael Bell, David Brunori, and Joan Youngman, 243–284. Cambridge, MA: Lincoln Institute of

Land Policy.

Hendrick and Crawford / Municipal Fiscal Policy Space and Fiscal Structure 49

Peterson, Mitchell A. 2009. “Estimating Standard Errors in Finance Panel Data Sets: Comparing

Approaches.” The Review of Financial Studies. 22 (1): 435–448.

Pew Center on the States. 2010. State of the States 2010: How Recession Might Change States. Washington,

DC: The Pew Charitable Trusts.

Plümper, Thomas, and Vera E. Troeger. 2004. “The Estimation of Time-Invariant Variables in Panel Analyses

With Unit Fixed Effects.” Paper presented at the annual meeting of the Society for Political

Methodology, Palo Alto, July 29–31.

Pollock, Richard and Jack P. Suyderhoud. 1986. “The Role of Rainy Day Funds in Achieving Fiscal Stability.”

National Tax Journal. 39 (4): 485–497.

Rafuse, Robert W. Jr, and Laurence R. Marks. 1991. A Comparative Analysis of Fiscal Capacity, Tax Effort,

and Public Spending among Localities in the Chicago Metropolitan Region. Washington, D.C.: U.S.

Advisory Commission on Intergovernmental Relations.

Schaffer, Mark E. 2010. XTIVREG2: Stata Module to Perform Extended IV/2SLS, GMM and AC/HAC,

LIML and k-class Regression for Panel Data Models. Accessed October 27, 2012. http://ideas.repec.org/

c/boc/bocode/s456501.html

Shadbegian, Ronald J. 1999. “The Effect of Tax and Expenditure Limitations on the Revenue Structure of

Local Governments, 1962–87.” National Tax Journal. 52 (2): 221–237.

Shannon, John P. 1987. “State Revenue Diversification—The Search For Balance.” In The Quest for Balance

in State-Local Revenue Structures, edited by Frederick D. Stocker, 9–37. Cambridge, MA: Lincoln

Institute of Land Policy.

Simon, Herbert. 1973. “The Organization of Complex Systems.” In Hierarchy Theory: The Challenge of

Complex Systems. edited by Howard Hunt Patee, 1–28. New York, NY: George Braziller.

Sobel, Russell S. and Russell G. Holcombe. 1996. “The Impact of State Rainy Day Funds in Easing State

Fiscal Crises during the 1990–91 Recession.” Public Budgeting & Finance. 16 (3): 28–48.

Springer, Job D., Aaron K. Lusby, John C. Leatherman, and Allen M. Featherstone. 2009. “An Evaluation of

Alternative Tax and Expenditure Limitation Policies on Kansas Local Governments.” Public Budgeting

& Finance. 29 (2): 48–70.

Thompson, James. 1967. Organizations in Action. New York, NY: McGraw Hill.

Thompson, Fred, and Bruce L. Gates. 2007. “Betting on the Future With a Cloudy Crystal Ball? How Financial

Theory Can Improve Revenue Forecasting and Budgets in the States.” Public Administration Review. 67

(5): 825–836.

Tukey, John W. 1977. Exploratory Data Analysis. Reading, MA: Addison-Wesley.

Wang, Wen and Yilin Hou. 2009. “Pay-as-You-Go Financing and Capital Outlay Volatility: Evidence From

the States Over Two Recent Economic Cycles.” Public Budgeting & Finance. 29 (4): 90–107.

Weiner, David. 2004. “Where Did the Revenues Go? or What Goes Up.” National Tax Journal. 57 (3): 513–

523.

Wolkoff, Michael. 1987. “An Evaluation of Municipal Rainy Day Funds.” Public Budgeting & Finance. 17

(2): 52–63.

Wooldridge, Jeffrey M. 2001. “Applications of Generalized Method of Moments Estimation.” The Journal of

Economic Perspectives. 15 (4): 87–100.

———. 2002. Econometric Analysis of Cross Section and Panel Data. Cambridge, MA: The MIT Press.

50 Public Budgeting & Finance / Fall 2014

Copyright of Public Budgeting & Finance is the property of Wiley-Blackwell and its content may not be copied or emailed to multiple sites or posted to a listserv without the copyright holder's express written permission. However, users may print, download, or email articles for individual use.