Wenchi Wei

Renmin University of China

Municipal Structure Matters: Evidence from Government Research Article

Fiscal Performance

Abstract: Evidence for the fundamental presumption that municipal structure matters for government performance Wenchi Wei is assistant professor in
the School of Public Administration and
is smaller and weaker than many might expect. This research contributes evidence to the literature by focusing on how Policy at Renmin University of China. He
municipal structure affects government fiscal performance. Municipal structure is operationalized by constructing an holds a doctoral degree in public policy
index on a political-administrative scale based on seven essential structural characteristics. As the index score increases, and administration from the Martin School
of the University of Kentucky. His research
the municipal structure becomes less political and more administrative or professional. Theoretically, municipal focuses on public management and
structure can affect government performance by influencing managerial professionalism, managerial strategy stance, performance, public and nonprofit financial
and the relative attention of government officials to managerial accountability and efficiency. The empirical evidence management, and policy evaluation.
His research has appeared in Public
shows that municipalities adopting a structure with an imbalance toward administrative characteristics are more Administration Review, the American
likely to achieve the best fiscal conditions in cash solvency, dependence on intergovernmental revenues, and outstanding Review of Public Administration, and
debt level. Moreover, municipal structure moderates the influence of external environmental factors on government Public Finance Review, among other
journals.
fiscal conditions. Email: weiwenchi@ruc.edu.cn

Evidence for Practice

• The prevalent cross-adoption of structural characteristics among municipalities in recent decades has
made the dichotomous classification of primary municipal structure as mayor-council or council-manager
problematic.
• A continuous index can be constructed based on the essential structural characteristics of mayor-council and
council-manager governments to measure municipal structure on a political-administrative scale.
• Municipalities adopting a structure with an imbalance toward administrative or professional characteristics
are more likely to achieve the best fiscal conditions in cash solvency, dependence on intergovernmental
revenues, and outstanding debt level.
• Adjusting structural characteristics to be more administrative or professional is one option for municipal
officials to improve fiscal performance and, potentially, government performance in other respects.

M
unicipal structure is a basic institutional
factor that organizes local politics and
influences public policy formulation
and implementation (Judd and Swanstrom 2015;
Wheeland, Palus, and Wood 2014; Zhang 2014;
Zhang and Feiock 2010).1 Most U.S. municipalities
use either the mayor-council or the council-manager
form of government.2 The critical characteristics of
mayor-council municipalities, as summarized by the
National League of Cities, include the following3: the
mayor is separate from the council and elected directly
by voters; the mayor has significant administrative
and budgetary powers4; the council is elected and
responsible for legislation; a professional manager is
appointed in some mayor-council municipalities to
share limited administrative authority. The essential
characteristics of council-manager municipalities
include the following: the council has legislative
powers and retains ultimate responsibility for the
government; the council appoints a professional
manager as the chief administrative officer (CAO)
to manage administrative affairs; the mayor is
the presiding officer of the council and handles
ceremonial duties.5
Cross-adoption of structural characteristics among
mayor-council and council-manager municipalities
has been prevalent since the 1950s (Choi, Feiock,
and Bae 2013), motivating scholars to offer various
reclassifications of municipal form or structure
that recognize these modifications. One of the
contributions of this research is the introduction
of an index on a political-administrative scale to
reclassify municipal structure based on several
essential structural characteristics linked to
managerial professionalism, separation of powers Public Administration Review,
Vol. 00, Iss. 00, pp. 1–14. © 2020 by
and checks and balances between officials, and local The American Society for Public Administration.
electoral systems. DOI: 10.1111/puar.13183.

Municipal Structure Matters  1

A long-running literature has attempted to evaluate the effect of
municipal structure on government management and performance.
Carr (2015) systematically reviews the literature and concludes
that the evidence for the fundamental assumption that municipal
structure makes a difference in government performance is
smaller and weaker than many might expect. Carr encourages
scholars to advance the research by developing theories to explain
why municipal structure matters for government performance
and by gathering empirical evidence to assess the theories. This
article follows Carr’s recommendations by theoretically linking
municipal structure to government performance and by empirically
investigating the effect of municipal structure on government
performance in terms of fiscal management.
The primary research question is whether and how municipal
structure affects government fiscal conditions. The first section
introduces the related literature. The following section constructs a
political-administrative index to reclassify municipal structure, and
then hypotheses based on the theoretical analyses are proposed. The
next sections introduce the methodology and discuss the empirical
findings, respectively, and the last section concludes and discusses.
Prior Works on Government Fiscal Conditions
Studies in the related literature focus primarily on how government
fiscal conditions are measured, the validity and reliability of fiscal
condition indicators, and the determinants of government fiscal
conditions. The most frequently cited indicators of government
fiscal conditions include the Fiscal Trend Monitoring System by
the International City/County Management Association (ICMA)
(Groves, Godsey, and Shulman 1981; Nollenberger 2003); Brown’s
10-point test (Brown 1993; Maher and Nollenberger 2009); the
four-solvency approach by Wang, Dennis, and Tu (2007); and the
10-point scale approach by Kloha, Weissert, and Kleine (2005).
These indicators capture a comprehensive picture of government
fiscal conditions by examining revenues, expenditures, budgetary
balance, outstanding debt, and government assets and liabilities.
McDonald and Bruce (2015) construct a fiscal condition
determinant model in an open system framework that involves
the explanatory factors of politics, economics, demographics,
and government characteristics. The empirical results for Florida
counties show that local political ideology, educational attainment,
income, racial composition of citizens, unemployment rate,
government revenue sources, population size, and charter form
influence government fiscal conditions. Gorina, Maher, and Joffe
(2018, 81) construct an action-based measure of local government
fiscal distress and operationalize fiscal distress as “actions, often
disruptive and politically unpopular, that a government takes
because it is unable to meet its fundamental operating needs and
service requirements.” They use cash solvency, budgetary solvency,
and long-term solvency to predict governments’ disruptive and
politically unpopular actions taken to mitigate fiscal distress.
Jimenez (2017) investigates the effect of managerial networking on
local government fiscal health from a perspective of management-
performance linkage. He finds an inverted U-shaped relationship
between city managers’ networking and governments’ budgetary
solvency. Specifically, city managers’ ties with external stakeholders
help improve budgetary solvency initially. However, frequent
2  Public Administration Review  •  xxxx | xxxx 2020
interactions with external stakeholders harm managers’ decision-
making ability and exacerbate governments’ fiscal difficulties.
There are several important limitations in the related literature. The
research samples in empirical studies are often derived from a single
or a few states (Gorina, Maher, and Joffe 2018; McDonald and
Bruce 2015). The use of one-year cross-sectional data is common.
Moreover, few studies draw on the rich literature on government
management and performance to explain the effect of municipal
structure on government fiscal conditions (Carr 2015). This
research contributes to the related literature in these respects.
Measuring Municipal Structure on a Political-
Administrative Scale
Previous studies generally have adopted a binary variable to
indicate the primary municipal structure as mayor-council or
council-manager. However, prevalent cross-adoption of structural
characteristics among municipalities renders a dichotomous
classification problematic. Scholars have attempted to use different
methods to reclassify municipal form or structure (Carr and
Karuppusamy 2008, 2009, 2010; DeSantis and Renner 2002;
Frederickson, Johnson, and Wood 2004a, 2004b; Nelson and
Svara 2010). The typology of Nelson and Svara (2010) is based on
the stated mayor-council or council-manager form and examines
three critical structural features: how the mayor is selected, whether
a CAO exists, and who appoints the CAO. They divide the
council-manager form into three types: council (mayor)-manager,
mayor-council-manager, and empowered mayor-council-manager.
Likewise, the mayor-council form has four variations: mayor
and council-administrator, mayor-council-administrator, mayor-
administrator-council, and mayor-council. The seven subcategories
of municipal form “represent a progression from low centralized
political leadership and high professional autonomy to high political
leadership and low professional autonomy” (Nelson and Svara 2010,
558).
The approaches of Frederickson, Johnson, and Wood (2004a,
2004b) and Carr and Karuppusamy (2008, 2009, 2010) examine
a broader range of structural characteristics. The “Type III
City” framework of Frederickson, Johnson, and Wood (2004a,
2004b) classifies municipal structure into three groups: political
structure, administrative structure, and adapted structure. Carr
and Karuppusamy (2008, 2009, 2010) modify the Type III City
framework and classify the mayor-council form as political, adapted
political, or conciliated political, and they classify the council-
manager form as conciliated administrative, adapted administrative,
or administrative. Similarly, these reclassified municipal forms range
in order from the most political to the most administrative.
Municipal Structure Political-Administrative Index
Our approach to classifying municipal structure absorbs the
merits of the two methods introduced earlier. We aim to examine
the essential structural characteristics that are linked to the
degree of managerial professionalism, separation of powers and
checks and balances between officials, and local electoral systems
and to measure the general feature of municipal structure on a
political-administration scale. Table 1 presents the seven structural
characteristics to be examined. When choosing the appropriate
structural characteristics to construct the municipal structure index,
it is a challenge to strike a balance between comprehensiveness Index Reliability and Validity
and data availability. The more structural characteristics involved, The seven elements that make up the index have been thoroughly
the lower the possibility of collecting all the necessary data. discussed in previous studies (Carr and Karuppusamy 2008, 2009,
The basic principle guiding our choice is to put more priorities 2010; Frederickson, Johnson, and Wood 2004a, 2004b). The data
on institutional designs than managerial practices based on the on the structural characteristics of municipalities are collected from
presumption that the former can more fundamentally shape local the Municipal Form of Government surveys conducted by the ICMA
politics and policy formulation and implementation. in 2001, 2006, and 2011. The ICMA uses uniform survey questions
across time, thereby improving the consistency and reliability of
The seven structural characteristics measure three dimensions of the data (Clark 2015). Reliability is also defined by the correlation
the political or administrative features of municipal structure. of multiple measures of an underlying concept. A high Cronbach’s
The stated nominal municipal form and the presence or alpha indicates that the index elements refer to the same concept
absence of a CAO reflect a government’s degree of emphasis on and can be usefully combined in an index. Cronbach’s alpha in the
managerial professionalism. The method of electing the mayor, the sample is 0.76, indicating a close correlation between the index
independence of the mayor from the council, and the veto power elements as a group and acceptable reliability of the index (Berman
of the mayor on council decisions determine the separation of and Wang 2017; Clark 2015). In addition, table 2 reports the
powers and checks and balances between officials. The two features Pearson correlation coefficients for each pair of index elements
of council election capture the influence of parties and special and the statistical significance level. The results show that the
interests on electoral politics. Note that CAOs under different index elements are positively correlated with each other at the 0.01
statutory municipal forms (mayor-council or council-manager) significance level.
and appointment mechanisms may possess different authority and
influence. (Refer to appendix S1 in the Supporting Information Clark (2015) suggests assessing the validity of an index in three
for further discussion of the existence of a CAO as an index respects: content validity, criterion-related validity, and construct
component.) validity. Content validity requires that components of the index
should include all possible items that can measure the core concept
We code each structural characteristic as 0, 0.5, or 1. A lower (Kerlinger and Lee 2000), and there should be evidence from the
value indicates a more political characteristic, and a higher value literature demonstrating the efficacy of these items in measuring
means a more administrative characteristic. Factor analysis of the the concept. There may be many more structural characteristics
seven coded values yields a principal factor, which is adopted as to examine. However, it is vital to balance thoroughness with
the municipal structure political-administrative index. A lower parsimony and take into account data availability. As noted earlier,
index indicates a more political municipal structure, and a higher the seven structural characteristics are key to examine the political
index means a more administrative municipal structure. Among or administrative features of municipal structure. Criterion-related
all observations in our database, the municipal structure political- validity requires the constructed index to be correlated with
administrative index ranges from −1.92 to 1.72, with a mean of 0 other potential measures of the same concept. Results in the first
and standard deviation of 0.86. column of table 2 show that statutory municipal form, which is
generally used by previous studies to measure municipal structure,
Table 1  Standards of Coding Structural Characteristics of Municipalities
is positively and statistically significantly correlated with the other
index elements, thus supporting the criterion-related validity of the
Structural Characteristics/
Coded Values
0 0.5 1 index.
Stated nominal form Mayor-council Council-manager
(N = 2,544) (N = 4,242) Construct validity is usually assessed by theoretically relating
Existence of chief No Yes the index elements to the concept to be measured. The grounds
administrative officer
on which the seven structural characteristics are political or
(N = 1,062) (N = 5,724)
Mayor election method Direct election Nondirect election administrative have been discussed thoroughly in previous studies.
(N = 5,539) (N = 1,247) The more political municipal structure underscores separation of
Mayor is independent of Yes No
council
(N = 2,144) (N = 4,642) Table 2  Correlation Matrix of Index Components
Authority of mayor to veto Yes No
council decisions a b c d e f g
(N = 2,058) (N = 4,728) a 1
Partisan or nonpartisan Partisan Nonpartisan b 0.5445*** 1
election of council c 0.3352*** 0.1887*** 1
members d 0.4834*** 0.3311*** 0.3118*** 1
(N = 1,047) (N = 5,739) e 0.4731*** 0.3291*** 0.2907*** 0.5839*** 1
At-large or by-district By-district Combination At-large f 0.3164*** 0.2663*** 0.1258*** 0.2055*** 0.1992*** 1
election of council g 0.2225*** 0.1814*** 0.1835*** 0.2897*** 0.3070*** 0.1168*** 1
members
(N = 1,138) (N = 1,453) (N = 4,195) Notes: The sample size is 6,786.
a = stated nominal form; b = existence of chief administrative officer; c = mayor
Notes: The number of municipalities in the final sample of analysis with each election method; d = mayor is independent of council; e = authority of mayor to
structural characteristic is reported in the parentheses. A lower coded value veto council decisions; f = partisan or nonpartisan election of council members;
indicates a more political characteristic, and a higher coded value means a more g = at-large or by-district election of council members.
administrative characteristic. *** p < .01.

Municipal Structure Matters  3

powers and checks and balances between the mayor and council
members, direct responsiveness and accountability of the mayor
to voters, and influence of parties and special interests on local
elections. On the contrary, the more administrative municipal
structure highlights professional management of municipal affairs,
the concentration of powers in the council, and reduction of
partisan and special interests’ influence on local elections.
Finally, factor analysis of the seven coded values reports only one
principal factor with an eigenvalue higher than the threshold value
of one, which is 2.24, demonstrating that the index elements
are unidimensional and measure one single underlying concept.
Appendix S2 in the Supporting Information shows that the
principal factor has positive loadings for all index elements.
Theoretical Analyses and Hypotheses
We can propose at least three theoretical mechanisms through
which municipal structure matters for government performance.
Municipal structure shapes managerial professionalism, which
can substantially influence government performance. Municipal
structure shapes the managerial strategy stance by determining
the allocation of powers and the behavior motivations of officials.
Moreover, government officials under different municipal structures
ought to give different relative attention to managerial efficiency
and accountability, both of which are important for government
performance.
Managerial Professionalism and Government Performance
The statutory municipal form—mayor-council or council-
manager—provides a basic institutional setting in which
government officials manage local affairs. Scholars generally
hold that the council-manager form helps promote managerial
professionalism. A professional or administrative municipal
structure is also characterized by the existence of a professional
CAO, who holds a substantial degree of professional autonomy
in policy making and implementation (Demir and Reddick 2012;
Nalbandian 1999; Nelson and Svara 2015; Selden, Brewer, and
Brudney 1999; Zhang and Feiock 2010). Additionally, fierce
political competition between the mayor and council members
and partisan and special interests’ influence on local elections can
presumably prevent officials from investing adequate time and
resources to advance professional management.
This section elaborates on how distinctions in the career paths
and behavior motivations of elected officials and appointed
professional managers can result in different degrees of managerial
professionalism. Widely used to explain individual behavior and
human decision-making in social science studies, rational choice
theory assumes that self-interested individuals make decisions based
on their rational calculation of the induced costs and benefits by
a certain behavior (Dunleavy 2014). Accordingly, the primary
motivation of elected officials, whose careers are generally determined
by electoral results, is to win popularity by gratifying voters. Thus,
responding to the preferences of voters is presumably the work that
consumes the most time and energy of elected political officials.
The vote maximization motivation of elected officials is also
manifested in Frant’s (1996) theory of high-powered and low-
powered incentives of government officials. Frant (1996) contends
4  Public Administration Review  •  xxxx | xxxx 2020
that the market in the private sector is the basic institution for
determining the allocation of resources; as a counterpart, in the
public sector of representative democracy, politicians’ desire for
reelection functions as the basic institution for resource allocation.
The desire for reelection is the high-powered incentive that may
inspire elected officials to invest a substantial amount of time
and resources in public programs that might not produce the
maximum level of marginal benefits, thereby causing government
inefficiencies. In sharp contrast, professional managers play a dual
role in government management (Hassett and Watson 2002).
They must satisfy elected officials while, at the same time, they are
confronted with the pressure of building an outstanding reputation
among peers (Dunn and Legge 2001; Jimenez 2020). It is also fairly
possible that the formal education and occupational experiences of
professional managers have made them believe in the professional
values they pursue and the effectiveness of professional practices for
their communities.
Previous studies typically have found that political stability and the
degree of management latitude are the essential determinants of
career choice among long-serving professional managers (Hassett
and Watson 2002; Watson and Hassett 2003, 2004). Professional
managers are simultaneously confronted with political pressures
from elected officials and the professional values of government
management, and they are able to practice various degrees of
management latitude depending on their relationship with elected
officials. Specifically, in municipalities where elected officials rely
heavily on professional managers to operate the government and
elected officials are willing to confer a high degree of management
latitude on professional managers, managerial professionalism
is substantially advanced (Hassett and Watson 2002; Watson
and Hassett 2003). Moreover, local governments usually select
professional managers based on their experiences and reputations,
which can facilitate professional government management.
Managerial professionalism helps promote government fiscal
performance by improving the quality of budget making, capital
planning, and public funds management. Professional managers are
more capable of preparing government budgets based on revenue
and expenditure forecasts as well as the priority of local social and
economic development (Frank and Gianakis 1990; Reddick 2004).
Also, they are more likely to forecast revenues and expenditures
accurately with professional techniques, implement multiple-year
budgeting and capital plans, make use of capital markets rationally,
and diversify revenue sources strategically (Reddick 2004). These
capacities have great potential to improve government fiscal
conditions.
Managerial Strategy Stance and Government Performance
Miles and Snow (1978) classify organizational strategy stances as
prospecting, defending, or reacting. Scholars widely use this typology
to study public sector organizations. Prospectors are prone to
be policy pioneers and to search for innovative approaches to
deal with emerging environmental trends (Andrews, Boyne, and
Walker 2006). Defenders usually take a conservative viewpoint
toward innovations and “focus on a narrow range of services,
their core activities, to retain their existing portfolio of activities
and protect their share of the public budget from attacks by
predatory organizations” (Andrews, Boyne, and Walker 2006, 53).
Lastly, reactors generally do not maintain a consistent stance in
organizational operation and tend to adjust management strategies
according to environmental pressures (Boyne and Walker 2004,
2010). Many previous studies have found a positive (negative)
correlation between the prospecting (reacting) strategy stance and
organizational performance (Andrews et al. 2005; Andrews et al.
2008; Andrews, Boyne, and Walker 2006).
Under a more political municipal structure, separation of powers
forces the mayor and council to negotiate and compromise with
each other to achieve their desired policy goals. Council members
selected on a partisan or district basis have greater motivation to
cater to specific interest groups and political supporters. Therefore,
officials under a more political municipal structure will be more
likely to be reactors in terms of their strategy stance. In contrast,
under a more administrative municipal structure, collaboration
inside the managerial system and the influence of professional
managers facilitate innovations in government operation (Nelson
and Svara 2012). Moreover, in municipalities where at-large
elections are implemented and partisan influence is reduced,
professional autonomy is substantially promoted. Therefore,
officials are more likely to adopt innovative managerial practices
and techniques, and they can be described as prospectors in terms
of their strategy stance. The theoretical linkage between municipal
structure and managerial strategy stance, together with the empirical
evidence on the correlation between managerial strategy stance and
government performance, implies that municipalities adopting a
more administrative structure may perform better.
In summary, based on the two theoretical mechanisms described
here, a higher municipal structure political-administrative index
indicates a higher degree of managerial professionalism and a higher
probability of adopting prospecting as the managerial strategy stance
and, thus, better government fiscal performance. Accordingly, we
propose the first hypothesis as follows:
Hypothesis 1: There is a positive correlation between the
municipal structure political-administrative index and
government fiscal conditions.
Managerial Accountability and Efficiency
Based on previous studies, we define accountability as the
responsiveness and conformance of government officials to citizens’
preferences and demands for public policies and services (Dunn and
Legge 2001). We define efficiency as the capacity of government
to improve professional management and enhance public resource
utilization (Hayes and Chang 1990). Both managerial accountability
and efficiency can help improve government performance.
Assume that an optimal municipal structure that leads to the best
government performance can delegate government management
to professional experts for efficiency gains while holding officials
accountable to citizens (Putnam, Leonardi, and Nanetti 1994;
Shepsle 1988). Combining the merits of managerial efficiency and
accountability was the aim of the earliest advocates of council-
manager government (Svara 1989). Early Model City Charters
drafted by the National Municipal League (now the National Civic
League) endorsed a classic council-manager government structure
characterized by the appointment by the council of a professional
manager on the basis of qualifications (Frederickson, Wood,
and Logan 2001). At-large election of council members is also a
recommended structural characteristic. Because of the problem
of weak political leadership and accountability, later Model City
Charters after the fifth edition in 1941 incorporated the mayor-
council as a “second-best alternative” and provided guidelines for
the election of a mayor, despite still favoring the council-manager
structure (Frederickson, Wood, and Logan 2001). Municipalities
in recent decades have been increasingly absorbing the advantages
of both political and administrative municipal structures. Some
have suggested that the resulting hybrid municipal structure echoes
the rationale of combining managerial efficiency and political
accountability advocated by early reformers (Frederickson, Johnson,
and Wood 2004a, 2004b).
Under a more administrative or professional municipal structure,
delegating government management to expert managers improves
managerial efficiency (Deno and Mehay 1987; Hayes and Chang
1990). However, it increases the difficulty of holding managers
accountable because they do not directly face ballot pressure, and
they hold information and expertise advantages over voters and
elected officials. Moreover, the concentration of powers in the
council weakens checks and balances between the mayor and the
council. Therefore, voters’ monitoring costs increase. On the other
hand, under a more political municipal structure, political officials
are more likely to be driven by the high-powered incentive of
winning elections to respond to the wishes of voters urgently. That
helps ensure the responsiveness and accountability of officials (Frant
1996; Jimenez 2020). However, managerial efficiency may decline
because of inadequate professional management and increasing
political struggles.
It should be noted that public demands for governments are
multidimensional, and managerial efficiency may be an essential
element of voters’ expectations of government officials. Ensuring
accountability may help enhance efficiency under certain
circumstances. We do not intend to place managerial accountability
in complete opposition to managerial efficiency; instead, the core
argument is that government officials under different municipal
structures put different priorities on adhering to the demands of
voters and promoting professional management. In this sense, an
optimal municipal structure may be the one that combines the
merits of the political and administrative structural characteristics.
The famous Friedrich-Finer debate of the early 1940s provides
an alternative perspective that emphasizes the importance of both
political and administrative municipal structures. Friedrich and
Finer developed contrasting viewpoints in terms of the mechanisms
through which responsibility is defined and accountability
is exercised in a democratic government. Finer endorsed the
compliance of public administrators with elected bodies. In
contrast, Friedrich claimed that the responsibility and accountability
of public administrators are established by professional knowledge
and the perception of public sentiments. Discussions of the
merits of political and professional management of public affairs
are ongoing, just as the Friedrich-Finer debate is still alive and
relevant today. The complex relationships between elected officials,
appointed administrators, and the public when it comes to properly
defining responsibilities and adequately assuring accountability call
Municipal Structure Matters  5
for an optimal municipal structure that combines the merits of both
the political and administrative characteristics.
We can reach a similar conclusion that a municipal structure that
absorbs both the political and administrative merits may result in
the best government performance based on the theory of managerial
strategy stance of Miles and Snow (1978). Although previous
studies have shown that reacting is detrimental to organizational
performance (Andrews et al. 2008; Andrews, Boyne, and Walker
2006), it may not be the worst strategy stance for governments
because of their public nature. Actively prospecting can improve
managerial practices, but failure to gratify the political principal
may lead to an accountability problem, loss of political support, and
poor policy performance. In contrast, reactor officials are hesitant to
innovate or prospect, but their adherence to the political principal
and prompt adjustment of actions in accordance with the external
environment help them achieve more political support and avoid
risk. Therefore, a strategy stance that mixes prospecting and reacting
may combine the advantages of both and lead to the best government
performance. According to the several lines of theoretical analysis in
this section, the second hypothesis is stated as follows:
Hypothesis 2: A municipal structure political-administrative
index in the middle, rather than at either extreme—indicating
that the municipal structure combines the merits of both
political and administrative features—results in the best
government fiscal conditions.
Lastly, local governments’ fiscal conditions are shaped by the
external environment interacted with internal government
management (Cabaleiro, Buch, and Vaamonde 2013; Hendrick
2004). In addition to its direct effect on government fiscal
performance, municipal structure can interact with factors in the
external environment to exert an indirect or moderating effect.
Accordingly, the third hypothesis is specified as follows:
Hypothesis 3: The municipal structure political-
administrative index moderates the effects of external
environmental factors on government fiscal conditions.
Methodology
Model Specification
Indicators of government fiscal conditions are first modeled as a
function of the municipal structure and the control variables to test
hypothesis 1.
FCi ,t    SEi ,t 1  Fi ,t 1  I i ,t 1  Si ,t 1  i ,t (1)
In this equation, FCi, t refers to the indicators of government fiscal
conditions in municipality i at year t. SE, F, I, and S denote the
factors of socioeconomic characteristics, fiscal environments,
institutional constraints, and the municipal structure political-
administrative index, respectively. The intercept is α; the estimated
parameters are β, γ, δ, and λ; and the error term is ε. The
explanatory variables are lagged by one year.6
Model 1 is expanded to include the square of the municipal
structure index to test the possible inverted U-shaped relationship
specified in hypothesis 2.
6  Public Administration Review  •  xxxx | xxxx 2020
2
  FCi ,t    SEi ,t 1  Fi ,t 1  I i ,t 1  1Si ,t 1   2 Si ,t 1  i ,t (2)
Lastly, the multiplicative interaction regression model is used to
test hypothesis 3. All the socioeconomic, fiscal, and institutional
factors can interact with the municipal structure index to
predict government fiscal conditions. All constituent terms of
the interaction items should be included in the multiplicative
interaction model (Brambor, Clark, and Golder 2005).
  FCi ,t    SEi ,t 1  1SEi ,t 1  Si ,t 1  Fi ,t 1  1 Fi ,t 1  Si ,t 1
 I i ,t 1  1I i ,t 1  Si ,t 1  Si ,t 1  i,t (3)
Data Collection
The variables, descriptive statistics, and data sources are presented
in table 3. Data on municipalities’ structural characteristics are
derived from the Municipal Form of Government surveys conducted
by the ICMA in 2001, 2006, and 2011. The ICMA sends survey
questionnaires to all municipalities with populations of 2,500 and
greater and to selected ones below this threshold. The survey sample
sizes are 7,867, 8,278, and 8,813, respectively, and the response
rates are around 40 percent to 50 percent. Data on socioeconomic
characteristics for 2000 are collected from the decennial census of
that year. The corresponding data for 2010 are collected from the
decennial census of that year and supplemented by the American
Community Survey five-year estimates for 2010. Data for 2005 are
linearly interpolated using the data points in 2000 and 2010, which is
a commonly used data-processing method (Coate and Knight 2011).
Data on government finance are collected from the U.S. Census
Bureau’s government finance surveys conducted in 2002, 2007, and
2012; the indicators of government fiscal conditions are calculated
based on these data. In years ending with 2 and 7, the surveys cover
almost all municipalities. The data on government finance for 2002
are matched with the municipal structure data for 2001 and the
census socioeconomics data for 2000. The corresponding data for
2007, 2006, and 2005 are matched, as are the corresponding data for
2012, 2011, and 2010. Some of the fiscal and institutional factors
are coded based on state-imposed regulations on municipalities,
according to Krane, Rigos, and Hill (2001). In addition, data on the
tax and expenditure limitations (TELs) stringency index are collected
and updated based on Amiel, Deller, and Stallmann (2009).
Like most previous studies, only mayor-council and council-manager
municipalities with populations of more than 2,500 are included
in the empirical analysis to mitigate selection bias among small
municipalities. The final data set is an unbalanced panel because the
municipalities responding to the ICMA surveys are different each
year. All income and fiscal variables are adjusted to real 2013 dollars.
Indicators of Government Fiscal Conditions
Considering data availability, we select a group of widely used
and intuitively reasonable indicators to capture several essential
elements of government fiscal conditions, including cash solvency,
dependence on intergovernmental transfers, the level of outstanding
debt, and the operating budget balance.
First, per capita total cash and securities at the end of a fiscal year
measures governments’ cash solvency, which indicates the ability
Table 3  Variables, Descriptive Statistics, and Data Sources
Variables N Mean SD Min Max Data Sources
Fiscal condition indicators
Cash solvency (real 2013 dollars) 6,767 1,454.23 3,061.02 0 102,628.50 CGFD
Dependence on intergovernmental transfers (percent) 6,750 16.74 13.19 0 94.80 CGFD
Debt level (ratio) 6,579 1.24 1.52 0 66.16 CGFD
Operating budget balance (percent) 6,785 1.15 21.76 −307.86 91.10 CGFD
Socioeconomic factors
Population size 6,786 29,104 80,186 2,501 3,694,820 DC and ACS
Average household income (real 2013 dollars) 6,786 71,257.68 36,857.98 26,223.57 431,349.70 DC and ACS
Unemployment rate (percent) 6,786 6.72 3.47 0.30 41.70 DC and ACS
Fraction of residents over 65 years old (percent) 6,786 14.46 5.93 1.44 68.70 DC and ACS
Local industry structure 6,786 0.94 0.03 0.34 0.98 DC and ACS
Fiscal factors
Function in public health (dummy) 6,786 0.45 0.50 0 1 KRH (2001)
Function in social services (dummy) 6,786 0.37 0.48 0 1 KRH (2001)
Function in public schools (dummy) 6,786 0.23 0.42 0 1 KRH (2001)
Structure of general revenues 6,786 0.81 0.14 0 1 CGFD
Structure of tax revenues 6,786 0.51 0.26 0 1 CGFD
Institutional factors
BBR (dummy) 6,786 0.62 0.49 0 1 KRH (2001)
Debt limit (dummy) 6,786 0.78 0.41 0 1 KRH (2001)
TELs stringency index 6,786 20.21 10.58 0 38 ADS (2009)

Notes: Cash solvency is measured by the per capita total cash and securities at the end of a fiscal year. Dependence on intergovernmental transfers is measured as the
percentage of intergovernmental revenues to total revenues. Debt level is measured as the ratio of total debt outstanding at the end of a fiscal year to general revenues.
Operating budget balance is measured by the percentage of general budget surpluses or deficits to general revenues. CGFD = Census government finance database;
DC = decennial census; ACS = American Community Survey; KRH = Krane, Rigos, and Hill (2001); ADS = Amiel, Deller, and Stallmann (2009).

to realize liquidity to fulfill short-term fiscal obligations. A higher
value of the indicator signals healthier fiscal conditions. Second,
undue reliance on revenues from other governments impairs the
recipient governments’ fiscal autonomy and generally indicates
weak fiscal ability. Therefore, a lower share of intergovernmental
revenues in total revenues reflects healthier fiscal conditions. Third,
governments usually resort to the capital market to maintain regular
fiscal operations or fund capital projects. The ratio of total debt
outstanding at the end of a fiscal year to general revenues is an
applicable indicator of the level of outstanding debt. A lower value of
the indicator implies healthier fiscal conditions. Lastly, the operating
budget balance reflects the ability of governments to collect adequate
revenues to provide general public goods and services in a certain
fiscal year. The operating budget balance is measured by the ratio of
general budget surpluses or deficits to general revenues, with a higher
ratio indicating healthier fiscal conditions.
ranges between 0 and 1, with a higher value indicating a more
diversified industry structure.
The fiscal control variables include municipal revenue structure
and a specified group of municipal fiscal functions. We use dummy
variables to indicate the fiscal functions in public health, social
services, and public schools to consider the effect of variations in
service responsibilities on government fiscal conditions (Krane,
Rigos, and Hill 2001).8,9 The model includes the diversity index
(HHI) of municipalities’ general revenues and tax revenues to
control for the effect of revenue diversification on government fiscal
conditions.10 Lastly, local governments generally do not possess
full fiscal autonomy (McDonald and Bruce 2015; Rose 2010).
The institutional constraint factors controlled for in the estimation
model include the TELs stringency index and dummy variables for
the balanced budget requirement (BBR) and debt limit.

Control Variables Empirical Results

Selection of the control variables is based on evidence from previous
studies and data availability. The socioeconomic factors include
population size, average household income, unemployment rate,
fraction of residents aged 65 and above, and local industry structure.
These factors affect government fiscal conditions by shaping revenue
collections and spending needs (Hendrick 2004). In particular, the
industry structure is measured by the Hirschman-Herfindahl Index
(HHI), which is specified as follows (Suyderhoud 1994):
 ,
2
1
n
X j ,i ,t
HHI i ,t 
j 1
1 total cash and securities increases by around 5 percent.
1
n
where i indicates municipality, t refers to time, and Xj specifies the
fraction of the labor force in industry j out of n industries.7 HHI
Direct Effect of Municipal Structure
The results obtained by regressing model 1 with year fixed effects
and robust standard errors are presented in the first four columns in
panel A of table 4.11 The dependent variables are the four indicators
of fiscal conditions. Column 1 shows that the municipal structure
political-administrative index has a positive and statistically
significant correlation with cash solvency. The dependent variable
takes a natural logarithm form, and the municipal structure
index in the sample has a standard deviation of 0.86. Therefore,
holding other factors constant, as the municipal structure political-
administrative index increases by one standard deviation, per capita
Column 2 of table 4 shows that the municipal structure index
is negatively correlated with governments’ dependence on
intergovernmental revenues, and the effect is statistically significant
Municipal Structure Matters  7
at the 0.01 level. Smaller values of the dependent variable in column
2 mean healthier fiscal conditions. Therefore, the results suggest that
a more administrative municipal structure is associated with better
fiscal performance based on dependence on intergovernmental
revenues. All else being equal, as the municipal structure political-
administrative index increases by one standard deviation, the share
of intergovernmental revenues in total revenues decreases by about
1 percentage point. Results in columns 3 and 4 of table 4 show that
the correlations between the municipal structure index and the level
of outstanding debt and general budget balance are not statistically
significant.
are positive and negative, respectively, and both are statistically
significant. Therefore, there is a quasi-inverted U-shaped
relationship between the municipal structure index and cash
solvency. Figure 1 illustrates that as the municipal structure adjusts
from the most political to the most administrative, governments’
cash solvency improves and then declines. According to the
estimated coefficients, the curve reaches its pinnacle at the index of
around 0.56. The results support hypothesis 2. Note that although a
quasi-inverted U-shaped relationship is demonstrated, municipalities
adopting a structure with an imbalance toward administrative
characteristics are more likely to achieve the greatest cash solvency.

Inverted U-Shaped Relationship Column 6 of table 4 shows that the effects of the municipal structure
Model 2 is estimated to test hypothesis 2, producing the results index and its square are negative and positive, respectively, and
reported in columns 5–8 in panel B of table 4. Column 5 shows both are statistically significant. Governments’ dependence on
that the effects of the municipal structure index and its square intergovernmental revenues declines until the municipal structure

Table 4  Direct Effect of Municipal Structure and the Inverted U-Shaped Relationship
Panel A Direct Effect of Municipal Structure (Model 1) Panel B Inverted U-shaped Relationship Test (Model 2)
1 2 3 4 5 6 7 8
Intergov. revenue General budget Intergov. revenue General budget
VARIABLES Cash solvency Debt level Cash solvency Debt level
dependence Surplus/deficit dependence Surplus/deficit
Municipal structure index 0.0602*** −1.275*** 0.0312 0.304 0.0362* −0.869*** −0.0844** 0.476
(0.0137) (0.210) (0.0208) (0.324) (0.0197) (0.292) (0.0395) (0.491)
Municipal structure index −0.0322* 0.546* −0.155*** 0.232
squared (0.0190) (0.288) (0.0357) (0.474)
Socioeconomic factors
Population size (log) 0.230*** −0.0813 0.0591*** −0.319 0.233*** −0.128 0.0725*** −0.340
(0.0105) (0.158) (0.0197) (0.247) (0.0105) (0.157) (0.0188) (0.249)
Average household income (log) 0.285*** 0.0850 −0.203*** 1.981* 0.285*** 0.0834 −0.205*** 1.979*
(0.0402) (0.627) (0.0622) (1.073) (0.0402) (0.626) (0.0621) (1.073)
Unemployment rate −0.126*** 2.323*** −0.145*** 0.738 −0.124*** 2.297*** −0.138*** 0.727
(0.0287) (0.403) (0.0475) (0.699) (0.0287) (0.403) (0.0478) (0.699)
Fraction of residents over 65 0.262*** 0.192 −0.254*** −1.762** 0.267*** 0.108 −0.231*** −1.797**
(0.0287) (0.373) (0.0409) (0.762) (0.0289) (0.376) (0.0420) (0.768)
Local industry structure −1.399*** −1.419 0.0879 −0.669 −1.417*** −1.097 0.00378 −0.534
(0.400) (5.275) (0.505) (8.064) (0.400) (5.273) (0.509) (8.085)
Fiscal factors
Function in public health 0.0300 1.843*** −0.0268 3.387*** 0.0329 1.795*** −0.0117 3.367***
(0.0299) (0.430) (0.0402) (0.816) (0.0300) (0.433) (0.0398) (0.821)
Function in social services −0.185*** −5.496*** 0.253*** −0.669 −0.188*** −5.455*** 0.242*** −0.652
(0.0290) (0.418) (0.0386) (0.788) (0.0291) (0.420) (0.0391) (0.793)
Function in public schools 0.0955*** 2.037*** 0.0147 −2.298*** 0.0922*** 2.091*** −0.000722 −2.274***
(0.0276) (0.453) (0.0427) (0.748) (0.0276) (0.455) (0.0430) (0.744)
Structure of general revenues 0.844*** 13.52*** 0.834*** −9.294*** 0.841*** 13.57*** 0.820*** −9.277***
(0.108) (1.593) (0.179) (1.995) (0.108) (1.592) (0.178) (1.995)
Structure of tax revenues −0.110** −7.564*** 0.0142 1.762 −0.113** −7.510*** −0.00157 1.784*
(0.0465) (0.678) (0.0790) (1.076) (0.0466) (0.676) (0.0778) (1.074)
Institutional factors
BBR −0.118*** −4.106*** −0.121*** −1.934*** −0.122*** −4.040*** −0.139*** −1.906***
(0.0256) (0.357) (0.0373) (0.594) (0.0257) (0.360) (0.0385) (0.598)
Debt limit −0.146*** −2.029*** −0.0913** −0.226 −0.146*** −2.029*** −0.0917** −0.226
(0.0277) (0.419) (0.0417) (0.700) (0.0277) (0.418) (0.0415) (0.700)
TELs stringency index 0.00516*** −0.134*** −0.00255 0.0388 0.00546*** −0.139*** −0.00115 0.0367
(0.00111) (0.0169) (0.00172) (0.0272) (0.00113) (0.0173) (0.00181) (0.0276)
2006 (2001 as reference) 0.446*** −2.374*** 0.146*** 1.140* 0.444*** −2.356*** 0.141*** 1.148*
(0.0266) (0.369) (0.0471) (0.670) (0.0266) (0.369) (0.0476) (0.669)
2011 (2001 as reference) 0.676*** −2.437*** 0.302*** 1.161* 0.673*** −2.393*** 0.289*** 1.181*
(0.0280) (0.413) (0.0546) (0.659) (0.0280) (0.414) (0.0557) (0.660)
Constant 1.469** 15.15* 3.066*** −8.180 1.465** 15.19* 3.077*** −8.143
(0.634) (9.078) (0.924) (14.85) (0.633) (9.079) (0.924) (14.85)
Observations 6,758 6,750 6,579 6,785 6,758 6,750 6,579 6,785
R2 0.205 0.120 0.026 0.015 0.205 0.121 0.029 0.015

Notes: Results in panels A and B are obtained by estimating models 1 and 2, respectively. Refer to table 3 for the measurements and descriptive statistics of variables.
Robust standard errors are reported in parentheses.
***p<0.01; **p<0.05; *p<0.1.

8  Public Administration Review  •  xxxx | xxxx 2020


Figure 1  Predicted Cash Solvency and 95% CIs along with
Municipal Structure Index
Figure 2  Predicted Intergovernmental Revenue Dependence
and 95% CIs along with Municipal Structure Index
index reaches around 0.80 and then increases, which is displayed in
figure 2. A lower value of the dependent variable indicates better fiscal
conditions. Therefore, hypothesis 2 is again supported. Similarly,
although a quasi-inverted U-shaped relationship between the
municipal structure index and government fiscal conditions exists,
municipalities adopting a relatively more administrative municipal
structure are more likely to achieve the best fiscal conditions.
Results in column 7 of table 4 show that the effects of both the
municipal structure index and its square on the level of outstanding
debt are negative and statistically significant. The results imply that
as the municipal structure becomes more administrative, the level of
outstanding debt declines at an increasing rate. Lastly, the results in
column 8 show that neither the municipal structure index nor its square
is statistically significantly related to governments’ general budget balance.
Moderating Effect of Municipal Structure
The results from estimating model 3 are reported in table 5. In the
multiplicative interaction model, the effects of the socioeconomic,
fiscal, and institutional factors on government fiscal conditions
are conditional on the municipal structure index. Most of the
interaction items show statistically significant effects, especially in
the first two columns. To save space, only the unemployment rate
and TELs are discussed as examples. The former is an important
indicator of local economic conditions, and the latter is an essential
rule to constrain governments’ fiscal policy making.
Results in table 5 show that the unemployment rate is negatively
and statistically significantly correlated with government fiscal
conditions for the first two indicators. The effects of the interaction
between the unemployment rate and municipal structure index
show that as the municipal structure becomes more administrative,
the negative effect of the unemployment rate on government
fiscal conditions is moderated. Possibly, increasing managerial
professionalism and efficiency help government officials mitigate
the negative effect of economic difficulties on government fiscal
conditions. The results also show that the TELs stringency index is
positively correlated with government fiscal conditions for the first
two indicators. Based on the effects of the interaction between the
municipal structure index and the TELs stringency index, as the
municipal structure becomes more administrative, the positive effect
of the TELs is moderated. One explanation is that a higher level of
professional autonomy helps government officials circumvent the
fiscal constraints imposed by the TELs.
Solving the Endogeneity Problem Using Instrumental
Variables
The potential endogeneity problem may arise from omitted variable
bias or reverse causality. For instance, unobservable or unmeasurable
factors in the error term of the estimation models, such as the
political preferences of local citizens and the cultural and historical
features of municipalities, can simultaneously affect municipal
structure and government fiscal conditions. It is also possible
that municipalities may appoint a CAO or adjust other structural
characteristics because of poor fiscal performance.
This study uses instrumental variables (IVs) for municipal structure
to address the potential endogeneity problem. Valid IVs should be
able to influence government fiscal conditions only by affecting
municipal structure, and any other possible paths should be
excluded. The selection of IVs is based on the existing evidence
on the determinants of municipal structure. Nelson (2011) finds
that states’ statutory or constitutional provisions on municipal
governments’ autonomy to make institutional changes influence
local choices of municipal structure. Similarly, Marando and Reeves
(1993) find that states’ constitutional or legislative decisions to allow
for structural changes substantially affect the structural reforms of
county governments.
We use the existence of the initiative and popular referendum
allowing citizens to place changes to the charter, ordinance, and
home rule on the ballot as IVs for municipal structure. Data
on the initiative and popular referendum are derived from the
ICMA’s Municipal Form of Government surveys. According to the
survey results, many structural changes are made through local
charters, resolutions, or ordinances. Moreover, there is no sufficient
or convincing evidence demonstrating that citizens’ autonomy
to change the charter and ordinance through the initiative and
Municipal Structure Matters  9
Table 5  Moderating Effect of Municipal Structure
1 2 3 4
Cash Solvency Intergov. Revenue Dependence Debt Level General Budget Surplus/Deficit
Municipal structure index −0.184 33.06*** −0.364 −10.60
(0.720) (10.99) (0.912) (16.79)
Socioeconomic factors
Population size (log) 0.225*** −0.235 0.0738*** −0.358
(0.0107) (0.152) (0.0188) (0.254)
Population size * Index −0.0725*** −0.779*** 0.0276 0.101
(0.0123) (0.199) (0.0173) (0.281)
Average household income (log) 0.297*** 0.336 −0.161*** 2.078*
(0.0408) (0.628) (0.0615) (1.099)
Average household income * Index 0.00250 −1.647** −0.0163 0.821
(0.0488) (0.816) (0.0596) (1.167)
Unemployment rate −0.114*** 2.068*** −0.109** 0.761
(0.0291) (0.408) (0.0496) (0.712)
Unemployment rate * Index 0.0687** −1.278*** 0.0373 0.726
(0.0320) (0.476) (0.0442) (0.767)
Fraction of residents over 65 0.272*** 0.381 −0.242*** −1.772**
(0.0292) (0.374) (0.0428) (0.798)
Fraction of residents over 65 * Index −0.148*** −0.443 −0.0430 0.166
(0.0386) (0.512) (0.0598) (1.085)
Local industry structure −1.373*** −3.531 0.0952 0.370
(0.388) (5.379) (0.503) (8.059)
Local industry structure * Index 1.216*** −12.09** 0.567 −3.191
(0.401) (5.685) (0.553) (9.527)
Fiscal factors
Function in public health 0.0171 3.012*** −0.0270 3.267***
(0.0313) (0.456) (0.0415) (0.860)
Function in public health * Index 0.0750** 2.070*** 0.0216 −0.829
(0.0333) (0.473) (0.0395) (0.879)
Function in social services −0.174*** −4.465*** 0.249*** −0.716
(0.0303) (0.447) (0.0412) (0.830)
Function in social services * Index −0.142*** 2.263*** 0.0202 0.840
(0.0338) (0.486) (0.0447) (0.897)
Function in public schools 0.0933*** −0.224 0.0210 −2.062**
(0.0311) (0.494) (0.0485) (0.823)
Function in public schools * Index −0.0526 −5.465*** 0.133*** 0.509
(0.0357) (0.597) (0.0507) (0.877)
Structure of general revenues 0.873*** 14.23*** 0.824*** −8.931***
(0.106) (1.532) (0.182) (2.050)
Structure of general revenues * Index 0.243* 3.836** 0.0320 0.784
(0.131) (1.920) (0.171) (2.192)
Structure of tax revenues −0.0810 −7.143*** −0.0417 1.734
(0.0494) (0.733) (0.0775) (1.131)
Structure of tax revenues * Index 0.0299 −0.218 −0.243*** 0.430
(0.0546) (0.879) (0.0743) (1.291)
Institutional factors
BBR −0.117*** −5.104*** −0.132*** −1.841***
(0.0285) (0.405) (0.0438) (0.632)
BBR * Index 0.0351 −0.916** 0.0292 −0.202
(0.0305) (0.446) (0.0428) (0.716)
Debt limit −0.132*** −2.589*** −0.0906** −0.303
(0.0283) (0.445) (0.0441) (0.722)
Debt limit * Index −0.114*** 2.453*** −0.0869* 1.563*
(0.0336) (0.557) (0.0477) (0.812)
TELs stringency index 0.00603*** −0.164*** −0.00218 0.0332
(0.00117) (0.0178) (0.00186) (0.0285)
TELs stringency index * Index −0.00400** 0.0561** −0.00363 0.00555
(0.00161) (0.0260) (0.00227) (0.0373)
2006 (2001 as reference) 0.444*** −2.400*** 0.143*** 1.174*
(0.0266) (0.363) (0.0472) (0.669)
2011 (2001 as reference) 0.677*** −2.305*** 0.287*** 1.197*
(0.0281) (0.406) (0.0559) (0.657)
Constant 1.246** 16.49* 2.407*** −10.08
(0.633) (9.084) (0.893) (15.23)
Observations 6,758 6,750 6,579 6,785
R2 0.214 0.154 0.030 0.016

Notes: Results are obtained by estimating model 3. Refer to table 3 for the measurements and descriptive statistics of variables. Robust standard errors are reported in
parentheses.
***p<0.01; **p<0.05; *p<0.1.

10  Public Administration Review  •  xxxx | xxxx 2020

Table 6  Effect of Municipal Structure Using Instrumental Variables
1 2 3 4
Variables
Cash Solvency Intergov. Revenue Dependence Debt Level General Budget Surplus/Deficit
Municipal structure index 0.491*** −6.931*** −0.494** 1.706
(0.139) (2.043) (0.247) (3.017)
Socioeconomic factors
Population size (log) 0.238*** −0.143 0.0419* −0.568**
(0.0130) (0.181) (0.0254) (0.276)
Average household income 0.106* 1.916* −0.0388 1.940
(0.0635) (1.020) (0.0966) (1.403)
Unemployment rate −0.152*** 2.505*** −0.115** 0.649
(0.0329) (0.491) (0.0574) (0.756)
Fraction of residents over 65 0.256*** −0.0535 −0.291*** −1.880**
(0.0327) (0.428) (0.0485) (0.846)
Local industry structure −1.335*** 1.175 0.307 2.494
(0.469) (5.969) (0.581) (8.553)
Fiscal factors
Function in public health 0.197*** −0.261 −0.215** 3.783***
(0.0628) (0.886) (0.104) (1.400)
Function in social services −0.211*** −4.969*** 0.227*** −0.334
(0.0341) (0.488) (0.0456) (0.838)
Function in public schools 0.0510 2.598*** 0.118* −2.363***
(0.0352) (0.577) (0.0608) (0.834)
Structure of general revenues 0.739*** 14.28*** 0.905*** −8.294***
(0.122) (1.762) (0.201) (2.096)
Structure of tax revenues −0.306*** −4.993*** 0.255* 0.861
(0.0784) (1.074) (0.147) (1.594)
Institutional factors
BBR −0.354*** −1.082 0.152 −2.966*
(0.0792) (1.153) (0.127) (1.647)
Debt limit −0.220*** −1.089** −0.0655 −0.456
(0.0336) (0.485) (0.0517) (0.797)
TELs stringency index −0.00241 −0.0358 0.00655 −0.0117
(0.00271) (0.0397) (0.00490) (0.0585)
2006 (2001 as reference) −0.640*** 2.180*** −0.308*** −0.815
(0.0320) (0.471) (0.0622) (0.714)
2011 (2001 as reference) −0.222*** −0.0997 −0.160*** 0.441
(0.0302) (0.429) (0.0475) (0.707)
Constant 4.568*** −15.45 1.030 −5.554
(0.959) (14.33) (1.361) (19.38)
Weak identification test
(Cragg-Donald Wald F-statistic) 39.1 40.1 39.3 40.3
Overidentification test 2.67 1.69 0.210 0.265
(Hansen J-statistic) (p = .102) (p = .193) (p = .647) (p = .607)
Observations 5,853 5,844 5,690 5,875

Notes: Results are obtained by estimating model 1, using the existence of the initiative and popular referendum allowing citizens to place changes of the charter, ordi-
nance, and home rule on the ballot as IVs for the municipal structure political-administrative index. Refer to table 3 for the measurements and descriptive statistics of
variables. Robust standard errors are reported in parentheses.
***p<0.01; **p<0.05; *p<0.1.

popular referendum has independent effects on government fiscal
conditions.
The results of the IV estimation of model 1 using the two-stage least
squares approach are reported in table 6. The two IVs perform well
from the statistical perspective. The Cragg-Donald Wald F-statistics
from the weak identification test shows that the “weak instruments”
hypothesis is statistically significantly rejected in all four regressions.
The Hansen J-statistics from the overidentification test suggests that
the hypothesis of no correlation between the IVs and the error term
in the model of interest cannot be statistically significantly rejected.
The results in table 6 show that the municipal structure index
positively and statistically significantly affects government fiscal
conditions for the first three indicators. Compared with the
results shown in columns 1 and 2 of table 4, the signs of the
effects of the municipal structure index remain unchanged, but
the magnitudes of effects are much larger in table 6. The effect of
the municipal structure index on the level of outstanding debt in
column 3 of table 4 is not statistically significant, but it becomes
statistically significant in column 3 of table 6. The results imply
that the positive effect of the administrative municipal structure on
government fiscal conditions may be underestimated in the original
estimation because the endogeneity problem.
Conclusion and Discussion
This research contributes to the literature by constructing an index
to reclassify municipal structure on a political-administrative scale,
theoretically addressing the mechanisms through which municipal
structure affects government performance, and obtaining empirical
evidence on the effect of municipal structure on government fiscal
performance. Moreover, we take into account and use IVs to solve
Municipal Structure Matters  11
the potential endogeneity problem in the empirical analysis. The
findings show that municipalities adopting a structure with an
imbalance toward more administrative or professional characteristics
are more likely to achieve the best fiscal conditions in terms of
cash solvency, dependence on intergovernmental revenues, and
outstanding debt. Moreover, municipal structure moderates
the influence of the external environment on government fiscal
conditions. Noteworthily, the evidence in favor of the administrative
municipal structure when it comes to improving government fiscal
conditions does not suggest that a more political municipal structure
might not be equally or even more efficacious in the pursuit of other
important conditions or values. That possibility should be tested.
It is important to acknowledge the potential limitations of this
study regarding the sample selection and external validity of the
empirical findings. Following the common practice of previous
studies, we focus on the municipalities responding to the ICMA
survey with populations over 2,500. External validity is therefore
restricted to that group of municipalities. The surveys have a
response rate of 40 percent to 50 percent, which is higher than most
other national surveys of local governments. However, whether
the respondents are representative of the selected survey samples
is still a concern. The results from the mean tests between the
respondents and nonrespondents in each survey show that the two
groups are not statistically significantly different with regard to
the basic socioeconomic and demographic features except for the
fraction of population aged 65 and above in 2001 and the average
household income in 2011.12 These two variables are controlled in
the estimation models. The results from the mean tests can relieve
the concern about the representativeness of the survey respondents
to a great extent.
Another limitation of sample selection stems from the self-
selection bias, which arises when selected samples intentionally sort
themselves into a certain group. Municipalities with professional
managers may be more likely to respond to the ICMA survey.
Unfortunately, we do not have data on the presence or absence
of professional managers in the nonrespondent municipalities.
However, according to the existing evidence that the forms
or structural characteristics of municipalities are substantially
determined by their socioeconomic and demographic features
(Choi, Feiock, and Bae 2013; Knoke 1982; Lineberry and
Fowler 1967; Simmons and Simmons 2004), the statistically
insignificant results in the mean tests introduced above imply that
the respondents and nonrespondents have similar probabilities of
having professional managers as well as other important structural
characteristics. Therefore, the concern about a higher response rate
among municipalities with professional managers is eased.
Notes
1. Many previous studies have used the terms “municipal structure” and “municipal
form” interchangeably. This research uses “municipal structure” as a broader
concept that includes the connotations of municipal form and other important
structural characteristics.
2. The other three statutory municipal forms are the commission, the town
meeting, and the representative town meeting. These forms have been adopted
by a small group of municipalities, mainly located in New England. Like most
previous studies, this research focuses on only mayor-council and council-
manager municipalities.
12  Public Administration Review  •  xxxx | xxxx 2020
3. The characteristics of mayor-council and council-manager municipalities are
summarized at https://www.nlc.org/forms-of-municipal-government.
4. Mayors’ powers may vary across municipalities, depending on the municipal charter,
resulting in strong-mayor and weak-mayor versions of mayor-council government.
5. The method of electing the mayor may change over time and varies across
municipalities depending on the municipal charter. A mayor selected by the
council was common in the early days of the council-manager government, but
it is much less common today.
6. Some of the socioeconomic and fiscal control variables are lagged by two years
(see the section on data collection).
7. Based on the classification of the Census Bureau, there are 13 categories of
industries in local economies: agriculture, forestry, fishing and hunting, and
mining; construction; manufacturing; wholesale trade; retail trade;
transportation and warehousing, and utilities; information; finance, insurance,
real estate, and rental and leasing; professional, scientific, management,
administrative, and waste management services; educational, health and social
services; arts, entertainment, recreation, accommodation and food services; other
services (except public administration); and public administration.
8. Krane, Rigos, and Hill (2001) classify municipal governments’ fiscal functions
using nine categories: general government; public safety; public health; public
works; social services; economic development; physical environment; culture and
recreation; and public schools. This study controls for three of them in the model
because almost all municipalities have functions in the other six categories.
9. Both public service responsibility and quality may affect government fiscal
conditions. Because of data availability, this research controls for the range of
service responsibilities but not service quality. Evaluating and comparing the
quality of public services of a large number of municipalities on a national scale
is an enormous challenge. One may argue that failure to control for service
quality in the empirical model might lead to omitted variable bias. However, we
use instrumental variables to reduce the potential endogeneity of municipal
structure and check the robustness of findings in the next section.
10. According to the Census Bureau, general revenues are classified as
intergovernmental revenues, total taxes, current charges, and miscellaneous
general revenues. Tax revenues are composed of property tax revenues, sales tax
revenues, license tax revenues, and other revenues.
11. The model controls for the TELs, BBR, and debt limit, which are essential
state-imposed fiscal constraints on local fiscal autonomy. To further alleviate the
concern that some other omitted state-level laws may affect the municipal
structure and government fiscal conditions simultaneously, thus leading to the
endogeneity problem, we also ran the regression with state fixed effects. We got
similar results to those reported in table 4 after adding state fixed effects.
Considering that state fixed effects automatically absorb the effects of the
explanatory variables coded based on the state-level data, including government
functions in public health, social services, and public schools and the three factors
of fiscal constraints, we report the regression results with year, but not state, fixed
effects. Note that we use instrumental variables to solve the endogeneity problem
of municipal structure in the next section, further mitigating the concern about
the endogeneity caused by the omitted state-level effects.
12. Based on data availability, we test the differences of means in population size,
average household income, unemployment rate, fraction of residents aged 65 and
above, and HHI of local industry structure. Because of space limitations, the results
of mean tests are not reported in the article but are available from the author.
References
Amiel, Lindsay, Steven Deller, and Judith Stallmann. 2009. The Construction of a
Tax and Expenditure Limitation Index for the US. Staff Paper 536. Department
of Agricultural and Applied Economics, University of Wisconsin–Madison.
https://aae.wisc.edu/pubs/sps/the-construction-of-a-tax-and/ [accessed March
19, 2020].
Andrews, Rhys, George A. Boyne, Kenneth J. Meier, Laurence J. O’Toole, Jr., and
Richard M. Walker. 2005. Representative Bureaucracy, Organizational Strategy,
and Public Service Performance: An Empirical Analysis of English Local
Government. Journal of Public Administration Research and Theory 15(4): 489–504.
Andrews, Rhys, George A. Boyne, and Richard M. Walker. 2006. Strategy Content
and Organizational Performance: An Empirical Analysis. Public Administration
Review 66(1): 52–63.
Andrews, Rhys, George A. Boyne, Jennifer Law, and Richard M. Walker. 2008.
Organizational Strategy, External Regulation and Public Service Performance.
Public Administration 86(1): 185–203.
Berman, Evan, and Xiaohu Wang. 2017. Essential Statistics for Public Managers and
Policy Analysts. Los Angeles: CQ Press.
Boyne, George A., and Richard M. Walker. 2004. Strategy Content and Public Service
Organizations. Journal of Public Administration Research and Theory 14(2): 231–52.
———. 2010. Strategic Management and Public Service Performance: The Way
Ahead. Public Administration Review 70(Special issue): S185–92.
Brambor, Thomas, William Roberts Clark, and Matt Golder. 2005. Understanding
Interaction Models: Improving Empirical Analyses. Political Analysis 14(1): 63–82.
Brown, Ken W. 1993. The 10-Point Test of Financial Condition: Toward an Easy-to-
Use Assessment Tool for Smaller Cities. Government Finance Review 9(6): 21–6.
Cabaleiro, Roberto, Enrique Buch, and Antonio Vaamonde. 2013. Developing a
Method to Assessing the Municipal Financial Health. American Review of Public
Administration 43(6): 729–51.
Carr, Jered B. 2015. What Have We Learned about the Performance of Council-
Manager Government? A Review and Synthesis of the Research. Public
Administration Review 75(5): 673–89.
Carr, Jered B., and Shanthi Karuppusamy. 2008. The Adapted Cities Framework: On
Enhancing Its Use in Empirical Research. Urban Affairs Review 43(6): 875–86.
———. 2009. Beyond Ideal Types of Municipal Structure: Adapted Cities in
Michigan. American Review of Public Administration 39(3): 304–21.
———. 2010. Reassessing the Link between City Structure and Fiscal Policy: Is the
Problem Poor Measures of Governmental Structure? American Review of Public
Administration 40(2): 209–28.
Choi, Cheon Geun, Richard C. Feiock, and Jungah Bae. 2013. The Adoption and
Abandonment of Council-Manager Government. Public Administration Review
73(5): 727–36.
Clark, Benjamin Y. 2015. Evaluating the Validity and Reliability of the Financial
Condition Index for Local Governments. Public Budgeting & Finance 35(2):
66–88.
Coate, Stephen, and Brian Knight. 2011. Government Form and Public Spending:
Theory and Evidence from US Municipalities. American Economic Journal:
Economic Policy 3(3): 82–112.
Demir, Tansu, and Christopher G. Reddick. 2012. Understanding Shared Roles in
Policy and Administration: An Empirical Study of Council-Manager Relations.
Public Administration Review 72(4): 526–35.
Deno, Kevin T., and Stephen L. Mehay. 1987. Municipal Management Structure and
Fiscal Performance: Do City Managers Make a Difference? Southern Economic
Journal 53(3): 627–42.
DeSantis, Victor S., and Tari Renner. 2002. City Government Structures: An
Attempt at Clarification. State and Local Government Review 34(2): 95–104.
Dunleavy, Patrick. 2014. Democracy, Bureaucracy and Public Choice: Economic
Approaches in Political Science. New York: Routledge.
Dunn, Delmer D., and Jerome S. Legge, Jr. 2001. US Local Government Managers
and the Complexity of Responsibility and Accountability in Democratic
Governance. Journal of Public Administration Research and Theory 11(1): 73–88.
Frank, Howard A., and Gerasimos A. Gianakis. 1990. Raising the Bridge Using Time
Series Forecasting Models. Public Productivity & Management Review 14(2): 171–88.
Frant, Howard. 1996. High-Powered and Low-Powered Incentives in the Public
Sector. Journal of Public Administration Research and Theory 6(3): 365–81.
Frederickson, H. George, Gary Alan Johnson, and Curtis Wood. 2004a. The
Changing Structure of American Cities: A Study of the Diffusion of Innovation.
Public Administration Review 64(3): 320–30.
———. 2004b. The Adapted City: Institutional Dynamics and Structural Change.
Armonk, NY: M.E. Sharpe.
Frederickson, H. George, Curtis Wood, and Brett Logan. 2001. How American
City Governments Have Changed: The Evolution of the Model City Charter.
National Civic Review 90(1): 3–18.
Gorina, Evgenia, Craig Maher, and Marc Joffe. 2018. Local Fiscal Distress:
Measurement and Prediction. Public Budgeting & Finance 38(1): 72–94.
Groves, Sanford M., W. Maureen Godsey, and Martha A. Shulman. 1981. Financial
Indicators for Local Government. Public Budgeting & Finance 1(2): 5–19.
Hassett, Wendy L., and Douglas J. Watson. 2002. Long-Serving City Managers:
Practical Application of the Academic Literature. Public Administration Review
62(5): 622–9.
Hayes, Kathy, and Semoon Chang. 1990. The Relative Efficiency of City Manager and
Mayor-Council Forms of Government. Southern Economic Journal 57(1): 167–77.
Hendrick, Rebecca. 2004. Assessing and Measuring the Fiscal Health of Local
Governments: Focus on Chicago Suburban Municipalities. Urban Affairs Review
40(1): 78–114.
Jimenez, Benedict S. 2017. When Ties Bind: Public Managers’ Networking Behavior
and Municipal Fiscal Health after the Great Recession. Journal of Public
Administration Research and Theory 27(3): 450–67.
———. 2020. Municipal Government Form and Budget Outcomes: Political
Responsiveness, Bureaucratic Insulation, and the Budgetary Solvency of Cities.
Journal of Public Administration Research and Theory 30(1): 161–77.
Judd, Dennis R., and Todd Swanstrom. 2015. City Politics. New York: Routledge.
Kerlinger, Fred N., and H.B. Lee. 2000. Foundations of Behavioral Research. San
Diego, CA: Harcourt College.
Kloha, Philip, Carol S. Weissert, and Robert Kleine. 2005. Developing and Testing a
Composite Model to Predict Local Fiscal Distress. Public Administration Review
65(3): 313–23.
Knoke, David. 1982. The Spread of Municipal Reform: Temporal, Spatial, and Social
Dynamics. American Journal of Sociology 87(6): 1314–39.
Krane, Dale, Platon N. Rigos, and Melvin Hill. 2001. Home Rule in America: A Fifty-
State Handbook. Washington, DC: CQ Press.
Lineberry, Robert L., and Edmund P. Fowler. 1967. Reformism and Public Policies in
American Cities. American Political Science Review 61(3): 701–16.
Maher, Craig S., and Karl Nollenberger. 2009. Revisiting Kenneth Brown’s 10-Point
Test. Government Finance Review 25(5): 61–6.
Marando, Vincent L., and Mavis Mann Reeves. 1993. County Government
Structural Reform: Influence of State, Region, and Urbanization. Publius: The
Journal of Federalism 23(1): 41–52.
McDonald, Bruce D., III. 2015. Does the Charter Form Improve the Fiscal Health
of Counties? Public Administration Review 75(4): 609–18.
Miles, Robert E., and Charles C. Snow. 1978. Organizational Strategy, Structure and
Process. New York: McGraw-Hill.
Nalbandian, John. 1999. Facilitating Community, Enabling Democracy: New Roles
for Local Government Managers. Public Administration Review 59(3): 187–97.
Nelson, Kimberly. 2011. State-Level Autonomy and Municipal Government
Structure: Influence on Form of Government Outcomes. American Review of
Public Administration 41(1): 542–61.
Nelson, Kimberly L., and James H. Svara. 2010. Adaptation of Models Versus
Variations in Form: Classifying Structures of City Government. Urban Affairs
Review 45(4): 544–62.
———. 2012. Form of Government Still Matters: Fostering Innovation in US
Municipal Governments. American Review of Public Administration 42(3): 257–81.
———. 2015. The Roles of Local Government Managers in Theory and Practice: A
Centennial Perspective. Public Administration Review 75(1): 49–61.
Municipal Structure Matters  13
Nollenberger, Karl. 2003. Evaluating Financial Condition: A Handbook for Local
Government. Washington, DC: International City/County Management
Association.
Putnam, Robert D., Robert Leonardi, and Raffaela Y. Nanetti. 1994. Making Democracy
Work: Civic Traditions in Modern Italy. Princeton, NJ: Princeton University Press.
Reddick, Christopher G. 2004. Assessing Local Government Revenue Forecasting
Techniques. International Journal of Public Administration 27(8–9): 597–613.
Rose, Shanna. 2010. Institutions and Fiscal Sustainability. National Tax Journal
63(4): 807–38.
Selden, Sally Coleman, Gene A. Brewer, and Jeffrey L. Brudney. 1999. The Role of
City Managers: Are they Principals, Agents, or Both? American Review of Public
Administration 29(2): 124–48.
Shepsle, Kenneth A. 1988. Representation and Governance: The Great Legislative
Trade-Off. Political Science Quarterly 103(3): 461–84.
Simmons, James R., and Solon J. Simmons. 2004. Structural Conflict in
Contemporary Cities. American Review of Public Administration 34(4): 374–88.
Suyderhoud, Jack P. 1994. State-Local Revenue Diversification, Balance, and Fiscal
Performance. Public Finance Quarterly 22(2): 168–94.
Supporting Information
Supplementary appendices may be found in the online version of this article at http://onlinelibrary.wiley.com/doi/10.1111/
puar.13183/full.
14  Public Administration Review  •  xxxx | xxxx 2020
Svara, James H. 1989. Progressive Roots of the Model Charter and the Manager
Profession: A Positive Heritage. National Civic Review 78(5): 339–55.
Wang, Xiaohu, Lynda Dennis, and Tu. Yuan Sen. 2007. Measuring Financial
Condition: A Study of US States. Public Budgeting & Finance 27(2): 1–21.
Watson, Douglas J., and Wendy L. Hassett. 2003. Long-Serving City Managers:
Why Do They Stay? Public Administration Review 63(1): 71–8.
———. 2004. Career Paths of City Managers in America’s Largest Council-Manager
Cities. Public Administration Review 64(2): 192–9.
Wheeland, Craig M., Christine Kelleher Palus, and Curtis Wood. 2014. A Century
of Municipal Reform in the United States: A Legacy of Success, Adaptation,
and the Impulse to Improve. American Review of Public Administration 44(4):
11S–28S.
Zhang, Yahong. 2014. The City Manager’s Role in Policy-Making: A Perspective
beyond Substitution and Collaboration Models. American Review of Public
Administration 44(3): 358–72.
Zhang, Yahong, and Richard C. Feiock. 2010. City Managers’ Policy Leadership in
Council-Manager Cities. Journal of Public Administration Research and Theory
20(2): 461–76.