Does Local Autonomy Breed Leviathans? An Empirical Examination of All Cities in the United States

• Author: Mikhail Ivonchyk
• DOI: http://doi.org/10.1177/10659129211049817
• Published date: 01 December 2022
• Date: 01 December 2022
• Subject Matter: Articles

Article

Political Research Quarterly

2022, Vol. 75(4) 1079–1095

© The Author(s) 2021

Article reuse guidelines:

sagepub.com/journals-permissions

DOI: 10.1177/10659129211049817

journals.sagepub.com/home/prq

Does Local Autonomy Breed Leviathans?

An Empirical Examination of All Cities in

the United States

Mikhail Ivonchyk

1



Abstract

US states grant theirlocal units different levelsof autonomy in several dimensionsincluding f‌iscal, functional,structural, and

legal discretion. This study uses a comprehensive, multidimensional measure of autonomy to test its association with the

f‌iscal behavior of over19,000 municipalities in the UnitedStates. Competing theoreticalpredictions range from signif‌icant

increases in government size (Leviathan model), to no effect (median-voter model), and even smaller governments

(institutional collective action model). Quantile regression analysis is implemented to test the association between au-

tonomy and f‌iscal behavior for different city sizes. The empirical f‌indings indicate that cities with more autonomy tend to

spend less and have lower taxes and debt. The strength of this relationship, however, varies by city size.

Keywords

local autonomy, f‌iscal behavior, Leviathan model, quantile regression, municipal budget, municipal debt, municipal taxes

Introduction and Literature Review

This study uses a multidimensional concept of autonomy

to test its inf‌luence on the f‌iscal behavior of all munici-

palities in the Unites States. The American Constitution

does not mention local governments delegating authority-

giving powers to states (1, 5). For historical, political, and

cultural reasons, states have been granted different de-

grees of local autonomy.

The implications of this variation for city policymaking

have been studied since the late 1800s, but much of the

literature has been “descriptive, classif‌icatory and often

historical”(Krane, Rigos, and Hill 2001, 14)and has relied

on crude measures of autonomy, such as the Dillon’s rule–

home rule dichotomy (Richardson, Gough, and Puentes

2003;Richardson 2011).

The concepts of Dillon’s rule and home rule do not

capture the amount of discretion available to local units

(Richardson 2011), and given the complexities involved in

the distribution of power between state and local units (Liner

1989), it is impossible to use several ordered categories to

accurately measure local autonomy in all states (Krane,

Rigos, and Hill 2001,15–16). Instead, researchers argue

that local autonomy is a continuous, multidimensional

concept without bright-line boundaries (Richardson 2011).

They distinguish legal, functional, personnel, structural, and

f‌iscal dimensions (Chapman 2003;Krane, Rigos, and Hill

2001;Woo d 201 1;Zimmerman 1981,1995)aswellasthe

importance of local governments in the state economy and

the capacity to use their discretion (Wolman 2008,Wolm an

et al. 2010).

Much of the literature has been preoccupied with only

one of these dimensionsof autonomy, f‌iscal discretion,and

has generated many valuable insights (e.g., Dove 2016;

Joyce and Mullins 1991;Maher, Deller, and Amiel 2011;

Mullins and Cox 1995;Mullins and Joyce 1996;Mullins

2004;Mullins andWallin 2004;Mullins andPagano 2005;

Park, Maher, and Ebdon 2018;Park, Park, and Maher

2018;Rown2000;Shadbegi an 1998,1999;Stallmann et al.

2017). The main premise of this study, however, is that all

dimensionsmust be tested in concert to fully understandthe

role of autonomy in local policymaking. There have been

several attempts to systematically operationalize and

measure local autonomy in different contexts (e.g.,Ladner

et al. 2019;Ladnerand Keuffer 2021;Vetter 2007;Wolm an

1

Rockefeller College of Public Affairs and Policy, University at Albany,

State University of New York, Albany, NY, USA

Corresponding Author:

Mikhail Ivonchyk, Rockefeller College of Public Affairs and Policy,

University at Albany, State University of New York, Milne 308, 135

Western Avenue, Albany, NY 12203, USA.

Email: mivonchyk@albany.edu

et al. 2010), but little progress has been made in testing

empirically the inf‌luence of multiple dimensions of au-

tonomy simultaneously. Wood ’s work (2011) is one of the

exceptionshere, which used a four-dimensionalmeasure of

municipal autonomy in the United States and found that

municipalities withmore autonomy tend to raise less taxes

and receive less state aid, with the level of spending and

debt not affected. However, this study used city data ag-

gregated at the state level and omits municipalities with

fewer than 25,000 residents.

In the meantime, there is little empirical evidence to

facilitate the debates about the appropriate level of local

autonomy and its implications for local policymaking.

Higher levels of governments in recent decades have been

imposing additional mandates on municipal governments

expanding their responsibilities without corresponding

increasesin their capacity and authority(Krane, Ebdon, and

Bartle 2004). State constraints on local authority may be

deemed necessaryto avoid local corruption,patronage, and

oppression of opposing views (Brennan and Buchanan

1980;Martin 1990), but the opponents note that lack of

autonomy can makeit diff‌icult for local off‌icials to react to

political, social, environmental, and economic changes

(Chapman 2003), and that more autonomy is necessary t o

be able to fulf‌ill the mandates and adequately meet the

desires of local constituents (Bunch 2014). In the absence

of clear empirical evidence that would help sort out these

opposing claims and elucidate the impact of multiple di-

mensions of autonomy, local policymakers are caught

between evergrowing demands and inadequateauthority to

meet them.

Current study aims to provide several contributions.

First, it uses a composite index incorporating seven di-

mensions ofmunicipal autonomy in all 50 states.Second, it

uses an individual city as the unit of analysis and includes

over 19,000 cities in the United States. Third, given the

diversity of cities in the country and that the effect of state

limitations may vary by jurisdiction size (Mullins 2004;

Rown 2000), this study implements a quantile regression

(QR) analysis allowing a richer characterization of the

testable relationship at different levels of the outcome

variables (Cameron and Trivedi 2009).

Relevant theoretical frameworks offer several alterna-

tive predictions with respect to the association between

municipal autonomy and f‌iscal behavior. The Leviathan

model posits that local off‌icials seek to maximize gov-

ernment revenuesand expenditures to distribute rewards to

their political allies, and more autonomy will be associated

with inf‌lated spending and higher taxes. The Institutional

Collective Action (ICA) framework suggests that more

independence from state restraints may provide the f‌lexi-

bility needed to create mutually benef‌icial collaborations

and generate cost saving, which may be ref‌lected in lower

spending and taxation, holding all elseequal. According to

the median-vote r model, however, democraticall y ac-

countable local off‌icials make policy choices consistent

with voter preferencesand the autonomy inf‌luence on f‌iscal

behavior may be insignif‌icant after accounting for citizens’

demands.

The f‌indings provide empirical support to the median-

voter and ICA models, and offer no support to the Leviathan

model. More specif‌ically, the results suggest that munici-

palities with more autonomy tend to have lower levels of

expenditures, taxation,and debt. These results, as predicted

by the literature, vary by the jurisdiction s ize, but no in-

dications of signif‌icant increases supportive of the Levia-

than model were found in any of the outcomes at any city

size. Consequently, this study supports the notion that,

assuming properly functioning democratic institutions an d

interjurisdictional competition, local off‌icials are unlikely to

runamokandinf‌late city spending in the absence of state-

imposed con straints.

The next section lays out the theoretical predictions

testedin this study.Section three describesthe data, variable

operationalization, and estimation methodology. Empirical

f‌indings are discussed in section four. The paper concludes

with a discussionof the results, relevantpolicy implications,

and opportunities for future studies.

Theory of Local Autonomy and Fiscal Behavior

How can autonomy inf‌luence city f‌iscal choices? Relevant

literature describes several models of local policymaking

providing alternative theoretical predictions about local

autonomy and f‌iscal behavior (Craw 2008;Krane, Ebdon,

and Bartle 2004;McCabe and Feiock 2005).

The Leviathan model ref‌lects the skepticism about

voters’ability to control the government (Brennan and

Buchanan 19 79,1980,1981,1984). This model predicts

that in the absenceof external constraints local off‌icials will

seek to maximize government revenues and use public

resources to reward their supporters and ensure reelection.

They will distribute favors to their political alli es and silence

their opponents (Martin 1990). Government in this model is

akin to an uncontrollable Leviathan with expansionary

tendencies that over time produce grossly inf‌lated budgets.

State-imposed tax and spending limitations (TELs) are the

externalconstraints necessary to preventthe Leviathan from

expanding beyond what the constituents prefer (Bren nan

and Buchanan1980). The f‌inding that TELsare effective in

reducing government size is “interpreted as supportive of

the notion that the Leviathan model may be operative”

(McCabe and Feiock 2005;McGuire 1999, 129). Thus, if

the Leviathan model is relevant, we should expect higher

taxes and expenditures with greater degrees of city au-

tonomy, holding all else equal.

The ICA frameworkpoints to another plausible scenario

(Andrew and Feiock 2010;Feiock 2007,2013;Feiock,

1080 Political Research Quarterly 75(4)