Close
About
FAQ
Home
Collections
Login
USC Login
Register
0
Selected
Invert selection
Deselect all
Deselect all
Click here to refresh results
Click here to refresh results
USC
/
Digital Library
/
University of Southern California Dissertations and Theses
/
Essays on mitigation, adaptation, and resilience to urban fiscal crises
(USC Thesis Other)
Essays on mitigation, adaptation, and resilience to urban fiscal crises
PDF
Download
Share
Open document
Flip pages
Contact Us
Contact Us
Copy asset link
Request this asset
Transcript (if available)
Content
Essays on Mitigation, Adaptation and Resilience to Urban Fiscal Crises
By Manita Rao
A Dissertation Presented to the
FACULTY OF THE USC GRADUATE SCHOOL
UNIVERSITY OF SOUTHERN CALIFORNIA
In Partial Fulfillment of the
Requirements for the Degree
DOCTOR OF PHILOSOPHY
(PUBLIC POLICY AND MANAGEMENT)
August 2022
Copyright 2022 Manita Rao
ii
ACKNOWLEDGEMENTS
I am very grateful to my colleagues and professors for the many lively conversations that
have, over these past years, helped shape and enrich my doctoral journey. I would like to express
my deepest appreciation to my advisor and dissertation committee members for their support and
continued guidance. Juliet, thank you for your unflinching belief in me and my work, for your
counsel in tough times, the appreciation of critical thought, helping me hone the skill of theoretical
minimalism, I truly cherish the friendship we have developed over these years and look forward
to further collaborations. John, I am deeply indebted to you, your mentorship has been instrumental
to my intellectual, and professional growth, thank you for challenging me, pushing me to think
deeper, your persistent focus on quality is inspirational, I am profoundly grateful. Mark, thank you
for being a great advisor, your encouragement of new ideas, and aptitude for fresh perspectives
were highly valuable.
I thank Gary Painter, Matthew Khan, Christian Weller, and Evgenia Gorina, who were
instrumental to helping me transform early-stage ideas into tractable research. I am obliged to
Hashem Pesaran for inspiring an appreciation of econometrics like no other. A huge thank you to
my cohort – Jocelyn, Thai, Sushant, Ann, Greg, Sue, Yougeng, Cynthia, Noah, Arnab, Andrew,
and Ben – for making this journey so much more fun!
None of this would have been possible without my incredibly supportive family. To my
dad, the Gandhian Mathematician, your vision for an equitable world continues to inspire me every
single day, I miss you so very much. I owe a great deal to my extended family of aunts, uncles,
and cousins who supported me in countless ways over the many years. To my kids, Sanna and
Ashu, know that we achieved this together, your compassion, humanity, and individuality are a
constant source of strength. Last but by no means the least, to my friend and confidant - my
husband, Amit – you have always been there for me, thank you for taking care of our family, you
made it look easy, words fall short of the appreciation I have for your effort that has made this
doctoral journey such a memorable experience for all of us. From Sanna, Ashu, and me, you are
our anchor, we love you so much, we have immense gratitude for all that you do every single day.
iii
TABLE OF CONTENTS
ACKNOWLEDGEMENTS .................................................................................................................................. ii
LIST OF TABLES ............................................................................................................................................... iv
LIST OF FIGURES ............................................................................................................................................. iv
ABSTRACT .......................................................................................................................................................... v
Chapter 1: Introduction ........................................................................................................................................ 1
I Inception of Urban Fiscal Crises .................................................................................................................2
II Political Economy of Urban Fiscal Recovery .............................................................................................3
III Overview of Contribution ............................................................................................................................5
Chapter 2: Mitigating Pension Stress? The Impact of Pension Underfunding on Local Government ............. 11
I Introduction ................................................................................................................................................13
II The Role of Pensions in the Public Sector .................................................................................................17
III Theory and Hypotheses .............................................................................................................................19
IV Data and Empirical Design ........................................................................................................................25
V Results and Analysis ..................................................................................................................................30
VI Robustness .................................................................................................................................................39
VII Discussion and Conclusion ........................................................................................................................40
Chapter 3: Performance and Equity Implications of Fiscal Adaptation to Negative Revenue Shocks ............. 42
I Introduction ................................................................................................................................................44
II Theoretical Perspective on Strategic Adaptation .......................................................................................47
III Hypotheses .................................................................................................................................................49
IV Data, Empirical Design, and Diagnostics ..................................................................................................54
V Results and Analysis ..................................................................................................................................60
VI Discussion and Conclusion ........................................................................................................................67
Chapter 4: Resist, Recover, Renew: Fiscal Resilience as a Strategic Response to Economic Recessions ....... 70
I Introduction ................................................................................................................................................72
II Theory: Dimensions and Determinants of Fiscal Resilience .....................................................................73
III Methods, Data and Empirical Strategy .....................................................................................................85
IV Findings and Analysis ................................................................................................................................94
V Concluding Discussion ..............................................................................................................................99
REFERENCES .................................................................................................................................................. 101
APPENDICES .................................................................................................................................................. 113
I Appendix to Chapter 2 .............................................................................................................................113
II Appendix to Chapter 3 .............................................................................................................................118
III Appendix to Chapter 4 .............................................................................................................................123
iv
LIST OF TABLES
Chapter 2:
Table 1. Summary Statistics ......................................................................................................... 29
Table 2. Spillovers to Public Service Expenditure ....................................................................... 31
Table 3. Spillovers by Stress Level ............................................................................................... 32
Table 4. Differential Spillovers by Service Type ......................................................................... 33
Table 5. Spillover Results for Subsamples of Over- and Under-Funded Plans ............................ 34
Table 6. Spillovers to Pension Generosity .................................................................................... 36
Table 7. Mediated Effect of Unionization on Pension Generosity ............................................... 38
Table 8. Robustness Tests ............................................................................................................. 39
Table 9. Group Stability Test ........................................................................................................ 59
Table 10. Summary Statistics ....................................................................................................... 60
Table 11. Fiscal Adaptation Strategy: Effect on Expenditure ...................................................... 62
Table 12. Fiscal Adaptation Strategy: Effect on Non-Tax Revenues ........................................... 65
Table 13. Fiscal Adaptation Strategy: Effect on Fiscal Solvency ................................................ 67
Table 14. Key Concepts in Fiscal Resilience Analysis ................................................................. 89
Table 15. Summary Statistics: Fiscal Recovery and Renewal ...................................................... 91
Table 16. Summary Statistics ....................................................................................................... 92
Table 17. Determinants of Fiscal Resilience [Cox Proportional Hazard Model] ......................... 95
LIST OF FIGURES
Chapter 2:
Figure 1. Differential Spillovers by Pension Stress ...................................................................... 37
Figure 2. Distribution of Referenda Vote Share ........................................................................... 58
Figure 3. Conceptual Framework of Fiscal Resilience ................................................................. 77
Figure 4. Stylized Figure of Fiscal Recovery ............................................................................... 86
Figure 5. Stylized Figure of Fiscal Renewal ................................................................................. 87
Figure 6. Fiscal Trends: Revenue, Expenditure, and Reserves ..................................................... 93
v
ABSTRACT
This dissertation presents three chapters on fiscal decision-making in local government. I
draw on interdisciplinary perspectives from local public finance, urban economics, and political
economy to propose new theoretical frameworks, and empirically evaluate the causal impact of
fiscal crises on urban economic development. The dissertation is anchored by the conceptual
framework of mitigation-adaptation-resilience, each chapter examines a different form of urban
fiscal crisis and the ways in which cities mitigate, adapt, or demonstrate resilience to those crises.
The first chapter traces the historical development of fiscal crises at the sub-national level
with specific attention to key theoretical frames that have informed literary discussions. Chapter
two investigates the impact that the $4.1 trillion in public pension underfunding has on local
governments functions and on public employees. Utilizing a causal framework, I identify the
extent to which unfunded pension liabilities spillover to local service provision and pension
generosity. The paper employs a novel identification strategy that isolates the spillovers, if any,
are affecting local governments and the conditions under which spillovers are mitigated. I find that
unfunded liabilities not only have significant spillovers to service provision, but the effects on
services focused on redistribution are more likely to be impacted. In addition, while there are
spillovers to employee benefits, these effects are, contrary to political economy theory, more likely
to be felt by firefighter and police employee unions that have typically enjoyed higher bargaining
power. The impact on employees, while counterintuitive at first, can be explained by the higher
pensions stress of these pension plans. This paper’s findings inform policy conversations on the
complexity and multiple pathways through which unfunded pension liabilities are impacting long-
term fiscal sustainability at the local level.
vi
The third chapter is a critical perspective on how local governments adapt to abrupt revenue
disruptions. I employ the theoretical frame of fiscal adaptations to identify the ways in which
exogenous negative revenue shocks affect spending, revenue, and fiscal solvency. The second part
of the paper is an assessment of the role of organizational factors such as professional management
and political pressure on fiscal adaptation to the negative revenue shocks. I examine fiscal
adaptations by exploiting a quasi-experimental setting in which the outcome of local fiscal
referenda constitutes an exogenous revenue shock that causes some local governments to witness
a decrease in tax revenues. The paper uses a regression discontinuity framework to estimate the
dynamic Average Treatment Effect (ATE) of negative revenue shocks on local governments. I find
that fiscal adaptations to revenue disruptions are limited to spending behavior with cities largely
reducing their allocations to core public safety services. However, the results diverge greatly by
the organizational characteristics that mediate fiscal adaptation. More specifically, while cities
with professionally trained managers are less likely to employ fiscal adaptation tactics that increase
revenue, cities experiencing political pressures are more likely to rely on fines and fees in order to
adapt to negative revenue shocks. These diverging forms of fiscal adaptations highlight differential
financial management practices in response to unanticipated revenue disruptions and raise
concerns for broader policy discussions on social equity in local public finance.
The final chapter is a theoretical and empirical treatment of fiscal resilience in response to
the pressure of economic recessions on local revenues. This paper which is coauthored with Juliet
Ann Musso and Matthew M. Young introduces a new conceptual framework of fiscal resilience
and empirically estimates determinants of resilience using a panel that cover two prior recessionary
periods. We decompose fiscal resilience in pre-crisis and post-crisis strategic decisions and
estimate cox proportional hazard models to identify factors that enable fiscal recovery over the
vii
short-term and fiscal renewal over the long-term. We find a nuanced pattern wherein the
effectiveness of strategic decisions associated with revenue diversification and countercyclical
fiscal policy are contingent on recession-specific characteristics while structural factors that induce
local investments have higher resilience payoff over the short- and long-term.
Keywords: Local public finance, urban fiscal policy, political economy, public pensions, direct
democracy, fiscal adaptation, fiscal resilience, social equity
1
Chapter 1:
Introduction
2
I Inception of Urban Fiscal Crises
In 1975, New York city was on the brink of bankruptcy. The city had a deficit of over $2
billion, nearly 17% of its general revenue (Gramlich, 1976). The city’s fiscal crisis was attributed
to forces of its political economy – the strength of public employee unions and the city’s welfarist
fiscal policies – and an economic recession. Despite its looming bankruptcy, political campaigns
by city administrators allowed the city to continue borrow from municipal markets at relatively
favorable cost. Further, much of the city’s debt was held by New York city banks leads to fears
that a bankruptcy would have a ripple effect causing a bank run and push the city into long-term
recession. Some scholars argued that the fiscal crisis facing New York city had less to do with
economic conditions or even its political economy but had been caused by racial conflict, class
struggle, an influx of immigrants, coupled with an exodus of wealthy White households which
ultimately eroded the city’s tax base (Castells, 1977; Lichten, 1980; Weaver, 2017).
Rationalizations on whether the urban fiscal crisis of New York city resulted from economic
factors, administrative laxity, political might, or shifting cultural conditions is, to an extent, an
artificial divide. The deterioration of New York city’s fiscal condition was a watershed event that
ushered in an era of market-based, austerity reforms at the institutional level and in municipal
administration.
Between 1970 and 2022, 31 other general purpose local governments faced extreme fiscal
crisis leading them to file for chapter 9 bankruptcy, and 2 cities have filed for bankruptcy more
than once. This includes the city of Detroit which at the time of its 2013 bankruptcy filing had
nearly $18 billion in outstanding debt. The number of similarly large fiscally strapped cities has
increased in recent years. California cities stand out in particular. Three populous California cities
– Stockton, Vallejo, and San Bernardino – filed for bankruptcy between 2008 and 2012 owing to
3
unsustainably high unfunded pension obligations and revenue shortfalls. In 2020, the most recent
bankruptcy filing by a general-purpose municipal government, the city of Fairfield, Alabama faced
structural deficits resulting from loss of population and tax revenue after U.S. Steel – its largest
employer closed operations – in close succession by the closure of a Walmart supercenter. These
bankruptcy filing portend to deeper issues that cause a city to experience extreme fiscal stress
which is in turn indicative of troublesome financial management practices at the state and local
level. While a great deal of research has sought to unearth the forces cause urban fiscal crises,
these discussions neglect an important piece of the puzzle.
II Political Economy of Urban Fiscal Recovery
The papers in this dissertation examine the how cities recover and demonstrate resilience
when facing fiscal crises. Gaining an understanding of the structural, strategic, and systemic
factors that contribute to recovery are important to both theoretical literature in public finance and
hold value for applied public administration. More specifically, I leverage the newer conceptual
frameworks of mitigation, adaptation, and resilience to identify factors that cause urban fiscal
recovery. The papers focus on administrative behaviors and financial decisions adopted by public
agencies to offset fiscal crises and reduce their impact on the core functions of local government.
The papers a key concern overlooked in previous studies – heterogeneity in the forms of
fiscal crises that local governments witness. Unlike several previous studies that have largely
studied the role of economic recessions on urban fiscal crises, I assess the impact of unfunded
pension obligations which are today the foremost fiscal challenge that state and local governments
face and revenue shocks that are induced in the institutional context of local direct democracy
where voters decisions determine continued access to local tax revenues. Furthermore, I employ
4
over 30-years of panel data and empirical strategies that overcome endogeneity bias which enrich
the analyses and provide new insights on the relationship between urban fiscal recovery from
markedly distinct forms of fiscal stressors.
Theories of urban fiscal recovery fall broadly into two categories: capacity-based
approaches and strategy-based frameworks. Under a capacity-based approach, the theory
emphasizes the role of local economic structure in fiscal recovery. Changes in industrial
composition, for example, play a vital role in aiding or inhibiting the ability of a local government
to recover from a fiscal crisis. In this approach, factors that contribute to tax or non-tax revenues
such as changes in changes in home values and business composition constrain fiscal recovery. In
extension, differences between cities in their underlying economic capacity helps explain
differential recovery capacities.
A strategy-based theoretical framework focusses on managerial and administrative
decision making to assess the ability of a government entity to demonstrate resilience and long-
term recovery. Previous research points to factors including whether or not a city has a diversified
revenue portfolio, if the city has adopted a reserve policy, and its service commitments as factors
that contribute to its potential to experience a fiscal crisis. Factors such as these point to an
administrators’ strategic approach to financial preparedness as an important aspect of fiscal
recovery. Such strategic decisions, however, are not divorced from the political contexts of local
public agencies.
Public organizations are inherently resource dependent on political institutions. Such
political dependencies affect revenue stability and make the government entity substantively more
fiscally vulnerable to crises as a result of changes in political leadership or policy priorities. The
papers in this dissertation advance theory and empirics by evaluating the intrinsic link between
5
capacity, strategy, and political economy as they relate to fiscal recovery and resilience. By
studying the success of local governments to recover consistently and from varied forms of fiscal
crises, I shed new light on the persistent issue of urban fiscal crises.
III Overview of Contribution
The three essays in this dissertation are anchored by the conceptual framework of mitigation
adaptation and resilience. I draw on interdisciplinary perspectives to investigate the effects of three
distinct forms of fiscal crises on fiscal recovery and resilience.
The first essay “Mitigating Pension Stress? The Impact of Pension Underfunding on Local
Government” examines the ability of local governments to meet the fiscal challenge of financing
unfunded public pension liabilities. Total unfunded public pensions of are a staggering $5 trillion.
The magnitude of pension underfunding is a cause for concern among both policymakers and
academics. Despite a well-developed literature on factors that caused public pensions to become
underfunding, no previous studies have examined how those unfunded obligations are affecting
other aspects of local government functioning. Further, local governments have received limited
academic attention despite the concentration of unfunded liabilities at the local level.
In this paper, I investigate if unfunded pension liabilities cause fiscal stress and the extent
to which pension stress is mitigated or spills over to the provision of local public services and
whether such spillovers, if any, are secular or concentrated to specific service categories. I further
examine how the stress of unfunded pension liabilities affect the generosity of benefits that public
employees receive. In addition to conducting several robustness tests, these analyses help tease out
the short- or long-term effects of mitigating pension stress and how it differs across the three
outcomes of interest – total expenditure, services that increase social equity, and public employee
6
benefits. Apart from the direct benefit of mitigating pension stress to local fiscal health, the ability
of local governments to adequately finance public pensions has implications on the retirement
security of the nearly 18 million public employees in local governments, many of whom are
covered solely by local public pension plans and do not have access to social security.
The empirical challenge to estimating the effect of unfunded public pensions on any
potential outcome of interest is separating correlation from causation. Unfunded pension liabilities
are endogenous to local expenditure and pension generosity. This makes causal identification a
serious challenge. In this paper, I innovate by employing a novel identification strategy which
helps estimate the causal effect of public pension underfunding on services, equity, and employees.
I provide clean estimates on spillovers to each outcome and the conditions under which those are
mitigated. More specifically, the aper exploits exogenous changes in returns to pension fund
investments as an instrument. This instrument is used to predict the spillover effects of pension
liabilities to all services, equity-enhancing services, and to pension generosity.
I use new panel data on city-sponsored pension plans acquired under a public records
request from the largest pension fund in the US – the California Public Employees’ Pension
Retirement System (CalPERS). These new data contain detailed information on 1,872 distinct
pension plans sponsored by 480 municipal governments in the state of California. This data cover
three-decades from 1991-2018. Over half-a-million public employees including firefighters, police
officers, and general employee receive pension benefits from the plans covered in the dataset. I
estimate fixed effects models and provide forward-looking estimates for over four years after the
initial unfunded liability is incurred by the plan sponsor. .
The findings indicate significant spillovers from pension underfunding to all three
outcomes of interest. Unfunded pension liabilities induce budgetary tradeoffs wherein every $1
7
allocated to paying down those liabilities, causes 0.15 to 0.17 cent decrease in allocation to provide
other public services. These spillovers are observed for all services including public safety and
economic development, but the effect is highest on redistributive spending. Cities are more likely
to increase spending on redistributive services like public housing and transit that enhance equity
when pensions are overfunded compared to if pension obligations are underfunded pointing to the
distributive consequences of unfunded public pension liabilities.
Pension underfunding also affects pension generosity. I find significant spillovers to all
employee groups including firefighters and police officers who have relatively strong unions and
have traditionally enjoyed higher bargaining power. This finding supports the hypothesis that
spillovers to employees depends on the level of pension stress rather political economy theories
which argue that the bargaining power of employee unions insulates them from the adverse impact
of fiscal stress.
I find that spillovers are partially mitigated by asset returns to pension fund investments.
When markets perform well, investment returns increase and mitigate spillovers to services,
equity, and employees. The mitigating effect, however, varies for service expenditure and
employee benefits. Spillovers to services decline while those to pension generosity increase when
liabilities persist for longer periods. Taken together, the findings paint a picture of significant fiscal
concerns resulting from consistent underfunding of public pensions. Spillovers to public safety,
economic development, and equity-enhancing redistributive services has the potential to cause
further strain on the local tax base, and low-income families, in particular. I conclude by discussing
the implications for pension management and human capital in local government.
The second essay “Performance and Equity Implications of Fiscal Adaptations to Negative
Revenue Shocks” critically examines adaptation strategies of local governments to negative
8
revenue shocks to local governments. The abrupt onset of COVID-19 pandemic resulted in a
downward spiral in local revenues. In just over a month, state and local governments lost nearly
40% of total revenue collections. Some of these losses were reversed when congress approved the
Coronavirus federal assistance package. However, the unexpected negative revenue shock shed
light on a underexamined question in local public finance.
How do local governments adapt to abrupt disruptions that cause substantial revenue loss?
An even more interesting scenario occurs when institutional forces such as voters’ decisions on
local fiscal referenda induce negative revenue shocks. When institutionalized a fiscal disruption is
both, paradoxically, unexpected and routinized. In this paper, I exploit a quasi-experimental setting
to examine the strategies that cities adopt when adapting to institutionally induced negative
revenue shocks. I assess the direct effect of the shock as well as the mediating effect of political
economy factors in the agency’s internal environment on expenditure, non-tax revenues, and fiscal
solvency.
Drawing on theories of adaptative governance, where an ongoing debate concerns the
benefits to public organizations from adapting to revenue disruptions as opposed to its potentially
harmful effects on internal stability, the paper proposes hypotheses that help disentangle the role
of shocks vs. political economy variables on fiscal adaptation.
The empirical strategy leverages a quasi-experimental setting in which voters decide on
local tax and bond referenda proposed by the concerned local government. Referenda that receive
favorable votes from a majority or supermajority of the voters are approved while the remaining
are rejected. A dynamic regression discontinuity model is estimated which uses referenda vote
share as the running variable. The dataset consists of the universe of 2,350 tax and bond
9
referendums proposed by 480 California municipal governments and covers a three-decade period
from 1991 to 2018.
Findings indicate that fiscal adaptation through the direct channel is confined to decline in
total expenditure with no significant effect detected for either non-tax revenues or fiscal solvency.
However, when negative revenue shocks are mediated by political economy factors, the results
differ considerably. I examine three mediating factors – professional management, political
pressure from employee unions, and the stress of unfunded pension liabilities. The findings reveal
that professionally managed cities are more likely to adapt by decreasing spending levels and this
includes reducing expenditures public safety services and are less likely to increase non-tax
revenues from either enforcement fines or service fees. Political pressure, on the other hand, is
associated with increase in service expenditure as well as higher enforcement fines and significant
decline in fiscal solvency. A similar pattern of results is revealed when two stressors co-occur –
negative revenue shocks and pension underfunded - are observed. This pattern of findings indicates
that while professionally managed cities are more likely to adopt adaptative strategies that confer
with prudent financial management practices, both political pressure and co-occurring fiscal stress
obviates those gains.
A particularly important finding is that abrupt negative revenue shocks are associated with
significant increase enforcement fine revenues. Since enforcement fines are disproportionately
borne by low-income households and racial minorities, they are not only regressive but also hurt
social equity goals and indicate a form of fiscal maladaptation. The paper concludes with a
discussion on deeper examination of the potentially tenuous relationship between local fiscal
policy, social equity, and fiscal responses to crises.
10
The third essay “Resist, Recover, Renew: Fiscal Resilience as a Strategic Response to
Economic Recessions” is co-authored with Dr. Juliet A. Musso and Dr. Matthew M. Young.
Economic recessions have, in recent years, increased in frequency, length, and intensity. There are
not only more recessions but economic recessions such as the one induced by COVID-19 has
persisted for a longer period. Given the fiscal uncertainty that recessions cause, this paper centers
the theoretical frame of resilience and investigates the factors that contributed to local fiscal
resiliency from two prior recessions – the dotcom recession in 2001 and the Great Recession which
lasted from 2008-2010. We provide a conceptual framework that decomposes fiscal resilience into
pre-crisis resistance capacity, post-crisis recovery, and long-term fiscal renewal. The empirical
models analyze three distinct types of determinants – strategic, structural, and economic. This
approach identifies the marginal impact of each contributing factor to enhancing fiscal resiliency.
The findings indicate that strategic decisions associated with revenue diversification and
countercyclical capacity facilitate fiscal resilience but only when recessionary pressures affect a
single revenue stream. These strategies are less effective when the overall local economy slows
down and multiple revenue streams are impacted. We also find that local revenue and service
structures play an important role in enabling fiscal resilience. An important finding is the nuanced
relationship of property taxes and financial recovery. While previous studies indicate that the
stability provided by property taxes to local revenue portfolios aid recovery, we find that these
beneficial effects were limited to the Dotcom Recession which did not affect local property values
but did not extend to the Great Recession which was caused by a property bubble. Capital
investment in local infrastructure, however, are a consistent positive predictor of fiscal resilience.
This dissertation advances theory and empirical knowledge on public finance, political
economy, and urban economics.
11
Chapter 2:
Mitigating Pension Stress?
The Impact of Pension Underfunding on Local Government
12
Mitigating Pension Stress?
The Impact of Pension Underfunding on Local Government
Manita Rao
Ph.D. Candidate, Public Policy and Management
Sol Price School of Public Policy
University of Southern California
Date: 01/11/2022
Abstract
The shadow of unfunded pension liabilities has raised concerns on its consequences for citizens
and cities. Some have argued that unfunded pensions will crowd out urban services while others
propose that the effects are limited to public employees. This article empirically examines the
extent to which the stress of unfunded pensions is mitigated or demonstrates spillovers to services,
equity, and employees. Employing a novel identification strategy and over two decades of new
data on city-sponsored pensions (1997–2018), I provide forward-looking causal estimates on the
short- and long-term effects of unfunded pensions on cities. I find that unfunded pensions have
significant spillovers to urban services including public safety and economic development with
the most acute impact on redistribution. Further, decline in pension generosity indicates spillovers
to public employees, and contrary to political economy theory, pension stress explains impact on
unions. Although spillovers are partially mitigated by asset returns to pension fund investments,
the effects are highly persistent.
Keywords: Public pensions, local government, public unions, pension management, social equity
13
I Introduction
The significant size of pension liabilities in American cities has given rise to the concern that
cities will cut back their essential services to meet their pension obligations. Over the past two
decades, the funded status of state and local pensions has declined by nearly 28%. In 2021, state
and local retirement systems had a median funded ratio of 72%
1
. The public pension funding gap
is estimated at $3.7 to $4.3 trillion, nearly a fifth of the US GDP (Rauh, 2017; Novy‐Marx &
Rauh, 2011 ).
Despite the magnitude of pension underfunding and the immense task that underscores the
management aspects of pension liabilities, there is limited research on the impact of underfunded
pensions on local government. This has left us largely in the dark about the manner in which
underfunding of the 5,900 locally sponsored pension plans may be affecting public spending,
more broadly, and specifically, essential services such as public safety and redistribution that
are provided by local governments.
The purpose of this paper is to provide forward-looking causal estimates on the extent to
which the stress of unfunded public pension liabilities is mitigated or exhibits spillovers to
public service expenditures and the implications of fiscal spillovers, if any, for social equity.
Further, I investigate whether underfunded pensions impact public employees by examining
spillovers to pension generosity and the mediating effect of unionization on mitigating spillovers
to employees. Using data on the three main municipal employee groups – firefighters, police
officers, and general employees – this paper contributes to political economy theory by testing the
effects of differential bargaining power on spillovers to public employees’ pension benefits.
1
Reason Foundation; https://reason.org/data-visualization/how-state-pension-funding-ratios-have-declined-over-
time
14
The empirical challenge in estimating the impact of unfunded pension liabilities on local
government is separating correlation from causality. The simple observation that a city that
experiences underfunding and cuts spending at the same time does not establish that the
underfunding caused the cuts - it could have been the case, for example, that a recession caused a
low return on fund assets and at the same time a decline in tax revenues. This would increase
unfunded pension liabilities while also decreasing the tax revenues required to finance public
services. In this paper, I solve the empirical problem by exploiting exogenous shocks to pension
fund investments as an identification strategy for unfunded pension liabilities. I instrument for
increase (decrease) in pension liabilities with market value of pension plan assets to identify the
causal effect of underfunded pensions on urban outcomes. This strategy provides clean causal
estimates in addition to isolating the relative salience of unfunded public pension liabilities to
citizens, low-income households, and public employees.
I use new data on 1,842 unique city-sponsored pension plans that provide benefits to
half a million firefighters, police officers, and general municipal employees and cover a two-
decade period from 1997 to 2018. These data which were procured through a public records
request from the largest public pension fund in the US – the California Public Employees’
Pension Retirement System (CalPERS) – are the most fine-grained and detailed information on
local sponsors of public pensions. Given the paucity of research on local government pensions,
this new data source considerably advances current knowledge on local public pensions. The
analysis is conducted by constructing a plan-level dataset that integrates detailed pensions data with
fine-grained fiscal information on sponsoring city governments. It is important to note that that
since the dataset is comprised primarily of contracted plans, potential confounds from the discount
rate and investment portfolio are obviated which strengthens the identification strategy.
15
The main finding is that an unexpected increase in a pension fund’s assets spills over to
more public spending, and conversely for an unexpected decline in asset values. I find that the
spillover effect is in the range of 1.4 to 2.2 percentage points. This translates $1.7 million decrease
in public service expenditure for a mid-sized city with 50,000 residents and $14 million for larger
cities that have 400,000 residents. The estimated budgetary tradeoff induced by unfunded pension
liabilities is 15% to 17%. In other words, for each $1 allocated to unfunded pensions, the
allocation to public services will decrease by 15 to 17 cents.
The spillover effect extends to service categories including spending on public safety,
economic development, and redistribution. The average spillover to public safety is 1.7 percentage
points and 1.8 percentage points for economic development. However, the direct spillovers to
service categories provide a partial picture of the spillover effects. A more complex picture
emerges when I allow spillovers to depend on the level of pension stress – a pension plan’s
funded ratio. Using subsamples of pension plans that are overfunded vs. underfunded, I find
that cities are more likely to provide redistributive services when pensions are overfunded, the
marginal propensity of budgetary allocation to redistribution is significantly lower for cities
with one or more underfunded pension plan. The findings also reveal decline in redistributive
expenditures across all levels of pension stress - severe, high, and low - indicating that unfunded
pension liabilities have distributive consequences.
Unfunded pension liabilities also impact public employees’ pension benefits, and
spillover effects are magnified over the long-term. More importantly, contrary to political
economy theory which conjectures that unionization insulates public employees from potential
adverse impacts of decline in fiscal health (Anzia & Moe, 2015; Dippel & Sauers, 2019;
DiSalvo & Kucik, 2018), I find that employee groups with high relatively high bargaining
power – firefighters and police officers – have witnessed decline in generosity. These results
16
can be explained by the higher level of pension stress associated with firefighter and police
pensions.
On the positive side, spillovers to both service expenditures and pension generosity are
partially mitigated by asset returns to pension fund investments. Asset returns are especially
salient to preserving public employees’ pension benefits. While the mitigating effect for service
expenditures is 0.22 percentage points, the effect on pension benefits is substantively higher and
nearly 7 percentage points.
This study sheds light on an unexamined research question on public pension puzzle in
local government: how unfunded pension liabilities are impacting cities, citizens, social equity,
and public employees. While prior research has examined the factors that led to the rise of
unfunded pension liabilities including fiscal stress (Chaney et al., 2002; Splinter, 2017; Thom &
Randazzo, 2015); interest group pressure (Anzia & Moe, 2015; Frandsen & Webb, 2017; Hoang
& Goodman, 2018); and political opportunism (Anzia & Moe, 2019; Chen, 2018; Stalebrink &
Donatella, 2021), these theories are challenged by empirical evidence which indicates pension
liabilities are largely an outcome of legacy decisions (Dye & Gordon, 2012) and economic
downturns (Munnell et al., 2008), as opposed unreasonable union demands or imprudent
investment decisions (Weller & Wenger, 2009; Wang & Peng, 2018).
I argue, in contrast, that notwithstanding the determinants of unfunded pension liabilities,
the extent to which asset returns increase (decrease) pension stress have consequences for citizens
when liabilities spillovers to service expenditures, on low-income households as a result of the
impact on redistributive services, and for human capital in the public sector due to spillovers to
employee benefits. The findings in this paper reveal that spillovers caused by unfunded public
pension liabilities are adversely impacting all groups including citizen-taxpayers, low-income
households, and public employees indicating implications for pension sustainability.
17
II The Role of Pensions in the Public Sector
Public pensions are vital to human capital in the public sector. Wages in the public sector have
typically been lower than those of similarly qualified private sector employees (Katz & Krueger,
1991; Lewin et al., 2011). These wage differentials are compensated for under the deferred
compensation model of public sector defined benefit plans that provide guaranteed income over
the course of an employees’ post-employment retirement years. Previous research shows that
state level pension reforms that reduced pension generosity have negatively affected the ability
of public employers to attract and retain high quality talent (Gorina & Hoang, 2020; Quinby &
Sanzenbacher, 2020).
Apart from the salience of public pensions for human capital, pension benefits are the
primary source of retirement income for public employees, most of whom are excluded from
receiving benefits under the federal social security program. More importantly, adequate
financing of public pensions is vital to increasing retirement security of racial minorities. It is
estimated that while 48% of White households are at risk of retirement security, 54% of African
American and 61% of Latinos households experience substantive financial risks associated with
retirement security (Munnell et al., 2018). In addition, given the high representation of racial
minorities in the public sector workforce - 45% of all Black women and 21%
2
of all Black men
– fully funded public pensions are essential to reducing the racial wealth gap. Underfunded
pension liabilities undermine the ability of the public sector to contribute toward these broader
racial equity goals.
2
This information is provided by the Center for American Progress for the period October 2020. Accessed at
https://www.americanprogress.org/article/public-work-provides-economic-security-black-families-communities/
18
There are three main sources of public pensions revenues: returns from pension fund
investments, employee contributions, and employer contributions. 64% of pension revenues are
generated from investment returns and the remaining 36% are from employer and employee
contributions (Matkin et al., 2019). Among these, investment returns are the most volatile,
employee contributions the most stable and, although, in theory, employer contributions should
bridge the gap between returns and employee contributions, evidence suggests that employers
decrease contributions during economic recessions to manage other budgetary demands (Chaney
et al., 2002; Thom & Randazzo, 2015). Underfunded pension liabilities will escalate in response
to sustained levels of employer under-contribution or when recurrent economic recessions lead
to lower-than-expected returns to pension investments.
Apart from the impact of economic conditions on pension revenues, the political
economy context of public pensions can also increase pension generosity and lead to pension
underfunding. The influence of public unions through collective bargaining, political
contributions, and the political clout of public employees as a voter base are found to be
associated with higher pension benefits (Anzia & Moe, 2015; Rauh, 2009; Dippel & Sauers,
2019). In addition, the technical complexity of pension benefits obscures pension benefits from
voter oversight thereby expanding the scope for corruption and political opportunism (DiSalvo
& Kucik, 2018; Glaeser & Ponzetto, 2014; Liu et al., 2021).
These political economy perspectives are challenged by the theory that public pensions
are underfunded due to legacy liabilities that resulted from political inference, and myopic
legislative decisions during the late 1990s (Eaton & Nofsinger, 2004; Munnell et al., 2015;
Weller & Wenger, 2009). For example, Eaton and Nofsinger (2004), cite the case of California
in the 1990s, when then Governor Pete Wilson threatened to gain control over California’s
retirement system, CalPERS, by replacing employee representatives on the governing board with
19
political appointees. CalPERS offset the bid for political control by offering $1.6 billion to
balance the state budget.
Although such overt instances of political interference are relatively infrequent, politically
motivated investment decisions such as directing Economically Targeted Investments (ETIs) to a
politician’s electoral precinct, or investment in local Real Estate Investment Funds (REITs) will
alter the optimal return maximizing investment portfolio thereby increasing unfunded pension
liabilities (Sundén & Munnell, 1999, Andonov et al., 2018).
Pension liabilities are also highly susceptible to discount rate choices of pension
governance boards. Technically, a higher discount rate will reduce actuarial pension liabilities
while lower discount rates increase liabilities. Research shows that governing boards comprised
retired employees and those with a higher share of unionized employees select higher discount
rates while boards with political appointees and active employees are more likely to select lower
discount rates (Wang & Peng, 2018; Brooks, 2019; Anzia & Moe, 2019).
Despite the politicized nature of discount rates and pension investment portfolios, in
principle, nominal pension liabilities are less of a policy challenge while real pension liabilities
are a fiscal challenge to the extent that liabilities outpace economic growth. The intrinsic
complexities of public pension systems require disentangling the mechanisms that potentially
transmit unfunded pension liabilities to urban outcomes.
III Theory and Hypotheses
Why would the stress of unfunded pension liabilities affect public services and in what
ways can unfunded pensions spillover to citizen-taxpayers and public employees? Theoretically,
the disassociation between decision-making and liability risks creates moral hazard problems
20
in public pension systems. A principals’ fiduciary responsibilities should normatively precede
political considerations in governance, discount rates choices, or investment portfolios.
Infractions from those normative goals will result in unfunded liabilities that increase financial
risks to plan sponsors who can in turn transfer those risks to either public employees, or citizen-
taxpayers, or both parties. I examine how liabilities risks are transferred to either party, I
disentangle two distinct mechanisms - budgetary tradeoffs, and employee benefits – to isolate their
independent impact on citizen-taxpayers and public employees.
3.1 Underfunded Pensions, Fiscal Pressure, and Budgetary Tradeoffs
The majority of public employers provide Defined Benefit (DB) pensions that provide
employees post-retirement guaranteed monthly payment. In the ideal scenario, public pensions
are fully funded and impose no fiscal pressure on plan sponsors. In the event that pensions are
underfunded, plan sponsors are under financial pressure to make funds available in order to
fulfill promised benefits. However, maintaining adequately funded pensions are one among the
multitude of budgetary demands on local governments.
Arguably, the primary role of local governments is to ensure stable provision of essential
public services as opposed to maintaining adequately funded pensions. As underfunded pension
liabilities increase, in a hypothetical zero-sum scenario, budgetary tradeoffs induce the marginal
dollar allocated to pensions to be associated with decline in the marginal allocation to service
expenditures. However, because the complexity of local allocative decisions surpasses a
simplistic zero-sum scenario, there is conflicting evidence on how pension obligations affect
budgetary decisions. In principle, fiscal pressure from unfunded pension liabilities can have
negligible impact on service expenditures when revenue shortfalls are offset by reductions in
pension generosity. Alternatively, pension liabilities will increase when service expenditures
21
and employee benefits remain unchanged. In practice, budgetary decisions are influenced by
both institutional and market forces.
On the institutional side, DB benefits are constitutionally protected mandates which
makes it essential for plan sponsors to maintain fund sufficiency and also limits sponsors’
flexibility to modify promised benefits. Chaney et al. (2002) find that fiscal institutions such as
balanced budget requirements are associated with higher unfunded pension liabilities in addition
to the selection of optimistic discount rates which in turn increase budgetary pressures. And,
Eaton and Nofsinger (2004) show fiscally stressed plan sponsors attempt to reduce short-term
budgetary pressures by adopting more lenient actuarial assumptions such as a longer
amortization period which shifts pension liabilities to future periods. These forms of accounting
maneuvers to offset short-term budgetary pressures are pervasive in cities with low fiscal
capacity that not only have poorly funded plans but also a weaker economic base (Thom &
Randazzo, 2015; Splinter, 2017). Because ensuring compliance with institutional mandates
while at the same time managing pension contributions increases fiscal pressures, plan sponsors
also encounter budgetary tradeoffs.
Capital markets play a critical role in public financing of municipal governments.
Research shows that the consequence of a municipal sponsors high unfunded pension liabilities
are deterioration in credit quality and higher debt costs. Raman and Wilson (1990) find that
credit rating agencies assess municipal issuers with high pension liabilities to be at increased
default risk resulting in issuers facing higher debt costs across all issuances including general
obligation, revenue, as well as short-term bonds. It is estimated that the credit rating for plan
sponsors with severely underfunded pensions – 40% or lower funded ratio – is nearly 80% lower
than that for plan sponsors with overfunded plans (Martell et al., 2013). In the worst case
scenario, plan sponsors that issue high-risk instruments such as pension obligation bonds will
22
incur overleverage risks due to the risk premia associated with these instruments (Calabrese,
2010; Aubry et al., 2017). Therefore, market factors function in a manner similar to institutional
mandates, both induce budgetary tradeoffs when unfunded pension liabilities escalate.
Although there is an extensive body of literature on the consequences of underfunded
pension liabilities on a plan sponsor’s fiscal health, there is limited research on how local
governments manage such budgetary tradeoffs. Traditional cutback management theory
suggests that in times of fiscal strain, local governments should minimizing disruption to core
functions such as public safety and economic development while shifting expenditure cutbacks
to non-core functions including providing fewer redistributive services
3
(Hendrick & Degnan,
2020; Jimenez, 2014). Such a strategy is economically rational from a fiscal federalism
perspective, the negative externalities associated with local provision of redistributive services
as opposed to public safety underpins the normative assignment of redistributive provision to
higher level governments and public safety provision to lower level governments (Peterson &
Rom, 2010; Tiebout, 1956; Musgrave, 1971; Oates, 1999). Previous research shows that during
economic recessions, cutbacks to public safety are a last resort strategy, commonly preceded
by alternate strategies including increasing tax revenues, debt issuance, and reduction in capital
spending (Nelson, 2012; Scorsone & Plerhoples, 2010; Shoag et al., 2019; Kim & Chen, 2020;
Raudla et al., 2017). Although cities with concentrated poverty have the highest need for
redistribution, these cities are especially prone to reducing redistributive services and retaining
public safety or economic development in response to declining revenues and increased fiscal
pressure (Joassart-Marcelli et al., 2005; Collins, 2008). Taken together, there is wide consensus
3
In general, locally provided public services fall into three broad categories: public safety, economic development,
and redistribution (Peterson, 1981). Public safety incudes policing, fire, and Emergency Medical Services (EMS).
Economic development incentivizes economic growth and tax base expansion. Redistributive services include
services such as public transit, public housing, and public health that meet the needs of low-income households.
23
that unfunded pension liabilities impose fiscal pressure on local plan sponsors. To the extent
that underfunded pensions also induce budgetary tradeoffs, plan sponsors can be expected to
engage in strategic fiscal responses that vary by service functions. The following hypotheses
are proposed:
H1(a): Spillovers from underfunded pensions will decrease public service expenditures.
H1(b): Spillovers from underfunded pensions are less likely to decrease expenditure on public
safety and economic development; and more likely to decrease redistributive expenditure.
3.2 Pension Stress, Union Pressure, and Benefit Generosity
Nearly 40% of public employees are unionized pensions are negotiated through collective
bargaining. This makes unions important stakeholders in pension policy. Public choice
theorists argue that demand-side pressures exerted by public unions has led to increase in
pension generosity and subsequently to the fiscal challenge of adequately financed pensions
(Glaeser & Ponzetto, 2014; Kelley, 2014). This is borne out by evidence which indicates that states
with collective bargaining rights have a larger public sector and higher wages compared to states
without collective bargaining (Zax, 1989; Freeman & Valletta, 1987; Valletta, 1993). Apart from
collective bargaining, pension generosity has been attributed to factors such as the relatively high
technical complexity of pensions which limits voter oversight, union campaign contributions,
and public employees’ voter base (Anzia & Moe, 2015; Glaeser & Ponzetto, 2014). Recent
studies estimate an average public sector union wage premium of 6 to 11 percent when not
accounting for employee specific human capital (Brunner & Ju, 2019; Crowley & Beaulier,
2018).
A major component of pension generosity is Other Post-Employment Benefits (OPEB)
that include health care, occupational injuries, and related costs. OPEBs alone constitute over a
24
third of all unfunded pension liabilities at the local level (Munnell & Aubry, 2017). Between 1960
and 2010, OPEB spending increased by 21% in collective bargaining states, or by $3.5 billion
compared to $2 billion in non-collective bargaining states (Frandsen & Webb, 2017). OPEB
spending has increased across partisan lines, union political contributions are associated with
higher OPEB expenditures in both Democratic and Republican states (DiSalvo & Kucik, 2018). In
addition, because protective service employees have more generous OPEBs, cities that have both
collective bargaining and in-house protective services have witnessed increased OPEB spending
(Matkin & Krivosheyev, 2013).
Although there is relative consensus that unionization is associated with higher public
sector wages and employee benefits, there is also evidence that unions increase prudent
management of public pensions (Weller & Wenger, 2009; Mitchell & Smith, 1991), in addition to
research that shows unionized cities are no less likely than non-union cities to decrease pension
contributions even under fiscal strain (Thom & Randazzo, 2015). Factors such as fiscal capacity,
revenue diversification, and fiscal constraint are more likely than unionization to predict
contribution volatility and pension underfunding (Gorina, 2018; Thom & Randazzo, 2015;
Coggburn & Kearney, 2010). Pension benefits are also highly sensitive to economic conditions
including revenue shortfalls during an economic recession even for employee groups that have
relatively high bargaining such as firefighters, police officers, and teacher unions (Hoang &
Goodman, 2018).
This evidence conflicts with the notion that union pressure is one-sided and raises the
possibility that pension underfunding is more likely to adversely affect public employees. Recent
legislative reforms targeted toward restricting collective bargaining rights such as Wisconsin’s
Act 10 passed in 2011 have led to debates on the true intent of such interventions. For example,
Freeman and Han (2012) argue that Wisconsin’s bid to close state budget deficits by restricting
25
collective bargaining rights was motivated by political opportunism rather than pubic union
intransigence in responding to state fiscal pressures. In addition, the absence of union resistance
to pension reforms provides evidence of union support for the prudent management of public
pensions (Abott, 2015; Munnell et al., 2011). Taken together, there is contrary evidence on the
role of public unions in the puzzle of pension underfunding. The following hypotheses are
proposed:
H2(a): Spillovers from underfunded pensions will decrease pension generosity.
H2(b): Spillovers from underfunded pensions to pension generosity will diverge by union
bargaining power.
IV Data and Empirical Design
4.1 Data
This study uses new administrative data on 1,842 distinct city-sponsored pension plans covering
a two-decade period from 1997 to 2018. The data was procured through a public records request
from the largest pension fund in the US – the California Public Employee’s Pension Retirement
System (CalPERS). 96% of city governments in California contract with CalPERS for pension
governance and administration.
4
The remaining 4% are self-administered pensions that are
provided by large cities including Los Angeles, San Diego, San Jose, Fresno, and Oakland. The
analysis in this paper is conducted on a plan-level panel dataset that consists of both CalPERS-
contracted and self-administered plans. The dataset integrates detailed plan-level pensions data
with fine grained fiscal information on sponsoring local governments. The information on
pensions includes the present value of pension benefits, Market Value of Assets (MVA), plan
4
It administers pension benefits to over 1.6 million current and retired public employees of city, county, and special
district employees in California. The fund also has over $160 million in unfunded pension liabilities. The average
funded level since the Great Recession has been around 70%.
26
membership, and the Unfunded Actuarial Accrued Liabilities (UAAL). These plans provide
benefits to over half-a-million firefighters, police officers, and general municipal employees.
I analyze spillovers to total expenditure on all public services in addition to spillovers to
service categories. Based on Peterson (1981), three functional service categories are examined:
(a) public safety which includes city spending on policing, fire, and Emergency Medical Services
(EMS); (b) economic development that includes spending on redevelopment, highways, land use
planning, and infrastructure; and (c) redistribution which covers local expenditures on public
housing, transit, health, parks, libraries, and museums. The dataset also includes an extended set
of control variables from the American Community Survey (ACS) and historical data from the
IPUMS NHGIS files.
5
Plan-years that were outliers on employer were excluded. The analysis
also excludes San Francisco, which is both a city and a county as well as smaller cities with fewer
than 100 residents: Sand City, Vernon, and the City of Industry.
4.2 Empirical Strategy
The empirical challenge of estimating the impact of underfunded pensions on public services is to
identify its causal effect. Both service expenditure and benefit levels are simultaneously
determined which causes an endogeneity problem. For example, an economic recession could
cause both decline in local revenue which would lead to decrease in service expenditures while at
the same time increase unfunded liabilities due to decline in asset returns. Identifying the causal
effect of underfunded pensions requires an instrument that is correlated with pension liabilities but
not expenditure on public services. I solve the empirical problem by exploiting exogenous shocks
to pension fund investments as an identification strategy. More specifically, I instrument for
5
NHGIS data is publicly available at https://www.nhgis.org/
27
unfunded pension liabilities with asset returns to pension fund investments which provides clean,
causal estimates. Given the exogeneity of asset returns, the identification strategy helps isolate
spillovers from underfunded pensions to services, equity, and employees in the study context city-
sponsored public pensions.
However, endogeneity will be a concern when a plan sponsor decides on the discount rate
the pension investment portfolio. I solve this problem by conducting the analyses on the subsample
of pension plans that are contracted with CalPERS for governance and administration. Because
the local sponsors of contracted plan outsource discount rate and investment decisions to the
contracted agency – CalPERS – confounds from these sources are obviated.
I also conduct several robustness tests and estimate spillovers for several future periods to
test for directional stability and effect size. The findings indicate results are highly robust across
all specifications and all future period that were tested.
The main empirical strategy is estimating dynamic fixed effects models to examine the
direct effect and mediated effects of pension underfunding on the outcomes of interest.
Specification 1: Direct Effect of Unfunded Pensions on Outcomes
𝑌
!(#$%)
= ⍺+𝜏𝐴𝑆
!#
+𝜆𝑈𝑃
!#
+𝜃(𝐴𝑆 × 𝑈𝑃)
!#
+𝛽𝑋
!#
+𝜂
#
+𝛿
'
+𝜀
!#
…(1)
where 𝑌
!(#$%)
is the dependent variable for pension plan 𝑖 in year (𝑡+1). 𝑌
!(#$%)
takes
the value of the following outcomes of interest - total expenditure, public safety expenditure,
economic development expenditure, redistributive spending, and the present value of pension
benefits per member. Asset Shocks (𝐴𝑆
!#
) are proxied by the Market Value of Assets (MVA) for
plan 𝑖 at time 𝑡. 𝑈𝑃
!#
are Unfunded Actuarial Accrued Liabilities (UAAL) for plan i at time t.
The interaction of Asset Shocks and Underfunded Pensions is given by the coefficient of
28
interest, 𝜃, which estimates the causal effect of unfunded pension liabilities on outcomes of
interest. 𝑋
!#
is a matrix of institutional and control variables. 𝜏, 𝜆, and 𝛽 are coefficients to be
estimated for asset shocks, underfunded pensions, the vector of control variables. I also include
fiscal year fixed effects (𝜂
!
) and fixed effects for a 2014 pension reform which introduced a
new tier of pension plans for new hires (𝛿
'
). 𝜀𝑖𝑡 is the idiosyncratic error term.
Specification 2: Union-Mediated Dynamic Fixed Effects Model
𝑌
!( (#$%)
= ⍺+𝜏𝐴𝑆
!#
+𝜆𝑈𝑃
!#
+𝜃(𝐴𝑆 × 𝑈𝑃)
!#
+⍴
%
𝑈𝑛𝑖𝑜𝑛
(
+⍴
)
(𝑈𝑃
!#
×𝑈𝑛𝑖𝑜𝑛
(
)
+ 𝛽𝑋
!#
+𝜂
#
+ 𝛿
'
+𝜐
!(#
…(2)
equation (2) introduces an indicator variable for unions (𝑈𝑛𝑖𝑜𝑛
(
). More specifically,
the study covers the three main unions at the municipal level - firefighters, police officers, and
general municipal employees.
6
⍴
%
estimates the main effect of unionization and coefficient of
interest, ⍴
)
, is the mediating effect of unions conditional on plan underfunding. 𝜐
!(#
is the error
term. The remaining terms are similar to equation (1).
4. 3 Summary Statistics
Table 1 presents summary statistics. The average pension plan in the study sample has an
underfunded liability that 26 times higher than service expenditures. Underfunding in firefighter
pensions is 22 times higher than total expenditure, it is 16 times higher for police pension plans
and 8 times lower for general municipal employee plans.
6
I exclude pension plans in which both firefighters and police employees are beneficiaries which permits estimating
the independent influence of each union.
29
Pension plans in the sample differ between employee groups in membership and
generosity. On average, plans have 396 participants, but firefighter plans have an average of 72 72
Table 1. Summary Statistics
Mean SD Min Max Definition
Total Expenditure 1,626 1,114 10.6 14,227 Total expenditure per capita
Functional Services
Public Safety 597 451 0 4,381 Public safety expenditure per
capita
Economic Development 346 251 364 6,035 Economic development
expenditure per capita
Redistribution 315 345 0 6,510 Redistributive expenditure per
capita
Public Employees
Pension Benefits 230,871 177,414 370 1,296,400 Present value of pension benefits
per member
Pension Performance
Asset Shocks
138,358 125,157 0 902,395 Market Value of Assets (MVA)
per member
Unfunded Pensions
26,120 52,282 -219,448 446,588 Unfunded Actuarial Accrued
Liabilities (UAAL) per member
Political Pressure
Firefighter union 0.15 0 1 Dummy = 1 for firefighter union
Police union 0.22 0 1 Dummy = 1 for police union
General employee union 0.36 0 1 Dummy = 1 for general municipal
employee union
Safety employee plans 0.27 0 1 Dummy = 1 for safety employees
Institutional Factors
Home Rule 0.21 0 1 Dummy = 1 for charter city
Control Variables
Population 49,809 168,252 198 4,013,170 City size
Percent Democrat 0.43 0.11 0.16 0.82 Registered Democrats
Percent Hispanic 0.32 0.24 0.14 0.98 Percent Hispanic
Percent African American 0.038 0.05 0 0.53 Percent African American
Unemployment rate 0.08 0.04 0.008 0.45 Percent of unemployed persons
Median household Income 72,986 32,881 24,318 291,647 Median household income
Housing tenure 0.35 0.17 0.035 0.97 Percent owner-occupied housing
Note: Expenditure variables are per-capita 2018 constant dollars. Pension plans that include both firefighter and police employees
are excluded from the empirical analysis that identify union-specific effects. Negative value on pension liabilities occurs when plans
are overfunded.
30
participants and police officer plans have 88 participants which is substantively lower than the
average 497 participants in pension plans for general employees. In addition, pension benefits for
general employees receive is, on average, 265% lower than benefits for protective service
employees.
V Results and Analysis
5.1 Effect of Underfunded Pensions on Public Services and Social Equity
Table 2 presents results from models estimated based on equation (1). The regression in
column (1) shows the unconditional effect of a positive asset shock on public services. Asset
returns have a significant effect on total expenditure, the effect is significant at the 0.1% level.
The coefficient of 0.11 means that for each 1 percent increase in asset returns, total
expenditure will increase by 1.1 percentage points and conversely for a decline in asset values.
This translates to an average increase of $17 per capita at the mean and $21 per capita at the 75
th
percentile. Given that cities differ in size, the results indicate that a mid-sized city with 50,000
residents will increase total expenditure by an average of $850,000 and a larger city of 400,000
residents will increase its total expenditure on public services by $6.8 million. The converse will
occur in the case of negative asset returns. Total expenditure on public services will decrease by
a similar magnitude for each 1 percent loss in asset returns. These results are confirmed from
robustness tests presented in Column (2). Although the effect size for spillovers narrows, the
effect remains significant at the 0.1% level.
In column (3), I allow spillovers to depend on whether or not the pension plan is
underfunded. The negative coefficient for underfunded pensions confirms hypothesis H1(a)
which proposed that spillovers from underfunded pensions will cause a decrease in total
expenditure. Pension underfunding has significant negative spillovers, the effect is significant
31
at the 0.1% level. The coefficient of -0.022 means that for each 1 percent increase in unfunded
pension liabilities, total expenditure decreases by 2.2 percentage points. In terms of budgetary
tradeoffs, each additional dollar allocated to pay down an underfunded pension plan will
decrease allocation to public services by 0.17 cents. This translates an $35 less per capita at the
mean and $42 at the 75
th
percentile of the service expenditure distribution.
Table 2. Spillovers to Public Service Expenditure
Shock
Model
Robustness Shock +
Unfunded
Full Model
Pension Performance
Asset Shock
0.011
***
(0.001)
0.0074
***
(0.0008)
0.021
***
(0.002)
0.011
***
(0.002)
Unfunded Liability -0.022
***
(0.002)
-0.013
***
(0.002)
Asset Shock ×
Unfunded Liability
0.0028
***
(0.0002)
0.0019
***
(0.0001)
Institutional Factors
Home Rule
0.034
***
(0.006)
0.028
***
(0.006)
Control Variables
Population -0.011
***
(0.003)
-0.012
***
(0.003)
Percent registered Democrat 0.13
***
(0.03)
0.11
***
(0.02)
Percent Hispanic -0.11
***
(0.02)
-0.094
***
(0.02)
Percent African American -0.13
*
(0.05)
-0.12
*
(0.05)
Unemployment Rate -0.09
(0.09)
-0.05
(0.09)
Median Household Income 0.024
*
(0.01)
0.028
**
(0.01)
Housing Tenure -0.12
***
(0.03)
-0.11
***
(0.03)
Constant 1.92
***
(0.006)
1.81
***
(0.1)
1.87
***
(0.009)
1.76
***
(0.1)
N 22,424 22,424 17,180 17,180
R2 0.15 0.35 0.25 0.41
Note: Coefficients are in log points for [t+1]. Population and median household income are log transformed. All models include
fiscal year and pension reform fixed effects. Standard errors are clustered at the plan sponsor level.
***
p<0.001,
**
p<0.01,
*
p<0.05.
32
I also find that the spillover effect is partially mitigated by asset returns to pension fund
investments. The coefficient of interest, 𝜃, is significant at the 0.1% level. This means that each
1 percent increase in asset returns mitigates spillovers from underfunded pensions to service
expenditures by 0.28 percentage points or 0.02 cents. Column (4) presents the full model which
controls for a plan sponsors institutional, socio-economic, and demographic characteristics. The
effect size for the variables of interest – asset shocks, underfunded pensions, and their interaction
- are marginally smaller, but significant at the 0.1% level. The model also has a high R2 and
explains 40% of the total variation in service expenditure.
In the next analysis, I investigate differential spillovers by three levels of pension stress. A
pension plan is stressed when plan liabilities exceed its assets. The ratio of plan assets to liabilities
is given by a pension plan’s funded ratio. Differential spillovers are assessed for plans with severe
stress (that have a funded ratio between 50 to 65%), plans with high pension stress (these plans
have a funded ratio between 65 to 80%), and those with low pensions stress (with an 80% or higher
funded ratio). Results from analyses that use stress-level subsample are presented in table 3.
Table 3. Spillovers by Stress Level
Severe Stress:
50-65% Funded
High Stress:
65-80% Funded
Low Stress:
80% or Higher Funded
Pension Performance (1) (2) (3)
Asset Shock -0.01
(0.009)
0.016
**
(0.005)
0.007
**
(0.002)
Unfunded Liability 0.015
(0.009)
-0.022
***
(0.005)
-0.0061
**
(0.002)
Asset Shock ×
Unfunded Liability
0.0026
***
(0.0007)
0.0031
***
(0.0003)
0.0014
***
(0.0002)
Note: This table presents coefficients of interest, extended models are presented in the appendix. Coefficients are in log
points for [t+1]. All models include fiscal year and pension reform fixed effects. Standard errors are clustered at the plan
sponsor level.
***
p<0.001,
**
p<0.01,
*
p<0.05.
33
The results are largely similar to those presented in table 2 but differ in a few important
ways. I find that asset returns have a positive effect on service expenditures but this only occurs for
plans that are 65% or higher funded. Higher asset returns will not transmit to increased spending
on public services when a plan sponsor has one or more severely stressed pension plan. The main
effect of underfudned pensions is negative and significant for plans with high and low stress. In
addition, column (2) shows highest spillovers to service expenditures are observed for plan sponsors
with one or more plans with a 65 to 80 percent funded status. The effect size shows that a one
percent increase in unfunded pensions causes a 2.2 percentage point decrease in total expenditure
for plans in the high stress group and by 0.61 percentage points for plans in the low stress group.
These spillovers are partially mitigated by asset returns. As expected, the mitigating effect is larger
for severe and highly stressed plans compared to plans with low pension stress. The mitigating
effect is between 0.26 and 0.31 percentage points for plans that are 50 to 80 percent funded and
0.14 percentage points for those with 80% or higher funded status.
I now analyze spillovers from underfunded pensions to the three main service functions of
local governments: public safety, economic development, and redistribution. These results which
Table 4. Differential Spillovers by Service Type
Public
Safety
Economic
Development
Redistribution
Pension Performance (1) (2) (3) (4) (5) (6)
Asset Shock 0.015
***
(0.003)
0.013
***
(0.002)
0.015
***
(0.004)
0.014
***
(0.003)
0.011
*
(0.004)
0.0043
(0.004)
Unfunded Liability -0.002
(0.002)
-0.016
***
(0.002)
0.001
(0.003)
-0.016
***
(0.003)
0.0037
(0.003)
-0.0043
(0.003)
Asset Shock ×
Unfunded Liability
0.0024
***
(0.0002)
0.0022
***
(0.0003)
0.0019
***
(0.0003)
Note: This table presents the main coefficients of interest, extended models that include control variables are presented in the
appendix. Coefficients are in log points for [t+1]. Empirical models include fiscal year and pension reform fixed effects. Standard
errors are clustered at the plan sponsor level.
***
p<0.001,
**
p<0.01,
*
p<0.05.
34
are shown in table 4 indicate support for hypothesis H1(b), pension underfunding has significant
spillover effects on core services – public safety and economic development – as well as an impact
on non-core redistributive expenditure.
Columns (1) and (2) show results for spillovers to public safety. The findings suggest that
asset returns increase public safety expenditure while underfunded pensions decrease city spending
on public safety. Results in Column (2) shows that the spillover effect for public safety spending
is 1.6 percentage points and the is mitigated by 0.24 percentage points with a net spillover of 1.5
percentage points. Spillovers to spending on economic development are shown in columns (3)
and (4). The effect size is similar to public safety - 1.6 percentage points with a marginally lower
mitigating effect of 0.22 percentage points. Columns (5) and (6) show results for spillovers to
redistribution. The results indicate that positive effect of asset returns on redistributive spending is
sensitive to pension underfunding. More specifically, municipalities are less likely to spend on
redistribution when pensions are underfunded.
Table 5. Spillover Results for Subsamples of Over- and Under-Funded Plans
Public
Services
Public
Safety
Economic
Development
Redistribution
Over-
funded
Under-
funded
Over-
funded
Under-
funded
Over-
funded
Under-
funded
Over-
funded
Under-
funded
(1) (2) (3) (4) (5) (6) (7) (8)
Asset Shock
(AS)
0.016
***
(0.002)
0.013
***
(0.002)
0.017
***
(0.002)
0.014
***
(0.002)
0.016
***
(0.003)
0.016
***
(0.003)
0.019
***
(0.004)
0.0071
(0.004)
Unfunded
Liability
(UL)
-0.014
***
(0.002)
-0.016
***
(0.002)
-0.017
***
(0.003)
-0.0059
(0.004)
AS × UL
0.0018
***
(0.0002)
0.0022
***
(0.0002)
0.0021
***
(0.0003)
0.0017
***
(0.0003)
Note: This table presents the main coefficients of interest, extended models that include control variables are presented in
the appendix. Coefficients are in log points for [t+1]. The subsample of overfunded pensions consists of plans in which assets
> liabilities, the sample size 6600. The subsample of underfunded pensions consists of plans in which assets < liabilities, the
sample size is 15,300. All models include fiscal year and pension reform fixed effects. Standard errors are clustered at the
plan sponsor level.
***
p<0.001,
**
p<0.01,
*
p<0.05.
35
For the next analysis, I assess spillover effects of pension plans that are overfunded and
compare these results to spillovers from plans that are underfunded. Results are shown in table 5.
Columns (1) and (2) present results for total expenditure. Using observations from the
subsample of overfunded pensions, I find that a 1 percent increase in asset returns increases service
expenditure by 1.7 percentage points. The effect size decreases for the subsample of underfunded
pensions, asset returns increase service expenditures by 1.4 percentage points, a difference of 0.03
percent. This pattern of findings repeats for public safety as well (columns 3 and 4). The effect
size for asset returns drops by .03 percentage points. The effect size is unchanged in the case of
economic development (columns 5 and 6). There is a sharp decline in redistributive spending
(columns 7 and 8). The findings indicate that municipalities are more likely to provide
redistributive services when pensions are overfunded and less likely to do so when one or more
plans are underfunded. I also find that the mitigating effect of asset returns is highest for public
safety (0.22 percentage points) and lowest for redistribution (0.17 percentage points).
Overall, I find that spillovers from pension underfunding have a causal effect on decrease
in municipal expenditures on essential public services. In addition, because mitigating these
spillovers depends significantly on asset returns, return volatility is transmitted to instability in
public service provision. More importantly, spillovers to redistribution indicates that pension
underfunding has implications for social equity.
5. 2 Effect of Underfunded Pensions on Public Employees
Table 6 presents results for spillovers from underfunded pensions to public employees’
pension benefits. Column (1) shows the unconditional effect of asset returns on pension benefits.
The coefficient is positive and significant at the 0.1% level which means that asset returns increase
36
employee benefits. Column (2) presents results from robustness tests. The coefficient for asset
returns remains significant with a marginal decrease in effect size.
In column (3), I allow for the possibility that spillovers depend on whether pensions are
underfunded or overfunded. The findings confirm hypothesis H2(a) indicating employee
benefits decrease in response to pension underfunding. The coefficient of interest, 𝜃, shows that
asset returns mitigate spillovers, the mitigating effect is a 7.7 percentage points or nearly
$17,769 per employee. Column (4) shows the full model that controls for plan sponsors’
institutional, socio-economic, and demographic characteristics. The effect size decreases
marginally to 7 percentage points and is significant at the 0.1% level.
Figure 1 shows spillovers to employee benefits for plans with severe, high, and low pension
stress. As observed, the spillover effect is largely confined to severely stressed plans, there are
negligible spillovers to plans with high and low stress. In other words, employee benefits are more
likely to decrease when a plan sponsor has one or more severely stressed plans.
Table 6. Spillovers to Pension Generosity
Shock
Model
Robustness Shock +
Unfunded
Full
Model
(1)
(2)
(3)
(4)
Pension Performance
Asset Shock 0.47
***
(0.02)
0.41
***
(0.01)
-0.054
*
(0.02)
-0.009
(0.02)
Unfunded Pensions -0.88
***
(0.03)
-0.81
***
(0.03)
Asset Shock ×
Unfunded Pensions
0.077
***
(0.002)
0.069
***
(0.002)
Note: This table presents the main coefficients of interests, extended models that include control variables are
presented in the appendix. Coefficients are log points for [t+1]. All models include fiscal year and pension reform
fixed effects. Columns (2) and (4) include fixed effects and control variables. Standard errors are clustered at the
plan sponsor level.
***
p<0.001,
**
p<0.01,
*
p<0.05.
37
The next analysis presents results from equation (2) that examine the mediating effects of
unionization on employee benefits. Table 7 presents findings on interaction of unionization and
underfunded pensions for the three main employee groups at the local level – firefighters, police
officers, and general municipal employees. I find that while spillovers to employee benefits
diverge across the three groups, there is a negative effect on benefits levels for groups with high
bargaining power – firefighters and police officers – and a marginal increase in benefits for the
group with low bargaining power – general municipal employees. This pattern of findings is
contrary to political economy theory which stipulates that bargaining power will predict
spillovers to pension generosity and therefore fails to confirm hypothesis H2(b).
Columns (1) and (2) present results for the unconditional effect of unionization, the
findings show increase in benefits for groups with high bargaining power – firefighters and police
officers – while it declined for the low bargaining power group – general municipal employees.
Figure 1. Differential Spillovers by Pension Stress
10 11 12 13 14
predicted value
1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016 2017 2018
fiscal year
Severe Pension Stress:50-64% Funded Ratio
High Pension Stress:65-79% Funded Ratio
Low Pension Stress:80% or Higher Funded Ratio
38
In columns (3) and (4), I allow the effect of unionization to be conditional on underfunded
pensions. I find that when allowing for pension underfunding to the effect of unionization on
spillovers to employee benefits, there is decrease in benefits for employee groups with high
bargaining power – firefighter and police officers – and a marginal increase in benefits to the
group with low bargaining power – general employee unions. Spillovers to employee benefits are
highest for police unions – a 11 percentage point decrease in benefits – followed by firefighter
unions – 7.5 percentage point decrease in benefits. This translates to $32,501 less in pension
benefits per police officer and $23,338 less in pension benefits per firefighter employee. At the
same time, the benefit level for general municipal employees has increased by 6 percentage points
Table 7. Mediated Effect of Unionization on Pension Generosity
Shock
Model
Robustness Union
Mediation
Full
Model
(1) (2) (3) (4)
Pension Performance
Asset Shock -0.044
*
(0.02)
0.012
(0.02)
-0.046
*
(0.02)
0.0036
(0.02)
Unfunded Liabilities -0.66
***
(0.04)
-0.53
***
(0.03)
-0.75
***
(0.04)
-0.61
***
(0.03)
Asset Shock ×
Unfunded Pensions
0.062
***
(0.003)
0.049
***
(0.003)
0.068
***
(0.003)
0.055
***
(0.003)
Unionization
Firefighter Union 0.19
***
(0.03)
0.16
***
(0.03)
0.89
***
(0.17)
0.82
***
(0.15)
Police Union 0.27
***
(0.03)
0.24
***
(0.02)
1.27
***
(0.14)
1.22
***
(0.12)
General Employee Union -0.29
***
(0.02)
-0.35
***
(0.02)
-0.83
***
(0.09)
-0.66
***
(0.08)
Unfunded Liabilities × Union
Firefighter Union -0.075
***
(0.02)
-0.068
***
(0.01)
Police Union -0.11
***
(0.01)
-0.12
***
(0.01)
General Employee Union 0.062
***
(0.009)
0.038
***
(0.008)
Note: This table presents the main coefficients of interest. The extended model which includes control variables is
presented in the appendix. Coefficients are log points for [t+1]. All models include fiscal year and pension reform fixed
effects. Columns (2) and (4) include fixed effects and covariates. Standard errors are clustered at the plan sponsor level.
***
p<0.001,
**
p<0.01,
*
p<0.05.
39
or $5,092 per employee. These results are largely explained by the higher level of pension stress
for firefighter and police pensions as opposed to the greater bargaining power of these unions. In
other words, despite unionization, public employees are experiencing the adverse impacts of the
stress induced by pension underfunding.
VI Robustness
Table 8 presents results from robustness tests for spillovers from underfunded pensions.
The forward-looking estimates for both total expenditure and pension generosity are highly robust
and significant for up to four years after the initial unfunded liability is incurred.
Table 8. Robustness Tests
Total Expenditure Pension Generosity
[t+2] [t+3] [t+4] [t+2] [t+3] [t+4]
Asset Shock 0.012
***
(0.002)
0.012
***
(0.001)
0.011
***
(0.002)
-0.05
**
(0.02)
-0.11
***
(0.01)
-0.15
***
(0.02)
Unfunded Liabilities -0.015
***
(0.002)
-0.014
***
(0.002)
-0.013
***
(0.002)
-0.82
***
(0.02)
-0.83
***
(0.03)
-0.84
***
(0.03)
Asset Shock ×
Unfunded Liabilities
0.0022
***
(0.0002)
0.0023
***
(0.0002)
0.0025
***
(0.0002)
0.071
***
(0.002)
0.074
***
(0.002)
0.076
***
(0.002)
Note: This table presents the main coefficients of interest, extended model that include all control variables are presented in the appendix.
Coefficients are in log points. Asset shocks and unfunded pensions are per capita for public services and per member for pension benefits,
inflation adjusted to 2018 constant dollars and log transformed. All models include fiscal year and pension reform fixed effects. Standard
errors are clustered at the plan sponsor level.
***
p<0.001,
**
p<0.01,
*
p<0.05.
In considering results from future periods, I find a marginal decay in spillovers to service
provision while spillovers to pension generosity amplifies over future periods. Asset returns have
a consistent positive effect on expenditure and benefits. In addition, the relative salience of asset
returns is higher for employee benefits compared to service provision
40
VII Discussion and Conclusion
The underfunding of public pension systems is perhaps the biggest fiscal challenge that
state and local governments are facing. With total underfunded pension liabilities inching closer
to $5 trillion, a fifth of the total US GDP, pension stress aggravates the pressures on local budgets.
While previous research has considered the factors that led up to the rise of pension underfunding,
an unexamined question relates to the downstream effects of unfunded pension liabilities.
In this article, I provide the first evidence on how pension stress is affecting local
governments. Using two decades (1997-2018) of new data on 1,842 city-sponsored pension plans
and a novel identification strategy, I provide clean, causal estimates on spillovers from
underfunded pensions to total expenditure on public services, the impact on core and non-core
municipal services, and generosity of public employees’ pension benefits. I estimate dynamic fixed
models to identify both short- as well as long-term spillovers.
I find that underfunded pensions have significant spillovers to service provision, for every
$1 allocated toward paying down a plan’s underfunded liability, the plan sponsor will reduce
allocation to public services by 0.15 to 0.17 cents. The spillover effect extends to both core and
non-core services. The average spillover to public safety allocation is 0.6 cents. Cities are also
more likely to provide redistributive services when pensions are overfunded relative to when plan
sponsors incur unfunded liabilities. This result indicates that pension underfunding has distributive
implications. I also find that spillovers are while spillovers are pervasive across varying levels of
pension stress, there are higher spillovers to service provision for plans with a funded ratio of 65
to 80 percent, and spillovers to employee benefits are observed for severely stressed pensions plans
with a funded ratio that is 50% or lower.
41
The rise of underfunded pensions has brought the role of public sector unions into question.
From a political economy perspective, contrary to the conjecture that unions insulate public
employees from the potential adverse impact of fiscal pressures (Anzia & Moe, 2015; Kelley,
2014; Dippel & Sauers, 2019; DiSalvo & Kucik, 2018), I find that pension stress explains
spillovers to unions with high bargaining power.
This paper is a first step toward unpacking the effects of pension underfunding on cities,
and citizens. There is need for more research to examine how public pension liabilities are affecting
the ability of local governments to address some of the key societal challenges in the US including
homelessness, racial wealth gaps, and racialized policing. Further research is also needed on fiscal
responses of other local governmental entities with substantive unfunded pension obligations
including counties, school districts, and special districts. From fiscal federalism perspective, future
research can examine the relationship between fiscal disparities, intergovernmental transfers, and
pension underfunding, particularly to understand the intersection between concentrated poverty
and the provision of local redistributive services. Recent shifts to high-risk pension investment
portfolios and risky debt instruments increase both volatility and complexity of pension financing
which can have potential implications on access to essential public services that low-income
households are dependent on such as public transit and housing. The findings in this paper
emphasize a strategic balance between service provision and benefits in order to offset budgetary
tradeoffs that result from underfunding pensions. A balanced approach is vital to mitigating the
distributive consequences of pension stress.
42
Chapter 3:
Performance and Equity Implications of Fiscal Adaptation to Negative
Revenue Shocks
43
Performance and Equity Implications of Fiscal Adaptations to
Negative Revenue Shocks
Manita Rao
Ph.D. Candidate, Public Policy, and Management
Sol Price School of Public Policy
University of Southern California
Abstract
This article critically examines adaptations to unanticipated revenue disruptions and the
extent to which fiscal responses hinder social equity. The study employs a quasi-experimental
context in which some localities experience a negative revenue shock that results from voters’
decisions on local tax and bond referendums. I investigate three distinct fiscal adaptations to the
abrupt revenue disruption. Further, I examine how institutional factors mediate fiscal adaptation
and impact reliance on non-tax regressive revenue sources such as fines, fees, and forfeitures.
Using over three decades of data (1991-2018) and over 2000 distinct fiscal referendums, I estimate
the dynamic regression discontinuity effect on fiscal adaptations. The results reveal that although
the direct effect is adaptation through lower service provision, when mediated co-occurring
stressors such as political pressure and unfunded pension liabilities, local governments
demonstrate fiscal maladaptation. These findings inform policy discussions on financial
management of unanticipated disruptions, and social equity in local public finance.
Keywords: Fiscal adaptation, local direct democracy, municipal government, public services,
social equity
44
I Introduction
The disruption caused by Covid-19 was detrimental to fiscal stability of state and local
governments. With the abrupt onset of the pandemic, sub-national revenues plunged by nearly
40% as a result of a sudden drop in consumer activity (Chernick et al., 2020). While Covid-19 was
a one-time event, local governments routinely face similar unanticipated fiscal disruptions in the
institutional context of local direct democracy wherein voters’ decision on tax and debt
referendums cause abrupt revenue shocks. In this paper, I critically examine the ways in which
local public agencies adapt to negative revenue shocks and the extent to which adaptive
mechanisms affect social equity.
An enduring question in public administration pertains to whether adaptation helps or
hinders a public agency’s performance. The disruption caused by Covid-19 has renewed interest
in the concept of adaptive governance, defined broadly as the ability of public institutions to
respond to turbulence through creation of innovative, scalable solutions (Ansell et al., 2020). Some
scholars argue that adaptive organizations are better suited to manage disruptions with minimal
decline in the provision of essential public services (Warner et al., 2020; Hendrick & Degnan,
2020). An alternate perspective, however, emphasizes the destabilizing characteristics of
adaptation arguing that internal instability hinders performance and the unintended consequences
of adaptation are its disproportionate adverse impact on minorities and low-income households
(Boyne & Meier, 2009; Boin et al., 2017). These polarized perspectives on adaptation reveal
tensions that have received limited theoretical and empirical attention in public management
literature.
In this study I argue the impact of fiscal adaptation on local government performance and
its social equity implications are contingent on adaptation mechanisms and institutional factors
45
that mediate adaptive outcomes. The theory is motivated by interdisciplinary perspectives from
literature in strategic management and governance adaptation. These perspectives emphasize the
distinctiveness of adaptation sby public organizations as opposed to private organizations. Unlike
firms where strategic adaptation is driven by factors such as survival and competitive advantage,
the rationale for adaptation by public agencies is internal stability and service continuity (Glover
& Granberg, 2021; Zhang & Maroulis, 2021; Hrebiniak & Joyce, 1985). In addition, because
public agencies are, by design, embedded in political contexts, both internal and external
governance can be expected to impact fiscal adaptations. Therefore, I theorize that fiscal
adaptations are demonstrated in response to the direct impact of negative revenue shocks in
addition the mediated impact of governance institutions.
I test the theory in a quasi-experimental context in which local governments experience
unanticipated revenue disruptions in response to the outcome of local tax and debt referendums.
Employing over 2,600 distinct fiscal referendums proposed by municipal governments in
California, I critically examine three distinct adaptive mechanisms – adaptation through service
expenditures, adaptation through non-tax revenues, and adaptation through fiscal solvency. Using
a dynamic regression discontinuity design, I estimate the Average Treatment Effect (ATE) of
negative revenue shocks on local fiscal outcomes. The empirical strategy compares the ATE of
negative revenue shocks to effects mediated by the referendum sponsors’ institutional
characteristics. More specifically, I analyze the mediated effect of three distinct characteristics of
local government administration: professional management, political pressure, and pension stress.
Multiple data sources are used for analysis. The primary data on local fiscal referenda are
publicly available from the California State Department and the California Elections Data Archive
46
(CEDA).
7
I combine the fiscal referenda data with detailed financial information on the 480
proposer city governments. These data include fine grained information on revenue, expenditure,
and debt. Data on the institutional factors was extracted from collective bargaining contracts that
were obtained by placing a public records request with all 480 municipalities in California.
Professional management is an indicator variable for whether a city employs a professionally
trained manager. Political pressure, similarly, is an indicator variable, proxied by whether
municipal firefighter, police, and management employees are unionized. Information on the third
institutional variable – pension stress – is proxied by the Unfunded Actuarial Accrued Liabilities
(UAAL) incurred by the sponsoring local government; these data were obtained under a publoc
records request from the California Public Employees’ Pension Retirement System (CalPERS)
8
,
the largest pension fund in the US and the administrative agency for city-sponsored pensions in
California.
The main findings are that the direct causal effect of negative revenue shocks is associated
with decline in service provision but no statistically significant effect on non-tax revenues and
fiscal solvency. More specifically, the ATE of negative revenue shocks is a 4.5 percentage point
decrease in total service expenditures and a 4.4 percentage point decrease in spending on public
safety. These effects are highly persistent and, in the case of the effect on public safety, remain
significant for up to five years after the initial shock. However, the mediated effects of institutional
factors diverge significantly from the direct causal effect.
7
The California Elections Data Archive (CEDA) is maintained by the Sacramento State University and provides
information on all candidate and referenda elections in the state of California.
8
CalPERS is the largest public pension fund in the US and the fourth largest in the World. It has over $444 billion
in assets under management. The majority of pension plans sponsored by municipal, county, and special districts are
contracted with CalPERS. This novel data source data source has never been used before and allows this study to
identify the impact of public pension liabilities on local government fiscal behavior when institutional turbulence
occurs and more importantly when multiple fiscal stressors cooccur.
47
I find that while the mediated effect of professional management is associated with short-
term adaptation through decrease in service expenditures, no significant increase in non-tax
revenues, and a marginal increase in fiscal solvency. Unlike professional management, the
mediating impact of political pressure and pension stress demonstrate adaptation through higher
service expenditure coupled with increase in non-tax regressive revenues from fines, fees, and
forfeitures, and a significant decline in fiscal solvency. This pattern of findings indicates that
unlike the direct impact of negative revenue shocks, the mediated effect of institutional factors
such as political pressure and pension stress that cause increase in reliance on non-tax regressive
revenues have adverse impact on social equity and demonstrate fiscal maladaptation.
The causal evidence presented in this paper informs ongoing debates on the promise and
pitfalls of adaptation by public agencies. I disentangle the direct causal effect of negative revenue
shocks from the mediated effect of institutional factors to identify the source of maladaptation
particularly as it relates to social equity. The remainder of this paper contains a theoretical
discussion on strategic adaptation, a review of empirical evidence on local fiscal management of
revenue pressures, the empirical strategy, results, and conclusion.
II Theoretical Perspective on Strategic Adaptation
The development of literature on strategic adaptation can be traced back to the mid 1960s,
theoretical advancement during this period on open systems and resource dependence theory were
instrumental to the development of contrary schools of thought on environmental determinism,
strategic fit, and strategic adaptation. The environmental determinism perspective posited that
strategic organizational behavior was a function of resource dependent relations with competitors
as well as the regulatory environment (Lawrence & Lorsch, 1967; Thompson, 1967). From an
48
open systems perspective, a public agency’s resource dependence on political decision makers
such as politicians, and voters are critical to agency strategy (Bozeman & Slusher, 1979; Bozeman,
1981). For instance, changes in political leadership or revenue disruptions induce strategic
adaptation to the extent that resources have to realign with a new budget constraint or a new
political agenda (Kuipers et al., 2014). Such strategic adaptations will affect performance in
agencies that lack internal financial or personnel. For example, O'Toole and Meier (2010) find that
negative revenue shocks were less likely to affect student performance in school districts with
financial reserves to buffer against decline in instructional spending.
Strategic adaptation can also occur ex-post or after a disruption occurs. Ex-post adaptation
is limited by agents’ bounded rationality and involves proximal search for solutions (Simon, 1955;
Nelson, 2012). Using data on school fiscal referenda from Ohio, Kogan et al. (2017) find that
school districts where referenda failure resulted in negative revenue shocks were more likely to
reduce instructional spending that decreased student performance. Because bounded rational
adaptations are localized, the impact on performance outcomes depends on strategic alignment
with internal actors such as employees, and managerial perception of fiscal vulnerability (Kim et
al., 2018; Moynihan, 2012; Bjørnholt et al., 2016; Zhang et al., 2018).
While an open systems perspective of strategic adaptation pertains to internal decision
making, a governance perspective shifts the focus to a public agency’s political context. For
example, Klein et al. (2019) argue that internal political conflict that results from pursuing their
individual interests not on disenfranchises external stakeholders with limited voice but also affects
performance. Along a similar vein, Ansell and Trondal (2018) suggest that factional conflict, staff
turnover, conflicting organizational rules, or internal reforms increase the complexity of governing
disruptions. The performance of public organizations can also be affected by factors such as
49
electoral proximity and voters’ perception of government performing. For example, Dipoppa and
Grossman (2020) find that electoral proximity was associated with increase in agency response to
citizen complaints but this corresponded with a higher number of citizen complaints. Similarly,
Lavertu (2016) shows that in the context of direct democracy, voters were more likely to reject
ballot measures in schools with lower test scores which in turn affected school performance. This
indicates that the unique coupling of strategic and governance conditions can be expected to
influence not only how public agencies adapt to fiscal disruptions but also whether such
adaptations help or hinder social equity.
III Hypotheses
Conceptually, fiscal adaptation can occur through changes in service provision, changes to
how revenues are raised, or through a strategic balance between revenue and expenditure which
would be reflected in fiscal solvency. Adaptation through services can be associated with increase
or decrease in total expenditure, core services such as providing public safety services, or through
changes to peripheral services. Similarly, adaptation through revenue will involve a search for
alternate tax and non-tax revenue sources. When tax revenues are subject to a political approval
mechanism such as in the case of the current study, revenue adaptation will occur by accessing
non-tax revenues. Distinct from adaptation through expenditure and revenue, when both sides of
the fiscal ledger are strategically balanced, adaptation occurs through short- or long-term fiscal
solvency. Public organizations will, therefore, demonstrate fiscal adaptations to negative revenue
shocks through one or more of these approaches – service expenditures, non-tax revenues, or fiscal
solvency. Further, to examine the relationship between institutional characteristics and fiscal
adaptation, the paper develops and tests hypotheses on the mediating impact of professional
management, political pressure, and pension stress on adaptive responses.
50
3.1 Fiscal Adaptation and Professional Management
Cities in the US are governed either under a mayor-council system or a council-manager
system. In a mayor-council system, the administrative leader is an elected mayor while in a
council-manager system, a professionally trained manager is appointed by an elected council to
provide administrative leadership. Previous research shows that fiscal decisions of professional
managers differ from those of mayors and these differences are particularly critical in times of
fiscal strain. It is theorized that the politics-administrative dichotomy shields appointed managers
from electoral pressures thereby increasing the scope for adoption of prudent management
practices (Jimenez, 2020; Moynihan, 2012). Professional managers also differ from mayors in
their professional training which inculcates values of fiscal stewardship in addition to increasing
technical capabilities in public financial management (Hendrick, 2006; Schmidt & Groeneveld,
2021).
Differences in electoral incentives and professional training have been shown to affect the
financial management practices of professionally managed cities. For instance, cities managed by
professionally trained managers are shown to have higher fiscal reserves and adopt rainy day
policies, practices that have been shown to buffer against fiscal disruptions (Marlowe, 2005).
There is also evidence that professional managers demonstrate a preference for minimizing
disruptions to core organizational functions such as public safety or educational instruction (Wei,
2020; Boyne & Meier, 2009; Brien et al., 2021). In addition, studies on professionally managed
municipalities finds higher propensity for adoption of efficiency-enhancing practices and
deployment of a wider array of budgetary tools to counter revenue disruptions caused by economic
recessions. This not only reduces the need for cutbacks to service expenditure but also ensures
compliance with statutory balance budget mandates (Guo & Neshkova, 2018). It is estimated that
51
these differences between professionally managed vs. mayoral cities reduce fiscal wastage by
nearly $36 million (Jimenez, 2020). Taken together, existing evidence suggests that professionally
managed cities will not only adapt differently from mayoral cities but will demonstrate prudent
financial management. The hypothesis is stated as follows:
H1: Professionally managed cities will demonstrate fiscal adaptation through increase in service
provision, no change in non-tax revenues, and improved fiscal solvency.
3.2 Fiscal Adaptation and Political Pressure
An extensive literature in cutback management shows that fiscal scarcity creates the
conditions for internal political conflict and such conflict is particularly pervasive when internal
groups are organized (Levine et al., 1981; Bozeman, 2010). Because negative revenue shocks
increase fiscal strain and involve concessions, internal groups with stronger coalitions are better
positioned to resist cutbacks to their preferred services which will result in shifting reductions to
social programs with fewer champions. Such tradeoffs that emerge as a result of political pressure
from internal groups influence the extent to which fiscal adaptations occur strategies that result in
across-the-board cuts to all services as opposed to strategies targeted toward reductions cuts to
specific service categories (Cepiku et al., 2016; Jimenez, 2021). Although across-the-board
strategies maintain internal equity, such an approach can also increase inefficiency to the extent
that cuts efficient programs are equal to inefficient programs.
Some studies show that despite the potential to internal political conflict, fiscal decisions
under revenue strain are associated with management heavy handedness to maintain a balanced
budget, reductions to expenditure categories over which managers exercise discretion, and the
influence of political pressure from employee unions is largely relegated to staffing as opposed to
service decisions (West & Davis, 1988; Downs & Rocke, 1984; Berne & Stiefel, 1993; Bartle,
52
1996). Other studies, however, show that intense political pressure is offset by accessing alternate
sources of revenue to reduce the need for cutbacks to either services, staffing, or wages. In
particular, local governments employ tools such as issuing short-term bonds, competing for
intergovernmental grants, and increasing fees for services in a bid to offset the revenue loss
(Greenhalgh & McKersie, 1980; Jimenez, 2014; Warner et al., 2020). Recent studies also links
political pressure from employee unions to public financial practice that reduce fiscal transparency
such as obscuring funds by saving resources in less visible fund accounts thereby reducing the
scope for political conflict (Gore, 2015; Klasa et al., 2009). Because fiscal strain increases internal
political instability, fiscal adaptation is hypothesized to occur through both revenue and
expenditure channels.
H2: Cities with internal political pressure will demonstrate fiscal adaptation through increase in
service provision, higher non-tax revenues, and decline in fiscal solvency.
3.3 Fiscal Adaptation and Pension Stress
One of the biggest fiscal challenges facing state and local governments is the rise of
unfunded public pension liabilities. Although, public pensions are vital to public sector human
capital, unfunded liabilities are an institutionalized form of co-occurring fiscal stressor when cities
experience abrupt negative revenue shocks. More importantly, the institutionalized features of
public pensions such as the constitutional mandates which limit employer flexibility not only
increase fiscal stress but can also be expected to induce fiscal adaptation. Over the past two
decades, unfunded liabilities have mushroomed to $4.35 trillion
9
(Novy‐Marx & Rauh, 2011) and
9
The estimated amount of outstanding public pension liabilities varies from 1.35 trillion to 4.35 trillion depending
on the discount rate chosen.
53
have contributed to the bankruptcy of cities like Detroit, Stockton, Vallejo, and San Bernardino
(Abott & Singla, 2021).
Previous research shows that unfunded liabilities are increase during periods of fiscal strain
such as an economic recession (Chaney et al., 2002; Thom & Randazzo, 2015). When revenue
shortfalls occur over a prolonged period or are magnified by co-occurring stressors, local
governments are more likely to counter fiscal pressures by politically unpopular decisions
including increasing taxes and reducing the number of services that are provided free-of-cost
(Jimenez, 2021; Levine, 1978; Maher & Deller, 2007).
Recent research also shows that fiscally stressed local governments engage in financial
practices that harm social equity such as increasing revenue collections from regressive revenue
sources such as fines, fees, and forfeitures. Both cities and counties with declining revenues issued
more tickets for vehicle code violations and traffic citations. According to some estimates such
practice increase local revenues by as much as 8% (Su, 2020; Makowsky et al., 2019). Because
fines, fees, and forfeitures are disproportionately borne by racial minorities, such fiscal practices
are not only detrimental to social equity but are also shown to reduce public sector performance
(Pacewicz & N Robinson, 2020). For example, Goldstein et al. (2020) present causal evidence that
municipalities with high fiscal stress have lower crime clearance rates and this results from
reallocating police personnel from core crime fighting to enforcement activities. Based on the
existing evidence, to the extent that pension stress increases fiscal stress induced by negative
revenue shocks, the mechanisms through which agencies demonstrate fiscal adaptation can be
expected to affect social equity. The following hypothesis is tested:
H3: Cities with pension stress will demonstrate fiscal adaptation through decrease in service
provision, higher non-tax revenues, and decline in fiscal solvency.
54
IV Data, Empirical Design, and Diagnostics
The study uses a quasi-experimental design and estimates dynamic regression discontinuity
models to identify the direct causal effect of negative revenue shocks and the mediated effected of
institutional conditions on fiscal adaptation through service expenditures, non-tax revenues, and
fiscal solvency.
4.1 Data
The primary data source that the study uses is the universe of fiscal referendums on local
taxes and debt that were proposed by municipal governments in California over the two-decade
period from 1995 and 2018 by California’s 480 cities. The sample consists of 2,621 distinct
referendums proposed to raise revenues from sales and transient occupancy taxes, taxes on special
assessments, taxes on business licenses, and issue general obligation bonds. In a given jurisdiction,
voters decide on each fiscal referendum with an outcome that either approves or rejects the tax
measure. When a fiscal referendum is rejected by voters, it causes the proposer city government
to experience a negative revenue shock. For the majority of fiscal referendums in the sample,
approval requires a vote in favor by majority of voters in the proposer jurisdiction that cast their
ballot in the election in which the referendum is proposed. Some fiscal referendums such as special
assessment taxes and general obligation bonds are approved at a higher threshold that requires
two-thirds of the voters in the jurisdiction to vote in favor. The empirical strategy compares
potential outcomes for proposer governments in which a fiscal referendum was rejected with those
of in which a referendum was approved. The fiscal referendums data are publicly available from
the website of the California Secretary of State.
55
The analysis is conducted by constructing a referendum-level panel dataset that integrates
fiscal information on the outcome of the ballot measure with fine grained fiscal information on the
proposer municipality. The fiscal information is publicly available from the California State
Controllers’ office. The study has three outcomes of interest – fiscal adaptation through change in
service expenditures, fiscal adaptation through increase in non-tax revenues, and fiscal adaptation
through change in fiscal solvency. Fiscal data of the sponsoring local government are used to
estimate adaptation through expenditure and revenue. In the case of expenditure, I analyze the
impact on both total expenditure and spending on public safety. Adaptation through non-tax
revenue is estimated using data on service fees as well as fines and forfeitures. Following previous
research, fiscal solvency is measured over the short and long term. Short-term fiscal solvency is
given by share of unrestricted general fund revenue to total expenditure, and long-term solvency
is given by share of general fund assets net liabilities to total expenditure (Jimenez, 2020;
McDonald III, 2018).
I use two additional data sources to analyze the impact of institutional conditions –
professional management, political pressure, and pension stress - on fiscal adaptation. I construct
indicators variables for both professional management and political pressure using collective
bargaining contracts that were procured under a public records request placed with all 480 proposer
city governments. The indicator for professional management takes the value 1 if the city
employed a professionally trained manager in the year in which the ballot measure was proposed,
else the value is 0. The indicator variable for political pressure takes the value 1 if the city had a
firefighter, police officer and management union in the year in which the ballot measure was
proposed, else the value is 0. Pension stress is measured as a continuous variable. It is the per
member Unfunded Actuarial Accrued Liabilities (UAAL) incurred by a proposer city government
56
in the year that the ballot measure was proposed. These data were procured through a public
records request from the California Public Employee Retirement System (CalPERS).
The referendum-level dataset was constructed by identifying the focal election year in
which the ballot measure was proposed. For each focal election year, individual year datasets were
constructed by combining data on referendum outcome with the proposer governments’ fiscal
information. The multiple election years datasets were then stacked to construct the referendum-
level panel.
10
To execute the empirical strategy which is estimating the dynamic treatment effect
of negative revenue shocks, the year of election outcome constitutes the treatment year, years the
precede the election year are pre-treatment periods and years following the election year are the
post-treatment period. The referendum-level panel used for analysis includes four pre-treatment
periods and five-post-treatment periods [t-4, t+5] constructed around the focal election year. The
sample size for the dynamic panel ranges from 10,000 to 23,000 observations.
4.2 Empirical Design
This study uses a quasi-experimental design to analyze fiscal adaptations to negative
revenue shocks. The identification strategy exploits exogenous shocks to city revenues that result
from voters’ decisions on local fiscal referenda. The empirical strategy estimates the dynamic
effect of the revenue shock on fiscal adaptations. Using ballot measure vote share as the running
variable, I estimate Dynamic Regression Discontinuity (DRD) models
11
(Cellini et al., 2010). The
10
The approach used here follows closely from Cellini et. al. (2010) and Kogan et. al. (2017).
11
DRD designs are unique in allowing the estimation of treatment effects over time. The challenge of dynamic models
arises from potential shifts in the treatment and control group over time. For instance, a city that is treated by a negative
revenue shock will have the treatment reversed when voters approve the same ballot measure in a future period.
Therefore, there is a distinction between the Intent to Treat (ITT) and the Treatment on Treated (TOT). The
recommended correction is to reweight the coefficient by the probability of the treatment switch. The estimates
presented in the paper are ITT. Treatment switch corrected estimates are substantively similar to the ITT estimates
and are available on request.
57
running variable is zero-centered running variable for ease of interpretation. I estimate two
models, the causal effect of negative revenue shocks and the mediated effect of institutional
variables: professional management, political pressure, and pension stress.
Model 1: DRD model for causal effect of negative revenue shocks
𝐹𝑖𝑠𝑐𝑎𝑙 𝐴𝑑𝑎𝑝𝑡𝑎𝑡𝑖𝑜𝑛
!,#,+
= 𝑇
!
+𝜃
+
𝑇
!
𝜏
+
+𝑃
,
(𝜈
!
,𝜆
+
)+𝛿
#
+⍴
-
+𝜀
!,#,+
…(1)
where 𝐹𝑖𝑠𝑐𝑎𝑙 𝐴𝑑𝑎𝑝𝑡𝑎𝑡𝑖𝑜𝑛 is the outcome of interest; service expenditure, non-tax
revenue, and fiscal solvency for referendum 𝑖, in fiscal year 𝑡 and relative year 𝑘. 𝑇
!
is the
treatment variable – negative revenue shock. It takes the value 1 for failed fiscal referenda and 0
otherwise. 𝜏
+
is the year relative to the focal election year which estimates the dynamic effect of a
single treatment over five post-treatment periods. The Average Treatment Effect (ATE) is given
by 𝜃
+
. 𝑃
,
(𝜈
!
,𝜆
+
) is a flexible polynomial for the running variable – ballot measure vote share (𝜈
!
).
The model also includes fiscal year fixed effects (𝛿
#
) and election year fixed effects (⍴
-
). 𝜀
!,#,+
is
the error term.
Model 2: DRD model for mediated effect of institutions
𝐹𝑖𝑠𝑐𝑎𝑙 𝐴𝑑𝑎𝑝𝑡𝑎𝑡𝑖𝑜𝑛
!,#,.,+
= 𝑇
!
+𝜃
+
𝑇
!
𝜏
+
+𝛾
.
+
(𝑇
!
𝐷
.
𝜏
+
)+𝑃
,
(𝜈
!
,𝜆
+
)+𝛿
#
+⍴
-
+𝜂
!,#,.,+
…(2)
Equation (2) introduce a new term for the mediated effect of institutional variables. 𝐷
.
takes the value of each institutional variable – professional management, political pressure, and
pension stress. The coefficient of interest is 𝛾
.
+
which is estimates the mediated effects of
institutional variables on fiscal adaptations. 𝜂
!,#,.,+
is the error term. All other terms are similar to
equation (1).
58
4.3 Diagnostics
The treatment and control groups are examined to verify for non-manipulation and pre-
treatment stability, results from these tests are presented in figure 2 and table 9. Figure 2 shows
the histogram the zero-centered running variable which is the referendum vote share. A observed
the density is continuous around the zero-threshold which satisfies the non-manipulation
condition.
Figure 2. Distribution of Referenda Vote Share
0 .5 1 1.5 2 2.5
Density
-.6 -.4 -.2 0 .2 .4
margin of vote share
59
Table 9 presents results from pre-treatment covariate stability tests for the main outcome
variables. The first column shows the average and standard deviation for the treatment group which
are referendums that did not receive voter approval. The second column shows the mean and
standard deviation for the control group which are fiscal referendums that received approval. The
third column shows the difference in means between the treatment and control group and column
4 presents p-values which tests statistical validity of the difference in means. Because statistical
significance is observed for only one variable – expenditure on public safety, the results indicate
high pre-treatment stability between the treatment and control group.
Table 10 presents summary statistics for fiscal outcomes and the mediating variables. The
average per capita expenditure on public services is $1,826 and it is $515 on public safety. Cities
that proposed one or more fiscal referenda had an average of $236 per capita from service fees and
Table 9. Group Stability Test
Failed
Referenda
[Mean (SD)]
Approved
Referenda
[Mean (SD)]
Difference
[Failed -
Approved]
t-stat
(p-value)
(1) (2) (3) (4)
Public Services 5.09 (325.2) -0.9 (559.5) -5.99 -0.27 (0.7)
Public Safety 7.04 (73.6) -1.07 (84.4) 8.12 -2.1 (0.03)
Service Fees -1.75 (132.6) 6.29 (230.2) -8.04 0.91 (0.4)
Enforcement Fines 0.67 (11.1) 0.39 (7.3) 0.27 -0.58 (0.5)
Short-Term Solvency -0.006 (0.1) 0.0006 (0.1) -0.006 1.06 (0.3)
Long-Term Solvency 0.005 (0.1) -0.006 (0.1) -0.0004 -0.69 (0.9)
Pension Stress 72.7 (222.4) 92.2 (239.13) -19.5 1.05 (0.3)
Note: Covariate stability tests are estimated on a robust, bias-corrected, optimal bandwidth identified using
the rdbandwidth selection algorithm. Group means are change in treatment to year prior [t-1 to t]. Short-term
solvency is the general fund unrestricted balance as a share of expenditure. Long-term solvency is the general
fund assets net of liabilities as a share of expenditure. Pension stress is unfunded liability normalized by plan
membership.
60
raised an average of $15 per capita from enforcement fines. Short-term solvency measured as
unrestricted general fund reserves were 16% of total expenditure and long-term solvency measured
as assets net of liabilities were 34% of total expenditure. 43% of municipalities that proposed a
fiscal referendum employed professional managers, 25% had unionized firefighter, police officer,
and manager employees, and the average level of unfunded pension liability is $1,022 per member.
V Results and Analysis
This section presents results in three sub-sections. Sub-section 5.1 presents results on the
causal and mediated effect of negative revenue shocks on fiscal adaptation through service
Table 10. Summary Statistics
Mean SD Min Max Definition
Service Adaptation
Public Services 1,826 1,244 10 19,349 Expenditure on public services.
Public Safety 515 305 0 3,309 Expenditure on public safety.
Sourcing Adaptation
Service Fees 236 421 0 4,231 Per capita revenue from service
charges and fees.
Enforcement Fines 15 25 0 386 Per capital revenue from traffic citation
and fines from vehicle code violations.
Solvency Adaptation
Short-Term 0.16 0.26 -1.57 1.98 Ratio of unreserved general fund to
total expenditure.
Long-Term 0.34 0.38 -0.93 2.92 Ratio of general fund assets net of
liabilities to total expenditure.
Organizational Factors
Professional
Management
0.43 0 1 Dummy = 1 if city employs
professional manager.
Political Pressure
0.25 0 1 Dummy = 1 for if city has firefighter,
police, and management union.
Pension Stress 987 1,022 -1,085 8,798 Unfunded pension liability per member
Note: Revenue and expenditure variables are in per capita 2018 constant dollars. Negative values for
pension stress show overfunded pensions. Data for service fees cover the period from 2003-2016.
61
provision, sub-section 5.2 presents results on fiscal adaptation through non-tax revenue, and sub-
section 5.3 presents results on fiscal adaptation through short- and long-term fiscal solvency.
5.1 Fiscal Adaptation through Service Provision
Table 11 presents results for effects of negative revenue shocks on fiscal adaptation
through service expenditures. Column (1) shows the impact on total expenditure and column (2)
on public safety services. I find that negative revenue shocks are associated with decrease in
service levels and the ATE is 3.4 percentage points in relative year three and 4.5 percentage points
in relative year four. Column (2) presents estimates for public safety expenditure. I find that
negative revenue shocks decrease public safety spending, the ATE is 2.5 percentage points in
relative year two, 4.4 percentage points in relative year three and .3 percentage points three years
after the focal election year. Interestingly, the negative effect of core public safety spending
demonstrates high persistence, the effect is statistically significant up to five years after the initial
treatment. The effect size is highest for relative year three which suggests a two-year lag between
treatment and substantive drop in expenditure. On a per capita basis, total expenditure decreases
$80 and public safety spending by $22. For a mid-sized city with 100,000 residents, over a post-
treatment five-year period, the cumulative decline in expenditure is between 8 million for all public
services and 2.2 million for public safety.
62
Columns (3) and (4) show the mediated effect of professional management on fiscal
adaptation through change in service expenditures. The findings provide partial support hypothesis
1. Contrary to the hypothesized increase in service expenditure, I find that total expenditure
decreases by 6.7 percentage points in relative year two. The findings confirm the hypothesized
increase in public safety spending to the extent that spending levels remain largely unchanged
despite the negative revenue shock. These results indicate that although negative revenue shocks
affect the total level of public services provided, core public safety services are largely retained at
existing levels, and the revenue shock is managed through cuts to non-core services.
Table 11. Fiscal Adaptation Strategy: Effect on Expenditure
Mediating Effect of Institutional Factors
Negative Revenue
Shock
Professional
Management
Political
Pressure
Pension
Stress
Public
Services
Public
Safety
Public
Services
Public
Safety
Public
Services
Public
Safety
Public
Services
Public
Safety
(1) (2)
(3) (4)
(5) (6)
(7) (8)
1 year later -0.07
(0.01)
-0.01
(0.01)
-0.02
(0.02)
0.01
(0.01)
0.03
*
(0.02)
0.001
(0.01)
0.01
(0.01)
0.006
(0.004)
2 years later -0.03
*
(0.02)
-0.02
*
(0.01)
-0.06
**
(0.02)
0.03
(0.02)
0.06
*
(0.02)
0.02
(0.02)
0.003
(0.01)
0.0007
(0.006)
3 years later -0.04
**
(0.02)
-0.04
***
(0.01)
-0.02
(0.02)
0.01
(0.02)
0.04
*
(0.02)
0.02
(0.01)
0.01
(0.02)
0.005
(0.006)
4 years later -0.03
(0.02)
-0.03
*
(0.01)
0.03
(0.02)
0.02
(0.01)
0.03
(0.02)
0.04
*
(0.02)
0.005
(0.01)
0.01
(0.005)
5 years later -0.004
(0.02)
-0.03
*
(0.01)
0.01
(0.02)
0.01
(0.02)
0.03
(0.02)
0.07
***
(0.02)
0.02
(0.01)
0.007
(0.008)
Constant -0.02
*
(0.005)
-0.01
**
(0.004)
-0.012
(0.008)
-0.005
(0.008
)
-0.02
*
(0.006)
-0.02
***
(0.005)
-0.02
(0.02)
-0.02
(0.02)
N 23,587 23,605 23,500 23,520 23,587 23,605 10,734 10,758
R2 0.21 0.19 0.22 0.2 0.22 0.19 0.26 0.16
Note: Coefficients are in log points. Public services refer to total expenditure per capita and public safety refers
to per capita spending on fire, police, and Emergency Medical Services (EMS). Empirical models include fiscal
year and election year fixed effects. Standard errors are clustered at the city level. Pension stress is log
transformed.
***
p < 0.001
**
p < 0.01
*
p < 0.05
63
Columns (5) and (6) present results on the mediated effect political pressure. The findings
support hypothesis 2. Column (5) shows three continuous periods of growth in total expenditure
by 3.6 percentage points in the first post-treatment year, by 6.1 percentage points in the second
year, and by 4.7 percentage points in relative year three. Because the effect size is cumulative, the
biggest jump in spending occurs in the first year after the revenue shock. Between relative year
one and two, expenditure grows by 2.5 percent and between years two and three, spending drops
by 1.5 percent but is still significantly higher than the focal election year. The results in Column
(6) shows that expenditure on public safety remains largely unchanged in the first three years after
the negative revenue shock and increases by a statistically significant 4.4 percentage points in year
four. Columns (6) and (7) show that the mediating effect of pension stress on both total expenditure
and public safety spending is not statistically significant.
Taken together, the findings indicate that fiscal adaptation under political pressure is
associated with $112 per capita increase in total expenditure and $22 per capita increase in public
safety expenditure which diverges from the direct effect of the treatment that is associated with a
$80 per capita decrease in total expenditure and a $22 per capita decrease in public safety as well
as the $123 per capita decrease in total expenditure that is results from the mediating effect of
professional management. These distinctive fiscal adaptations in service provision point to the
salience of institutional mediating factors in shaping fiscal outcomes.
5.2 Fiscal Adaptation through Non-Tax Revenues
Table 12 presents results for the effect on non-tax revenues. I examine two different
sources of non-tax revenue – fees for services, and revenues raised through enforcement fines and
forfeitures. Columns (1) and (2) present results for the causal effect of negative revenue shocks.
64
As observed in column (1), the effect of negative revenue shocks on service fees is not statistically
significant. Column (2) reveal that, contrary to earlier research which finds fiscal stress increases
enforcement fines, I find that negative revenue shocks cause an average 10-percentage point
decrease in revenue from fines and forfeitures (Goldstein et al., 2020; Su, 2020). In relative year
one, revenue from fines and forfeitures decreases by 7.5 percentage points and by a further 3.5
percent to 11 percentage points two years after the negative revenue shock. However, the causal
effect diverges significantly from results for institutionally mediated effects.
Columns (3) and (4) show that the mediated effect of professional management on non-tax
revenues is not statistically significant. Columns (5) and (6) present results for the mediating effect
of political pressure on non-tax revenues and the findings confirm hypothesis 2. Political pressure
is associated with a statistically significant increase in enforcement revenues but not service fees.
Revenues from fines and forfeitures increase by 9.5 percentage points in relative year two and by
a cumulative of 17 percentage points in relative year four.
A similar pattern of findings that are observed for the mediated effect of pension stress
confirm hypothesis 3. Columns (7) and (8) show that the added fiscal stress induced by unfunded
pension liabilities is associated with 3.3 percentage point to 7.2 percentage points over a four-year
post-treatment period. The effect size translates to an average $60,000 increase in enforcement
revenues in a mid-sized cities with 70,000 residents.
The findings of fiscal adaptation through non-tax revenue channels have implications for
social equity. Racial minorities bear a disproportionate burden of enforcement fines and forfeitures
(Pacewicz & N Robinson, 2020; Singla et al., 2020), increase in enforcement revenues has an
adverse impact on the well-being of racial minorities and low-income households. The results
suggest that although social equity is not hindered by the direct effect of negative revenue shocks,
65
the heightened demand on local revenues that results from political pressure and the stress of
unfunded pension liabilities has significant adverse impact on social equity. These findings reveal
that both the mechanism of fiscal adaptation as well as the institutional factors that mediate
adaptation determine whether unanticipated disruptions affect social equity.
Table 12. Fiscal Adaptation Strategy: Effect on Non-Tax Revenues
Mediating Effect of Institutional Factors
Negative Revenue
Shock
Professional
Management
Political
Pressure
Pension
Stress
Fees Fines Fees Fines Fees Fines Fees Fines
(1) (2)
(3) (4)
(5) (6)
(7) (8)
1 year later 0.02
(0.02)
-0.07
*
(0.03)
-0.05
(0.04)
-0.04
(0.04)
-0.02
(0.02)
0.04
(0.03)
0.003
(0.01)
0.02
(0.02)
2 years later -0.03
(0.02)
-0.11
**
(0.03)
-0.07
(0.04)
0.03
(0.07)
0.03
(0.03)
0.09
*
(0.04)
-0.003
(0.02)
0.03
*
(0.02)
3 years later 0.003
(0.02)
-0.09
*
(0.04)
-0.04
(0.04)
-0.02
(0.06)
-0.005
(0.03)
0.09
(0.05)
-0.008
(0.02)
0.04
*
(0.02)
4 years later 0.006
(0.03)
-0.13
*
(0.05)
-0.02
(0.05)
-0.06
(0.07)
0.003
(0.03)
0.17
**
(0.06)
0.006
(0.02)
0.07
*
(0.03)
5 years later -0.02
(0.03)
-0.09
(0.05)
-0.01
(0.04)
-0.12
(0.06)
0.02
(0.03)
0.12
(0.06)
0.003
(0.02)
0.02
(0.02)
Constant -0.02
*
(0.01)
-0.03
**
(0.01)
-0.02
(0.01)
-0.07
(0.04)
-0.02
(0.01)
-0.03
**
(0.01)
0.01
(0.05)
-0.14
**
(0.05)
N 23,608 23,338 23,521 23,251 23,608 23,338 10,758 10,660
R2 0.46 0.15 0.46 0.15 0.46 0.15 0.59 0.09
Note: Coefficients are in log points. Empirical models include fiscal year and election year fixed effects.
Standard errors are clustered at the city level. Pension stress is log transformed.
***
p < 0.001
**
p < 0.01
*
p < 0.05
66
5.3 Fiscal adaptation through short- and long-term fiscal solvency
Table 13 presents results on fiscal adaptation through solvency. Fiscal solvency provides
insights on both short-term as well as long-term financial health. Columns (1) and (2) present the
direct effect of negative revenue shocks on fiscal solvency. As observed, there is no significant
effect on either short- or long-term fiscal solvency. Columns (3) and (4) which present results on
the mediated effect of professional management demonstrate a similar pattern that supports
hypothesis 1. While the effect is not statistically significant, the findings indicate that both short-
and long-term solvency remain positive, with the exception of single period in which draw down
in reserves causes a dip in short-term solvency.
Columns (5) and (6) present results on effect of political pressure and support hypothesis
2. The findings reveal that while the effect of political pressure on short-term solvency is not
statistically significant, there is a statistically significant effect on long-term solvency that ranges
from 2.3 and 3.5 percentage points over a three-year post-treatment period. This indicates that
general fund assets decrease in response to the combination of fiscal pressure from the negative
revenue shock and political pressure from unions. Columns (7) and (8) show that the mediating
effect of unfunded pension liabilities is not statistically significant.
The findings for the effect of negative revenue shocks on fiscal solvency are relatively
muted with the exception of the mediating effect of political pressure. When interpreted in
perspective of earlier findings which showed that political pressure is associated with higher
service expenditure as well as increase in non-tax revenues, these results suggest that expenditure
exceed revenues resulting in overall decline in long-term fiscal solvency.
67
VI Discussion and Conclusion
The recent disruption caused by Covid-19 upended the fiscal stability of state and local
governments. While Covid-19 has been described as a black swan event, the abrupt disruption it
caused raises a critical question on how local governments adapt to unanticipated fiscal
disruptions. The demand for adaptation is heightened when such forms of unanticipated
disruptions are routinized. When faced with fiscal uncertainty that is caused by a negative revenue
shock, I examine the mechanisms through which local governments demonstrate fiscal adaptation
in addition to analyzing the impact of institutional factors on adaptive outcomes. More importantly,
I investigate the extent to which fiscal adaptations affect social equity. I analyze how distinct
adaptive mechanisms – service provision, non-tax revenues, and fiscal solvency – interact with
Table 13. Fiscal Adaptation Strategy: Effect on Fiscal Solvency
Mediating Effect of Institutional Factors
Negative Revenue
Shock
Professional
Management
Political
Pressure
Pension
Stress
Short-
term
Long-
term
Short-
term
Long-
term
Short-
term
Long-
term
Short-
term
Long-
term
(1) (2)
(3) (4)
(5) (6)
(7) (8)
1 year later 0.008
(0.01)
-0.005
(0.01)
0.006
(0.01)
0.019
(0.01)
-0.006
(0.01)
-0.023
*
(0.01)
0.0002
(0.006)
-0.002
(0.006)
2 years later 0.002
(0.01)
0.016
(0.01)
-0.002
(0.013)
0.024
(0.01)
-0.011
(0.01)
-0.025
*
(0.01)
0.005
(0.006)
-0.002
(0.006)
3 years later 0.008
(0.01)
0.009
(0.01)
0.021
(0.019)
0.028
(0.01)
-0.022
(0.02)
-0.035
*
(0.01)
0.006
(0.008)
0.002
(0.008)
4 years later 0.005
(0.01)
0.003
(0.02)
0.017
(0.018)
0.006
(0.01)
-0.007
(0.02)
-0.02
(0.01)
0.002
(0.008)
0.006
(0.007)
5 years later 0.018
(0.01)
0.016
(0.02)
0.005
(0.019)
-0.001
(0.01)
-0.013
(0.01)
-0.02
(0.01)
-0.004
(0.01)
0.01
(0.009)
Constant 0.072
***
(0.01)
0.13
***
(0.02)
0.095
***
(0.016)
0.17
***
(0.02)
0.092
***
(0.01)
0.17
***
(0.02)
0.06
*
(0.03)
0.17
***
(0.03)
N 19,071 19,133 23,379 23,464 23,466 23,551 10,636 10,724
R2 0.036 0.08 0.035 0.09 0.034 0.098 0.046 0.153
Note: Coefficients are in log points. Empirical models control for log per capita expenditure, and fiscal year and
election year fixed effects. Standard errors are clustered at the city level. Pension stress is log transformed.
***
p < 0.001
**
p < 0.01
*
p < 0.05
68
institutional factors – professional management, political pressure, and pension stress – to identify
the direct causal effect of negative revenue shocks and the mediated effects of institutional
conditions.
The empirical strategy exploits exogenous shocks to local government revenues that result
from voters’ decisions on local tax and debt referendums. I estimate the dynamic effect of revenue
shocks for multiple post-treatment years. The main finding negative revenue shocks have a causal
effect on service provision, but non-tax revenue and fiscal solvency are unaffected. More
specifically, there is a significant decline in total expenditure and public safety expenditure that
persists up to four years after the shock.
The results diverge significantly when mediated by the institutional factors. Professionally
managed cities demonstrate a pattern of fiscal adaptation that confers with principles of prudent
financial management. A negative revenue shock is managed by decrease in service expenditures
and drawn down in reserves, these strategies obviate the need to increase non-tax revenues and
also preserve fiscal solvency. However, local governments under political pressure from internal
interest groups reveal a pattern of fiscal adaptation that aligns with maladaptation because of the
adverse impact on social equity. In particular, political pressure is associated with higher service
provision as well as a significant increase in enforcement revenues in addition decline in long-term
fiscal solvency. A similar pattern is observed effect for pension stress, significant increase in
enforcement revenues indicates fiscal maladaptation.
The conceptual and analytical framework of fiscal adaptation proposed in this paper also
inform ongoing debates on the promise and pitfalls of adaptation by public organizations. It is
argued that while adaptation is a necessity in today’s environment of constant change, studies show
that adaptive public organizations are subject to internal instability which not only decreases
69
organizational performance but is particularly detrimental to outcomes for minorities and low-
income households (Ansell et al., 2020; Singh et al., 1986; Boyne & Meier, 2009; O'Toole &
Meier, 2010). The findings presented here reveal the complexity of adaptation and more
specifically the complexity of fiscal adaptation. I show that the mechanisms through which fiscal
adaptation occurs as well as the institutional conditions that mediate adaptive outcomes have
significant bearing on the ability of public organizations to advance broader societal goals.
There is need for further research on how public organizations navigate the tenuous
relationship between fiscal crises and social equity. Future studies that employ disaggregated data
can help examine the impact of fiscal choices on racial and ethnic groups. Such research will be
particularly informative when assessing different forms of disruptive events - fiscal as well as non-
fiscal. For instance, events such as wildfires, floods, or hurricanes can be just as unanticipated as
a negative revenue shock induced by voters’ direct democratic decisions.
This study is also contributes to knowledge on local direct democracy, an area that has
received limited research attention (Matsusaka, 2018). In recent years, several policy innovations
such as increase in minimum wages, adoption of the sugar sweetened beverage aimed at reducing
childhood obesity, and cannabis regulation which expanded access to medical marijuana were
approved by voters through local direct democracy. Further research on the impact of direct
democracy on minority well-being and local public economies will contribute to strengthening
participatory policy making at the local level.
70
Chapter 4:
Resist, Recover, Renew:
Fiscal Resilience as a Strategic Response to Economic Recessions
71
Resist, Recover , Renew:
Fiscal Resilience as a Strategic Response to Economic Recessions
Manita Rao
PhD Candidate, Public Policy, and Management
Sol Price School of Public Policy and Urban Planning
University of Southern California
Juliet Ann Musso
Associate Professor
Sol Price School of Public Policy and Urban Planning
University of Southern California
Matthew M. Young
Assistant Professor
School of Governance and Global Affairs, Leiden University
ABSTRACT
Economic disruptions such as the COVID-19 pandemic have made clear how vulnerable
local governments are to fiscal shocks. This article centers the concept of fiscal resilience as a
frame for public financial management Boin & Van Eeten, 2013; Duit, 2016. It introduces and
empirically examines a dynamic framework of fiscal resilience, in which resilience is decomposed
into pre-crisis fiscal resistance, post-crisis fiscal recovery over the short-term, and long-term fiscal
renewal. The empirical modeling disentangles the strategic and structural determinants of fiscal
resilience with respect to two previous recessions – the Dotcom and the Great Recession. The
findings from Cox proportional hazard models indicate that strategic decisions such as revenue
diversification and countercyclical policies facilitate fiscal resilience. Moreover, local revenue and
service structures are critical to fiscal recovery and renewal. In addition, the underlying
characteristics of each recession affect whether institutional and economic conditions enable local
government resiliency. The work has broad implications for public financial management, which
should embed resiliency-based frameworks in fiscal planning.
Keywords: Fiscal resilience, local government, public financial management
72
I Introduction
This article develops and empirically models a dynamic construct of fiscal resilience, to
assess the fiscal viability of local governments confronting an environment of increasing economic
perturbation. A number of recent disruptions have altered the resource landscape of public
organizations globally, most recently the economic impact of COVID-19, but also the Great
Recession of 2008-2010. These more extreme disruptions punctuate a trend of pervasive fiscal
uncertainty associated with industrial globalization, waning citizen trust in government, and
unpredictable economic cycles. The shift to high uncertainty requires that organizations develop
the ability to resist adverse impact, bounce back from disruptions, and return stronger than before.
This paper argues that the construct of resilience, originating in the field of ecology, captures these
characteristics. It addresses the enduring puzzle posed by Boin and Van Eeten, who stated: “It is
not clear what resilience is exactly and it is hard to recognize resilience in action” Boin & Van
Eeten, 2013, p.430). This paper employs an interdisciplinary perspective that integrates
ecological, organizational, and regional economic theories in support of a dynamic conceptual
framework of fiscal resilience. It employs the model in an empirical examination of the fiscal
response of local governments to the two major economic recessions of the 2000s. The findings
emphasize the importance of both anticipatory capabilities as well as structural and institutional
features of the fiscal structure.
The empirical strategy is the estimation of Cox proportional hazard models that identify
the capacity for fiscal resilience as demonstrated through fiscal recovery and fiscal renewal during
the milder dotcom recession of the early 2000 and the high severity great recession in 2008-2010.
Adopting an equilibrium framework to operationalize fiscal recovery and fiscal renewal, the model
estimates the hazard of fiscal resilience as a function of local strategic decisions which include
73
revenue diversification, adoption of countercyclical policies, and the local governments’ revenue
and service structure. The study employs over three-decades of fine-grained fiscal data to identify
the dynamic impact of economic recessions on fiscal resiliency. Our findings provide new
evidence on how urban places recover from recessionary pressures and further the conversation
on designing resilient public institutions.
The rest of this paper is organized in five sections. The first section reviews resiliency
theory with a conceptual application to local government finances. The second section outlines
hypotheses on the determinants of fiscal resilience. Section three introduces the data and empirical
strategy; the final two sections present results and the concluding discussion.
II Theory: Dimensions and Determinants of Fiscal Resilience
This paper contributes to the literature on local government fiscal assessment, a literature in public
administration and policy that stretches back several decades. The longest-standing strand of this
literature, on fiscal capacity, reaches back to the 1970s with research by Ladd and Bradbury (1988),
Chernick (1998), Reschovsky (1980), and others. A central thrust of this literature is to determine
means of assessing the fiscal capacity or fiscal strength of cities, recognizing the differences in
resources, service demands, and institutional features. Extensions of this literature consider the
factors that explain fiscal disparities across regions or states, and fiscal stress experienced by some
cities. More recently, the framework of fiscal sustainability has emerged as an extension to the
fiscal capacity framework, acknowledging that organizations must balance service demands in
light of revenue portfolios confronting longer-term stresses, such as aging of the population and
infrastructure investment needs (Tang et al., 2014). This sustainability lens represents an
important improvement over the snap-shot approach characterizing some of the earlier fiscal
74
capacity literature. The current paper extends the concept of fiscal sustainability, considering the
dynamics of increasingly volatile fiscal environments. It builds on theories of fiscal capacity and
fiscal sustainability in considering resilience as an attribute of systems that can anticipate and adapt
to disruptive change.
The theory of resilience has been applied across disciplines as diverse of ecology,
economics, psychology, and organizational theory. The touchstone of resilience is its emphasis on
change – and the manner in in which an entity – an individual, organization, ecosystem – reacts
and recovers in response to a disruptive event. The response of the organization might involve
“bouncing back,” or recovery, in which the question is whether and how quickly the organization
can return to its previous growth trajectory. It has also been noted that disruption can create new
opportunities that allow an organization to “bounce forward,” moving to a new and more favorable
trajectory. The goal of this paper is to consider fiscal resiliency from a strategic perspective; in
other words, what can public financial managers do to build resilience in anticipating the likelihood
of disruptive events?
II A.1 Theoretical antecedents
The concept of resilience is widely applied in the interdisciplinary fields of ecology,
economics, and psychology. Resilience provides a useful conceptual tool to examine the behavior
demonstrated by systems, organizations or individuals in response to uncertainty and crises
Berkes, 2007. The theoretical lenses in resilience can be broadly grouped into two categories: an
ecological perspective that emphasizes recovery as a property of self-organizing systems, and
organizational perspectives where resilience is an outcome of strategic decisions. Ecological
resilience adopts a complex adaptive system perspective in which resilience is perceived as the
75
ability of systems to absorb changes and still persist Holling, 1973. Resilient systems demonstrate
recovery and persistence when disrupted by a shift in the equilibrium ‘state’ or the system’s
internal relationships. Recovery is the speed with which the system overcomes the disruption and
returns to its pre-disrupted state, while persistence is long-term resilience that is exhibited by the
ability to absorb internal and environmental changes while persisting along a growth trajectory.
The extent to which a system demonstrates recovery or persistence is theorized to be determined
by configurations of wealth (inherent potential), connectedness (control of internal functions), and
adaptive capacity Holling, 2001. The application of these ecological resilience concepts to fiscal
systems indicates that while a localities economic wealth is important to recovery, system
resiliency is enhanced through connectedness and capabilities to adapt to disruptive events. Han
& Goetz, 2015; Martin et al., 2015 .
Despite its wide application in economic and regional sciences, ecological resilience is
critiqued by organizational theorists for its non-agentic perspective of resilience. Duit (2016)
argues that ecological resilience presupposes those systems have self-organizing capacities that
enhance resiliency and are therefore both politically naïve and agent-agnostic. However, fiscal
systems are embedded in organizational contexts in which interventions by actors such as public
managers, political groups, and voters will influence outcomes. Comfort et al. (2001), for instance,
defines resilience as a solution to “the problem of how to increase the capacity of interdependent
public organizations to anticipate risk and respond to threats” (p.144; also see Wildavsky, 1988).
Alternatively, resilience is guided by change-oriented theories where disruptions are understood
to open windows of opportunity that develop new capabilities and increase organizational capacity
to bounce forward Barbera et al., 2021; Boin & Lodge, 2016.
76
Constructing resilience as an organizational capability highlights its strategic, and dynamic
dimensions. Resilience is a function of both the ability to build anticipatory capacities that
contribute to crisis preparedness as well as the speed of recovery after a crisis occurs (Boin & Van
Eeten, 2013. Recent research focusses on the development of meta-capabilities composed of
anticipation, coping, and adaptation as important elements of strategic resilience (Duchek, 2020.
Anticipation concerns the preventive aspects of resilience, while coping is the short-term response
to an unexpected event, and adaptability is the ability to embed learning and change management
processes that can facilitate long-term stability to weather recurrent disruptions. While there is
now a general consensus of resilience as a dynamic process, operationalizing the meta-construct
of resilience remains a challenge in the academic literature in public administration and, in
particular, public finance. This paper advances fiscal strategy theory by introducing a conceptual
framework of fiscal resilience, operationalizing its elements, and empirically evaluating factors
that contribute to fiscal resilience.
II A.2. Overview of the Fiscal Resilience Framework
A fiscally resilient public organization demonstrates the strategic capabilities to absorb
shocks with minimal revenue and service disruptions. Fiscal resilience is a dynamic process that
embeds strategic resilience capabilities in key fiscal functions to achieve revenue adequacy,
service continuity, and balanced fiscal growth. A dynamic theory of fiscal resilience is
conceptualized as a sequential process of pre-crisis fiscal resistance, fiscal recovery that coincides
with the crisis, and post-crisis fiscal renewal, as shown in Figure 3. The dimensions of fiscal
resilience are adapted from Martin’s (2012) process-based theory of regional economic resilience
77
which is comprised of the capability to resist (or absorb) shocks, the capacity to recovery from
disruptions, and ability to demonstrate renewal, or longer-term fiscal growth.
Figure 3 presents the fiscal resilience framework. The framework is a 3 x 3 matrix that
unifies resilience dynamics, resilience dimensions and the determinants of fiscal resilience. Fiscal
resilience is a capability that is generated dynamically through feedback between pre-crisis, crisis
response, and post-crisis periods. The dimensions of fiscal resilience are adapted from a Martin’s
(2012) process-based theory of regional economic resilience which is comprised of the capability
to resist (or absorb) shocks, the capacity to recovery from disruptions, and ability to demonstrate
renewed fiscal growth.
Fiscal resistance is determined by strategic decisions before a crisis occurs such as building
fiscal reserves before a crisis to resist the adverse impact of an economic slowdown on local
revenues and potential protect service provision. Fiscal recovery, in turn, depends on decisions
that occur during a crisis period, that affect the ability to recover from revenue loss and maintain
service levels. Economic recessions have typically had a stronger impact on local revenues as
observed, for example, during the Great Recession (2008-2010) when several local governments
witnessed substantive shortfalls in property tax collections Kim, 2019. Recent evidence from the
Covid-19 pandemic, however, indicated that pressures on public services are also important factors
Figure 3. Conceptual Framework of Fiscal Resilience
Resilience Dynamics Pre-Crisis Crisis Period Post-Crisis
Resilience Dimensions Fiscal Resistance Fiscal Recovery Fiscal Renewal
Resilience Determinants Buffer and Boost Service Recovery
Revenue Recovery
Balanced Fiscal Growth
78
that can affect economic recovery. Some cities witnessed a 75% increase in demand for health and
social services during Covid-19, a level unprecedented in earlier recessions highlighting the need
to consider both revenue and service features to assess fiscal recovery Gordon et al., 2020. Fiscal
renewal is a post-crisis phenomenon, defined as the ability to achieve a balance between revenue
growth and service demands. In principle, the three key components of fiscal resilience –
resistance, recovery, and recovery – will be affected by both public managers’ strategic decisions
as well as structural elements of public budgets.
II B. Determinants of Fiscal Resilience: Hypotheses
This section elaborates on four distinct pathways that affect fiscal resiliency: buffering,
boosting, revenue structure, and service structure. The notion of buffering is broadly defined as an
organization’s internal capacity to mitigate the adverse impact a disruptive event can have on
fiscal outcomes . Buffering is enhanced when accumulated internal resources either personnel or
financial are utilized to maintain service stability O'Toole & Meier, 2010. In contrast, boosting is
an outward response targeted toward deliberate increase in public spending in order to stimulate
economic growth. Boosting occurs through the adoption of countercyclical fiscal policies at either
the local, state, or federal level Hou & Moynihan, 2008. However, policies that boost local
economic growth can be curtailed by local fiscal capacity and a weak economic structure.
Hypothetically, because both buffering and boosting can shift a system away from its pre-crisis
status-quo, these mechanisms operate parallel to structural features that aid fiscal resilience.
79
II B.1. The Buffering Mechanism: Revenue Diversification
Revenue diversification is theorized to buffer against disruption, because a diversified
portfolio stabilizes investment returns by minimizing loss and maximizing returns Markowitz,
1976. Cities with diversified revenue are also more fiscally stable Carroll, 2005; Hendrick &
Degnan, 2020. However, achieving a diversified revenue portfolio is determined by factors such
as city size, economic structures and on institutional constraints all of which affect tax revenues.
Smaller cities , for example, are more dependent on property taxes while larger cities can be
expected to have a more diverse portfolio mix. Some cities do nor levy sales taxes while other are
more dependent on personal income taxes. Such differences affect the extent to which revenue
diversification is a viable strategy for local fiscal resilience.
Institutional constraints also play an important role in revenue diversification. Constraints
imposed by California’s Proposition 13, for example, which caps property tax levies at 1% of
assessed property value, and 2% for annual increases in assessed values will affect property tax
collections and the contribution of this revenue source to the revenue portfolio McCubbins &
McCubbins, 2009. Similar institutional requirements are pervasive in other states. Massachusetts,
for instance, has proposition two-and-a-half and Michigan’s proposition A limit property tax levies
Skidmore & Scorsone, 2011. Whether the local economy depends primarily on manufacturing, or
the technology sector will affect portfolio composition. Further, suburban cities will differ
considerably from central cities in the types of revenues accessible to these cities. In general,
research shows that cities with diversified economic structures are more resilient and have higher
capacity to buffer against sector specific business cycle fluctuations that can affect individual
revenue streams Martin et al., 2015.
80
However, the benefits of economic diversification for resiliency do not translate directly to
disruptions that cause revenue fluctuations. A more diversified revenue portfolio may decrease the
need for local governments to levy taxes thereby stimulating consumer spending Jordan et al.,
2017; Suyderhoud, 1994. At the same time, cities most in need of revenue diversification – small
and mid-sized localities – may be least likely to have such a portfolio which could increase the
challenge of fiscal recovery Hendrick, 2002. Previous research also shows that the diversification
is attenuated by the stability of the local economic base and cities that are more susceptible to
business cycle fluctuations will experience higher benefits from diversification of the local
economy. In sum, while the ability to achieve a diversified revenue portfolio depends on local
economic capacity and the relationship to resilience is nuanced, on balance, revenue diversification
is expected to have important buffering effects. Expressed as a hypothesis:
H1: Revenue diversification buffers against decline in revenue which in turn facilitates fiscal
resilience (recovery and renewal).
II B.2. The Boosting Mechanism: Countercyclical Policy
Countercyclical policy is defined as the financial decisions during economic downturns
that stimulate economic growth through increased public spending. Countercyclical policies are a
staple macroeconomic tool that federal governments employ to dampen the impact of recessions
on economic activity. The Coronavirus Aid, Relief, and Economic Security (CARES) Act rolled
out during the Covid-19 recession was a federal countercyclical policy intervention that provided
financial transfers to state and local governments as well as households with the objective of
boosting spending and economic recovery. During recessions, state and local governments engage
in similar countercyclical policy interventions by utilizing reserves and increasing public service
81
expenditures Gorina et al., 2019. However, unlike the federal government, the ability of state and
local governments to undertake countercyclical spending depends on whether or not the
governmental entity has accumulated financial reserves it can utilize to offset revenue shortfalls.
In addition, contractual obligations related to debt, compensation, and pension obligations and
institutional balanced budget mandate can outweigh discretionary spending on public services
Gorina et al., 2019; Su, 2019. Financial reserves are more likely to contribute to countercyclical
spending when political economy factors such as politicians’ electoral incentives and election
proximity are less salient to fiscal decisions Kim & Park, 2021; Maher & Deller, 2007; Wang &
Hou, 2012.
In sum, countercyclical policies are a tried and tested intervention to boost economic
activity during a recession. Despite the limited flexibility of local governments to engage in
countercyclical spending, accumulated financial reserves could hold promise in helping localities
dampen recessionary pressures and boost the local economy. Stated as a hypothesis:
H2: Countercyclical policy boosts public spending which in turn facilitates fiscal resilience
(recovery and renewal).
II B.3. Mediating Mechanism: Revenue Structure
Fiscal resilience is mediated by revenue structure, the composite of taxes, fees, and
transfers that are provided in exchange for public services. Localities raise revenues by levying a
variety of taxes including property tax, sales tax, hotel and motel tax, business license tax, and in
recent years, cannabis tax. Following the property tax revolution, fees and programmatic transfers
have become more important components of the local fisc. There is limited evidence on the
differential impact of economic recessions on tax and non-tax, fee-based revenues. For example,
82
while property tax revenues declined during the Great Recession, Mikesell and Mullins (2013)
point out that this impact was surprisingly mild (also see Dadayan, 2012; Pagano et al., 2012. One
explanation is that legislative limits on property taxes create a wedge between market value and
assessed values thereby dampening the short-term impact of an economic recession and
contributing to revenue stability Ross & Mughan, 2016.
Consumption taxes such as sales and occupancy taxes are more susceptible to fluctuations
due to economic recessions than are property taxes. Early evidence from the recent Covid-19
recession indicates state and local sales tax revenues declined by 17 percent over a one the one
year period between April 2019 and April 2020 Clemens & Veuger, 2020. Because dependence
on sales tax increases revenue elasticity, localities that are more dependent on sales and motel
taxes experienced weaker recovery during the Great Recession Wang, 2015. This could partly
result from the interconnectedness between revenue sources Chernick et al., 2011. While most
studies suggest sales tax dependence hurts recovery, the effect could be neutral (Alm & Sjoquist,
2014) or under some special conditions aid in fiscal recovery (Anderson & Shimul, 2018).
Apart from tax revenues, intergovernmental transfers constitute an important revenue
source for most local governments, especially small and mid-sized cities. However, there is limited
investigation of how dependence on transfers affects fiscal recovery. Carrol and Stater (2009)
suggest that to the extent that intergovernmental transfers fluctuate during an economic recession,
high transfer dependence exposes cities to revenue volatility and reduces their capacity to resist
recessionary pressures. At the same time, federal stimulus payments during recessionary years
will likely assist with service provision. As a result, the relationship between dependence on
transfers and fiscal recovery is complex. For example, during the dotcom recession, state aid to
local governments dropped by nearly 10% Hoene & Pagano, 2003. In contrast, during the Great
83
Recession, cities that witnessed a decline in intergovernmental transfers were able to recover 27%
through alternate state and federal sources such as programmatic grants Chernick & Reschovsky,
2017. The CARES Act passed in 2019 witnessed the largest stimulus transfers to state and
localities, a cumulative transfer of $150 million Green & Loualiche, 2021. While these transfers
helped offset decline in public service expenditures, the mandates prohibited utilization of those
funds for expenses related to employee compensation and pension obligations which left fiscally
strained localities in a bind to either layoffs employees or defer pension contributions Wang &
Pagano, 2017. Importantly, transfers from states to localities can be expected to decrease during
an economic recession when states look to balance their own budgets thereby aggravating fiscal
pressures faced by localities Hendrick & Degnan, 2020. The hypothesized impact of revenue
structure on fiscal resilience is stated as follows:
H3(a): Dependence on property tax increases fiscal stability which in turn facilitates fiscal
resilience (recovery and renewal).
H3(b): Dependence on income elastic revenue sources (sales and transient occupancy taxes)
decrease fiscal stability which in turn hinders fiscal resilience (recovery and renewal).
H3(c): Dependence on intergovernmental transfers decrease fiscal stability which in turn hinders
fiscal resilience (recovery and renewal).
II B.4. Mediating Mechanism: Service Structure
Service structures are the composite of all public services and cost commitments in local
budgets. Localities provide a range of public services including public safety, economic
development, redistribution, sanitation, and infrastructure in addition to managing cost
commitments related to debt servicing, employee compensation and pension obligations. In
84
principle, the degree of expenditure flexibility will affect the capability of a locality to recover
from economic recessions. One approach to fiscal recovery is by adopting a stage-wise strategy in
which cutbacks to non-core services such as infrastructure and pension obligations preceded
reductions to core public safety services Raudla et al., 2013; Wolman & Peterson, 1980.
While rational, stage-wise cutbacks can be affected by stakeholder interests and the need
to minimize taxpayer burden Jimenez, 2014; Scorsone & Plerhoples, 2010. In particular, since
personnel costs constitute a significant portion of state and local budgets, these costs will typically
be managed by furlough, hiring freeze, or leaving vacancies unfulfilled which could result in
having unintended consequences on maintaining public service levels Hoang & Goodman, 2018;
Justice & Yang, 2018; Raudla et al., 2013.
These factors make cutbacks to capital infrastructure spending an attractive proposition
that can yield immediate budgetary savings Brien et al., 2021. Nevertheless, capital investments
are often the tool incorporated in stimulus packages such as under the American Recovery and
Reinvestment Act (ARRA) , the Build America Bonds (BAB), and the recent Coronavirus Capital
Projects Fund (CCPF). In keeping with the expectation that infusion of federal funds dedicated to
capital financing will ease fiscal pressures, Fisher and Wassmer (2015) find that during both the
dotcom recession in 2001 and the Great Recession, localities utilized stimulus grants to free up
monies from local budgets on capital expenditure and channel those revenues toward service
provision. However, because most grants and stimulus monies are time bound and mandated for
“shovel ready” projects, localities without projects under development are less likely to benefit
from federal stimulus packages earmarked for insfrastructure spending Hanak, 2009; Miller et al.,
2021. Based on these multidimensional features of local service structure, the hypothesized
relationship between service structure and fiscal resilience is stated as follows:
85
H4(a): Capital infrastructure and debt servicing increase economic activity which in turn
facilitates fiscal resilience (recovery and renewal).
B4(b): Compensation and pension obligations limit expenditure flexibility which in turn hinders
fiscal resilience (recovery and renewal).
III Methods, Data and Empirical Strategy
The lack of consensus on operational measurement of resilience has limited empirical
modeling such that it is generally used metaphorically Duit, 2016. This paper develops an
empirical model by creating operational measures of recovery and renewal for both revenues and
expenditures. We then estimate a dynamic model to identify factors in the pre-crisis stage and the
post-crisis phase that contribute to local fiscal resiliency. The empirical approach introduced in
this paper makes progress over earlier techniques used to measure financial resilience such as
sensitivity-analysis Lee & Chen, 2021, and a composite resilience index Barbera et al., 2021.
Unlike previous methods, the empirical approach used in this paper provides insights into the
relative salience of strategic and structural factors in facilitating fiscal resilience.
III A.1. Fiscal Equilibrium and Fiscal Recovery
The concept of fiscal equilibrium draws on an equilibrium-based approach in which pre-
crisis fiscal balances serve as the benchmark equilibrium to examine post-crisis change in revenues
and expenditure. We employ a two-stage process, the first step establishing a smoothed pre-crisis
fiscal equilibrium for revenues – revenue equilibrium and expenditure on public services - service
equilibrium. The second step identifies post-crisis recovery relative to the pre-crisis equilibrium.
Empirically, fiscal recovery was identified by evaluating if revenue and expenditures equaled or
86
exceed their pre-crisis equilibrium levels. In figure 4 we use the Great Recession as the nodal crisis
event and present stylized graphs for revenue and service recovery using the equilibrium-based
approach. On the revenue side, the three-year smoothed equilibrium is $1,462 per capita and
revenue recovery shows as having recovered by FY 201l, the first post-crisis year in which pre
capita revenue exceeded its pre-crisis equilibrium level. On the expenditure side, pre-crisis service
equilibrium is $1,011 per capita and recovery shows as having recovered by FY 2012, the first-
post-crisis year in which per capita spending exceeded the pre-crisis equilibrium level.
The stylized graph for revenue recovery is given by a revenue equilibrium at $1,462 per
capita, revenue recovery occurs in FY 2011, the first post-crisis year in which pre capita revenue
exceeds equilibrium. The service equilibrium is at $1,011 per capita, recovery occurs in FY 2012:
the first-year post-recession year where total spending exceeded pre-crisis equilibrium.
Figure 4. Stylized Figure of Fiscal Recovery
Note: Stylized figure of revenue and expenditure recovery. The crisis event (shown as the vertical dotted line) is the
Great Recession which began in 2008.
Revneue Equilibrium
Revenue Recovery
1400 1500 1600 1700 1800
per capita revenue
2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016
fiscal year
Revenue Recovery
Service Equilibrium
Service Recovery
800 1000 1200 1400 1600
per capita expenditure
2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016
fiscal year
Service Recovery
87
III A.2. Operationalizing Fiscal Renewal
Fiscal renewal focusses on the ability of a public entity to achieve a balanced fiscal growth
path after the crisis has passed. It is broadly defined as the system shift such that the annualized
difference between revenue and expenditure is positive with minimum revenue volatility. Figure
5 depicts a stylized version of fiscal renewal with two crisis events as examples - the dotcom
recession in 2001 and the Great Recession in 2008-2010.
As observed in Figure 5, the dotcom recession was followed by a U-shaped recovery with
sustained growth occurring after FY 2004, a three-year lag between the official start of the
recession in FY 2001 and post-crisis continued revenue growth. The Great Recession, however,
indicates a V-shaped recovery with a sustained growth path from FY 2010, and a two-year lag
from the start of the recession in FY 2008-2010. The stylized figures also demonstrate that
recovery can vary not only be recession but also between local government entities. The exact path
Fiscal Renewal
-.05 0 .05 .1
growth rate for fund balance
2000 2001 2002 2003 2004 2005 2006 2007
fiscal year
Fiscal Renewal
-.2 0 .2 .4
growth rate for fund balance
2005 2006 2007 2008 2009 2010 2011 2012 2013 2014
fiscal year
Figure 5. Stylized Figure of Fiscal Renewal
Dotcom Recession Great Recession
88
of fiscal renewal is contingent, as hypothesized, on both structural and strategic characteristics of
local government revenues and expenditures.
III B. Data and Variables
This study uses three decades of fine-grained fiscal data on 480 city governments in
California. The time period of the study – 1991 to 2018 – covers two previous recessions: the
dotcom recession (2001) and the Great Recession (2008-2010). Information on local government
revenues and expenditures are publicly available from the California State Controller’s open data
portal. The dataset used for analysis includes data on tax, non-tax, and transfer revenues,
expenditure-by-category, debt, and administrative spending. We also include an extended set of
institutional and control variables on city incorporation, size, demographic characteristics, political
affiliations, and economic conditions. San Francisco is excluded from the analysis because of its
status as a city and county. We also exclude the city of Sand City, the city of Vernon, the city of
Industry, and the city of Colma which have fewer than 100 residents and are largely industrial.
Date on control variables for fiscal years 2005 to 2018 comes from the American Community
Survey (ACS) and from the National Historical Geographic Information System (NHGIS) for
fiscal years 1991 to 2015. The 5-year ACS and NHGIS are used to linearly interpolate data on
control variables.
The three main dependent variables are – revenue recovery, service recovery, and fiscal
renewal. Table 14 summarizes how these measures are operationalized mathematically. Fiscal
equilibrium is empirically measured as each city’s pre-crisis three-year average revenue and
expenditure for each recession. For example, the revenue equilibrium for the Great Recession is
average revenue per capita for FY 2005-2007 and for the dotcom recession, it is for FY 1997-
89
2000. To measure fiscal recovery, we create a dummy variable which takes the value 1 in the first
post-crisis year in which revenue or expenditure pre-capita equal or exceed the pre-crisis level.
Fiscal renewal is an indicator variable which takes the value 1 over any post-crisis year of
annualized positive change in general fund balance (row 4 of Table 14).
Empirically, the model estimates the effect of resiliency (recovery or renewal) with two
mechanisms of fiscal resistance - buffering (through revenue diversification) and boosting (when
countercyclical spending is undertaken). We measure revenue diversification using the
Hirschman-Herfindahl Index (HHI) and countercyclical spending is proxied by creating a variable
for financial reserves which is per capita unreserved funds in the city’s general account. Revenue
Table 14. Key Concepts in Fiscal Resilience Analysis
Concept Mathematical Expression
Fiscal Equilibrium The components of fiscal equilibrium are defined for entity 𝑖 (city) as the
three-year pre-crisis average revenue or expenditure, mathematically denoted
as:
𝑅𝑒𝑣𝑒𝑛𝑢𝑒 𝑒𝑞𝑢𝑖𝑙𝑖𝑏𝑟𝑖𝑢𝑚
!
= /𝜇1𝑅𝑒𝑣𝑒𝑛𝑢𝑒
!,#
!
23
𝑆𝑒𝑟𝑣𝑖𝑐𝑒 𝑒𝑞𝑢𝑖𝑙𝑖𝑏𝑟𝑖𝑢𝑚
!
= /𝜇1𝐸𝑥𝑝𝑒𝑛𝑑𝑖𝑡𝑢𝑟𝑒
!,#
!
23 ;
where −3 ≥ 𝑛 ≥ −1 and 𝑡
$
is the crisis year.
Fiscal Recovery:
Revenue Recovery
Defined for entity 𝑖 (city) as equal to 1 if post-crisis revenue equals or
exceeds revenue equilibrium, mathematically denoted as:
𝑅𝑒𝑣𝑒𝑛𝑢𝑒 𝑅𝑒𝑐𝑜𝑣𝑒𝑟𝑦
!
= 1 if 𝑅𝑒𝑣𝑒𝑛𝑢𝑒
!,#
!
≥ 𝑅𝑒𝑣𝑒𝑛𝑢𝑒 𝑒𝑞𝑢𝑖𝑙𝑖𝑏𝑟𝑖𝑢𝑚
!
;
where 1 ≥ 𝑛 ≥ 5
Fiscal Recovery:
Service Recovery
Defined for entity 𝑖 (city) as equal to 1 if post-crisis service expenditure
equals or exceeds service equilibrium, mathematically denoted as:
𝑆𝑒𝑟𝑣𝑖𝑐𝑒 𝑅𝑒𝑐𝑜𝑣𝑒𝑟𝑦
!
= 1 if 𝑆𝑒𝑟𝑣𝑖𝑐𝑒
!,#
!
≥ 𝑆𝑒𝑟𝑣𝑖𝑐𝑒 𝑒𝑞𝑢𝑖𝑙𝑖𝑏𝑟𝑖𝑢𝑚
!
;
where 1 ≥ 𝑛 ≥ 5
Fiscal Renewal Defined for entity 𝑖 (city) as equal to 1 if post-crisis year-on-year change in
fund balance is positive .
𝐹𝑖𝑠𝑐𝑎𝑙 𝑅𝑒𝑛𝑒𝑤𝑎𝑙
!,#
!
= 1 𝑖𝑓
1𝑅𝑒𝑣𝑒𝑛𝑢𝑒
!,#
!
−𝐸𝑥𝑝𝑒𝑛𝑑𝑖𝑡𝑢𝑟𝑒
!,#
!
2
(𝑅𝑒𝑣𝑒𝑛𝑢𝑒
!,#
!"#
−𝐸𝑥𝑝𝑒𝑛𝑑𝑖𝑡𝑢𝑟𝑒
!,#
!"#
)
> 0; where 1 ≥ 𝑛 ≥ 5
Note: 𝑡
!
is the crisis year which is 2001 for the dotcom recession and 2008 for the Great Recession.
90
structure is a vector of three revenue-type variables expressed as a share of local own-source
revenue- share of property taxes, share of income elastic revenue sources (which includes total
revenue from sales tax and transient occupancy tax), and the share of intergovernmental transfers.
Similarly, service structure is a vector of four expenditure-category variables expressed as a share
of own-source revenue - share of expenditure on compensation, share of expenditure on pensions,
share of expenditure on capital infrastructure, and share of debt servicing. All predictor and control
variables are lagged one period to account for resilience dynamics.
III C. Empirical Strategy
We estimate a Cox proportional hazard model specified as follows:
𝐻(𝑡) = 𝐻
0
(𝑡) exp (𝑏
%
𝑅𝑒𝑣𝑒𝑛𝑢𝑒 𝐷𝑖𝑣𝑒𝑟𝑠𝑖𝑓𝑖𝑐𝑎𝑡𝑖𝑜𝑛
!#1%
+𝑏
)
𝐶𝑜𝑢𝑛𝑡𝑒𝑟𝑐𝑦𝑐𝑙𝑖𝑐𝑎𝑙 𝑃𝑜𝑙𝑖𝑐𝑦
!#1%
+𝑏
2
𝑅𝑒𝑣𝑒𝑛𝑢𝑒 𝑆𝑡𝑟𝑢𝑐𝑡𝑢𝑟𝑒
!#1%
+ 𝑏
3
𝑆𝑒𝑟𝑣𝑖𝑐𝑒 𝑆𝑡𝑟𝑢𝑐𝑡𝑢𝑟𝑒
!#1%
+𝑏
4
𝑋
!#1%
+𝜀
!,#1%
)
where the outcome 𝐻(𝑡) estimates the hazard rate for three dependent variables: revenue recovery,
service recovery, and fiscal renewal. All three dependent variables are computed as shown in table
1. 𝐻
0
(𝑡) is the baseline hazard. Tests for proportionality of baseline hazards were conducted and
the results show that the proportionality assumption is satisfied. The results from these tests are
presented in the appendix.
The main predictor variables are lagged (𝑅𝑒𝑣𝑒𝑛𝑢𝑒 𝐷𝑖𝑣𝑒𝑟𝑠𝑖𝑓𝑖𝑐𝑎𝑡𝑖𝑜𝑛
!#1%
) for city 𝑖 in time
𝑡−1, lagged unreserved funds per capita which proxies for 𝑐𝑜𝑢𝑛𝑡𝑒𝑟𝑐𝑦𝑐𝑙𝑖𝑐𝑎𝑙 𝑃𝑜𝑙𝑖𝑐𝑦
!#1%
, lagged
𝑟𝑒𝑣𝑒𝑛𝑢𝑒 𝑆𝑡𝑟𝑢𝑐𝑡𝑢𝑟𝑒
!#1%
and lagged 𝑠𝑒𝑟𝑣𝑖𝑐𝑒 𝑆𝑡𝑟𝑢𝑐𝑡𝑢𝑟𝑒
!#1%
. Revenue structure is a vector of the
share of property taxes to locally collected own source revenue, share of income elastic revenue
sources to own-source revenue, and the share of intergovernmental transfers to total revenue.
91
Service structure is a vector of the share of employee compensation to own source revenue, pension
obligations to own-source revenue, spending on capital infrastructure to own-source revenue , and
debt servicing to own-source revenue. 𝑋
!#1%
includes a time-invariant institutional variable on city
incorporation (charter vs. general law) and lagged control variables for population, median
household income, unemployment rate, percent of city population that are registered as Democrat,
and percent of renter occupied units. 𝜀
!,#1%
is the lagged idiosyncratic error term. The empirical
model is estimated by clustering standard errors at the county-level. Revenue and expenditure are
inflation adjusted to 2018 constant dollars.
III D. Summary Statistics
Table 15 summarizes the number of cities that recovered in each fiscal year
following the two recessions – the dotcom recession in 2001 (top panel) and the Great Recession
in 2008 (bottom panel).
Table 15. Summary Statistics: Fiscal Recovery and Renewal
𝑡
%&'
𝑡
%&(
𝑡
%&)
𝑡
%&*
𝑡
%&+
Dotcom Recession
Revenue Recovery 88 (425) 90 (433) 92 (440) 93 (445) 95 (456)
Service Recovery 91 (435) 91 (438) 93 (446) 93 (445) 95 (455)
Fiscal Renewal 27 (130) 46 (220) 47 (226) 58 (278) 50 (241)
Great Recession
Revenue Recovery 64 (307) 58 (278) 67 (321) 71 (342) 76 (365)
Service Recovery 82 (395) 76 (363) 76 (368) 77 (370) 81 (387)
Fiscal Renewal 34 (163) 50 (242) 60 (287) 53 (254) 55 (263)
Note: This table shows the percentage (count in parentheses) of cities with revenue recovery, service recovery
and fiscal renewal from the five post-recession year (𝑡
"#$
to 𝑡
"#%
). Percentages are computed relative to the full
sample of 480 cities. The number shown in parentheses are counts for fiscal recovery and renewal.
92
Fiscal recovery after the dotcom recession was relatively steady, more cities demonstrated
recovery each year following the end of the crisis. Fiscal recovery after the Great Recession differs
for revenues and expenditures, while revenues recovery was steady, fewer cities indicate
consistency in recovery on the service side.
Table 16 presents summary statistics for the determinants of fiscal resilience. Cities in the
sample have a revenue diversification score is 0.77 with a standard deviation is 0.13 indicating a
wide variation in portfolio diversification. The overall level of liquidity is relatively low, the
average city has only 27% of its total own source revenue as financial reserves in its general fund
account.
Table 16. Summary Statistics
Mean SD Minimum Maximum
Fiscal Resistance
Buffering: Revenue Diversification 0.77 0.13 0.04 0.99
Boosting: Countercyclical Policy 0.27 0.38 0 3.63
Revenue Structure
Property Taxes 0.12 0.09 0.002 0.57
Income Elastic Revenue Sources 0.15 0.1 0 0.56
Intergovernmental Transfers 0.04 0.04 0.002 0.66
Service Structure
Employee Compensation 0.26 0.11 0 0.95
Pension Contributions 0.09 0.06 0 0.54
Capital Infrastructure 0.13 0.12 0 0.92
Debt Servicing 0.04 0.05 0 0.65
Institutional Context
Home Rule [Charter City = 1] 0.24 0 1
Control Variables
Population 58,579 190,388 188 3,763,424
Median Household Income 76,953 39,167 27,726 291,647
Percent Registered Democrats 0.43 0.11 0.16 0.82
Unemployment rate 0.08 0.046 0 0.45
Percent Renter Occupied Units 0.36 0.13 0.28 0.81
Note: This table shows summary statistics for fiscal covariates using data from three-year pre-recession periods for
both the dotcom and the Great Recession. Income elastic revenue sources include sales and transient occupancy taxes.
Intergovernmental transfers are a share of Median household income is inflation adjusted to 2018 constant dollars.
93
The standard deviation is nearly double as high as diversification indicating that financial
reserve capacities vary greatly between cities. The average city has 12% of its revenue portfolio
from property taxes, 15% from income elastic revenue sources, and 4% from intergovernmental
transfers. Among the service structure variables, employee compensation has the highest share of
total own source revenues (26%). Pension contributions are 9%, capital infrastructure is 13%, and
debt servicing is 4% of total local revenue. 24% of the cities in the sample are charter cities and
the remaining are general law cities.
Figure 6 presents fiscal trends for revenue per capita, expenditure per capita, and financial
reserves over the two recessions. Over the study period from 1991 – 2018, cities in the sample
witnessed substantive change in their revenue and service portfolios. The figures demonstrate that
the Great Recession which was more severe than the dotcom recession witnessed substantial
decline in revenues, expenditure, and reserves. In addition, reserves began to decline even before
500 1000 1500 2000
per capita
1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016
fiscal year
Expenditure
Revenue
.2 .22 .24 .26 .28 .3
reserves share of own-source revenue
1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016
fiscal year
Figure 6. Fiscal Trends: Revenue, Expenditure, and Reserves
Note: The figures show fiscal trends for the two recessions identified by dotted lines: for the dotcom
recession and 2008 for the Great Recession. Revenue and expenditure are in per capita. Financial
reserves are the unrestricted funds in the city’s general fund account as a share of own-source
revenue.
Revenue and Expenditure Financial Reserves
94
recession year and utilization of reserves, and this is observed for the Great Recession began in
2008. There is also, on average, a two-period lag between start of both the dotcom and the Great
Recession.
IV Findings and Analysis
Table 17 presents results from Cox proportional hazard models that estimate the
determinants of fiscal resilience. More specifically, the analysis identifies the factors that facilitate
fiscal recovery - revenue recovery and service recovery – and fiscal renewal. Columns (1), (2) and
(3) use the subsample for the Dotcom recession and Columns (4), (5), and (6) are estimated using
the subsample for the Great Recession.
4.1 The Dotcom Recession and Fiscal Resilience
Column (1) shows results for revenue recovery. We find that revenue diversification does
not have a statistically significant effect on revenue recovery indicating that the hypothesized
buffering effect of diversification did not facilitate revenue recovery from the Dotcom recession.
The second fiscal resistance variable - countercyclical policy - has a statistically significant effect
on revenue recovery but the finding shows that cities with higher reserves were less likely to
recover indicating that the hypothesized boosting effect did not facilitate fiscal resiliency during
the Dotcom recession.
95
Table 17. Determinants of Fiscal Resilience [Cox Proportional Hazard Model]
Dotcom Recession Great Recession
Revenue
Recovery
Service
Recovery
Fiscal
Renewal
Revenue
Recovery
Service
Recovery
Fiscal
Renewal
(1) (2) (3) (4) (5) (6)
Fiscal Resistance
Buffering:
Revenue Diversification
0.99
(0.02)
0.98
(0.02)
1.001
(0.003)
1.02
**
(0.01)
1.02
(0.02)
1.01
(0.01)
Boosting:
Countercyclical Capacity
0.32
**
(0.14)
4.24
(4.48)
0.75
**
(0.07)
1.03
(0.2)
1.98
(0.87)
0.89
(0.09)
Revenue Structure
Property Taxes 1.07
***
(0.02)
0.99
(0.05)
1.01
*
(0.01)
1.004
(0.01)
0.99
(0.01)
0.99
(0.01)
Income Elastic
Revenue Sources
1.01
(0.02)
0.947
(0.03)
1.001
(0.003)
0.99
(0.01)
0.99
(0.03)
0.99
(0.01)
Intergovernmental transfers 0.97
(0.04)
0.85
**
(0.05)
1.013
(0.01)
0.95
(0.04)
0.93
(0.05)
0.96
(0.03)
Service Structure
Employee compensation 0.99
(0.02)
1.03
(0.03)
1.002
(0.002)
1.01
(0.01)
0.98
(0.02)
1.005
(0.004)
Pension contributions 1.11
*
(0.04)
1.12
*
(0.05)
1.001
(0.01)
0.99
(0.02)
1.01
(0.03)
1.00
(0.01)
Capital investment 1.01
(0.01)
0.98
(0.02)
1.01
***
(0.002)
0.99
(0.006)
1.01
(0.01)
1.02
***
(0.002)
Debt servicing 0.96
(0.04)
1.07
(0.06)
1.01
**
(0.004)
0.99
(0.02)
1.04
*
(0.02)
1.01
(0.006)
Institutional Context
Home rule
1.33
(0.46)
0.29
**
(0.14)
1.05
(0.08)
1.07
(0.24)
0.67
(0.22)
1.13
(0.11)
Control Variables
Population 1.08
(0.12)
1.44
*
(0.24)
0.95
(0.02)
0.87
(0.07)
0.79
*
(0.07)
0.98
(0.03)
Median household income 0.92
(0.9)
1.11
(0.9)
1.03
(0.2)
2.94
*
(1.3)
2.51
(2.14)
1.11
(0.17)
Registered democrats
0.99
(0.02)
0.96
(0.02)
1.001
(0.003)
1.001
(0.01)
1.01
(0.02)
1.004
(0.01)
Unemployment rate 1.08
(0.06)
1.02
(0.07)
1.02
(0.02)
1.04
(0.03)
0.97
(0.05)
1.009
(0.02)
Renter occupied units
0.98
(0.02)
1.05
(0.03)
0.99
(0.003)
1.03
**
(0.01)
1.036
(0.03)
0.99
(0.004)
Observations 195 180 623 627 309 443
Note: Coefficients are hazard ratios. All variables are lagged one year. Population and median household income are log
transformed. Standard errors are clustered at the county-level.
***
p<0.001,
**
p<0.01,
*
p<0.05.
96
Among the revenue structure variables, dependence on property taxes has a statistically
significant effect on revenue recovery. This finding supports the hypothesized positive effect of
fiscal stability from dependence on property taxes on fiscal resilience. Among the service structure
variables, the results show that the share of pension contributions facilitate revenue recovery which
suggests that despite limited expenditure flexibility, budgetary spending on local pensions do not
hinder fiscal resilience. The remaining revenue and service structure variables are not statistically
significant. In addition, whether a city is incorporated under home rule does not affect fiscal
resilience, the control variables do not have a statistically significant effect on revenue recovery.
Column (2) presents results for service recovery. The effect of the fiscal resistance
variables – revenue diversification and countercyclical policy – on service recovery is not
statistically significant. Among the revenue structure variables, we find that the hypothesized
negative effect of dependence on intergovernmental transfers on service recovery is supported.
Among the service structure variables, a higher share of pension contributions is associated with
service recovery, the remaining revenue and service structure variables are not statistically
significant. Cities incorporated as home rule demonstrate lower service recovery. City size has a
statistically significant effect on fiscal resilience. Larger cities are more likely to achieve fiscal
resilience through increase service expenditure or by the means of service recovery.
Column (3) presents results for fiscal renewal. Among the fiscal resistance variables,
countercyclical policy has a statistically significant effect on fiscal renewal. However, contrary to
the hypothesized relation, cities with higher fiscal reserves were less likely to demonstrate long-
term fiscal renewal during the Dotcom recession. This can be explained if cities with low fiscal
capacity are more likely to accumulate higher reserves. Because fiscal renewal is a function of
achieving balance between revenues and expenditure, if reserves are expended to increase
97
expenditure in the absence of a higher or equivalent increase in revenue, reserves will fail to
facilitate fiscal renewal and resiliency. Apart from fiscal resistance, dependence on property taxes
facilitates fiscal renewal indicating that the benefits of fiscal stabilization from property tax
dependence increases the propensity for short-term fiscal recovery as well as long-term fiscal
renewal. Capital investment and debt servicing also have a significant positive effect on fiscal
renewal which suggests economic growth contributes to fiscal resiliency. The remaining variables
are not statistically significant.
Overall, we find that during the Dotcom recession, the factors that facilitated fiscal
resilience included dependence on property taxes, the share of pension contributions, higher capital
investment and higher debt servicing as well as city size. On the other hand, the factors that
impeded fiscal resilience were higher share of reserves, higher dependence on intergovernmental
transfers, and home rule incorporation. This pattern of findings suggests that fiscal resiliency from
the Dotcom recession was a function of revenue (property tax stabilization) and expenditure
(capital investments) specific factors as opposed to strategic decisions associated with revenue
diversification and countercyclical capacity.
4.2 The Great Recession and Fiscal Resilience
Column (4) presents results for revenue recovery. Findings support the hypothesized
buffering effect of diversified revenue portfolios indicating that diversification decreases revenue
shortfalls that are an inevitable outcome of economic recessions. The results also show that
countercyclical capacity has a positive impact on revenue recovery, although the coefficient is not
statistically significant. None of the revenue structure and service structure variables are
statistically significant. Among the control variables, the findings indicate that cities with higher
98
median household income or fiscal capacity and those with a higher share of cities with a higher
share of renter occupied units were more likely to demonstrate revenue recovery.
Column (5) shows findings for service recovery. The fiscal resistance variables - revenue
diversification and countercyclical capacity - have a positive effect on service recovery but the
effect is not statistically significant. We find that the share of spending on debt servicing which
proxies for economic investments is associated with a higher likelihood of fiscal resilience.
Surprisingly, the findings indicate larger cities were less likely to demonstrate service recovery.
These results can be explained to the extent that cities responded to the Great Recession by
ensuring that contractually mandated spending on debt servicing was uninterrupted despite
revenue shortfalls that would have resulted from the recession.
The determinants of fiscal renewal are presented in column (6). We find that capital
investments have significant positive effect on fiscal renewal. This finding supports the
hypothesized enabling effect of economic investments in local infrastructure on long-term fiscal
renewal. Because investments in local infrastructure create the conditions for revenue growth,
cities with higher spending on capital investments are more likely to demonstrate fiscal renewal.
The remaining variables are not statistically significant.
The pattern of findings for fiscal resilience from the Great Recession provide stronger
support for strategic decisions associated with revenue diversification and countercyclical capacity
as opposed to revenue or expenditure specific features of local budgets. However, the clear finding
is that local economic investments in capital infrastructure has a significant positive impact on
fiscal resilience - short-term recovery as well as long-term renewal. In addition, given that the
Great Recession was a deeper recession that the earlier Dotcom recession, the results point to the
salience of local fiscal capacity which is proxied by median household income in facilitating fiscal
99
resilience. Taken together, fiscal resilience from the Great Recession was driven by strategic
decisions that buffered against revenue loss but also economic investments that strengthened fiscal
capacity.
V Concluding Discussion
The recent prolonged recession triggered by Covid-19 led to a nearly 40% decline in state
and local revenues in addition to significant increases in demand for local public services Gordon
et al., 2020. Although resilience is an increasingly important issue in responding to such
recessionary pressures, there is limited attention to conceptually rigorous frameworks that
operationalize the full scope of resilience analysis. This paper advances the theoretical and
empirical scholarship on fiscal resilience. We introduce a new conceptual framework, formally
operationalize its distinct components, and provide rigorous empirical analysis to predict the extent
to which strategic and structural factors predict fiscal resilience. The analyses uses data from two
major recessions – the Dotcom Recession in 2001 and the Great Recession in 2008. Estimates
from Cox Proportional Hazard models help identify the disaggregated impact of pre-crisis fiscal
resistance, revenue structure, and service structure on local fiscal resiliency.
The findings present a nuanced picture on resilience enablers, fiscal resilience is a function
of local fiscal capacity, fiscal strategy, and recession-specific features. More specifically, the
findings indicate that revenue diversification is a resilience enabler even during a severe recession
such as the Great Recession. The impact of countercyclical spending, however, is more muted
during a severe recession and stronger for milder recessions. Property taxes have a stabilizing
effect and contribute to fiscal resilience but the effect is determined by recessionary characteristics
that affect property values. We consistently find that city investments in building local economic
100
capacity through capital infrastructure has a positive impact on recovery and renewal from both
mild and severe crises. These findings indicate that cities with higher investments in local
infrastructure will demonstrate a quicker recovery from Covid-19 pandemic compared to those
that did not have such projects in the pipeline. In addition, even though cities may have
accumulated financial reserves before the pandemic, fiscal resilience is contingent on expending
reserves to boost local economic activity.
The concept of resilience has gained increasing attention from policymakers, practitioners,
and academics in recent years as a result of recurring economic recessions. Fiscal resilience is
particularly vital to the design of resilient public organizations that can withstand disruptions and
recover quickly to ensure stable provision of local public services. Resiliency enables public
organizations to less likely to be perturbed from a long-run growth path despite disruptions that
vary in frequency, intensity, and duration by building internal capacity to absorb demand spikes
and revenue cliffs. More importantly, the ability of local governments to develop the capacity for
resilience to particularly important to the well-being of low-income households that are more
dependent on public services to navigate recessionary periods. The findings from this study show
that designing resilient public organizations requires a combination of strategic decisions
associated with local revenue portfolios and consistent investments in local infrastructure. Fiscal
resilience can be achieved by embedding fiscal resilience management in strategic planning.
Further research is required to understand how the broad array of institutional factors that
govern local public financial management affect fiscal resilience. Research is also needed to
examine the relationship between differences between cities on service provisioning arrangements
and its potential impact on fiscal resilience and the extent to which market-based vs. in-house
provisioning mechanisms translate to resilience enablers.
101
REFERENCES
Abott, C. (2015). Public Sector Unions, Partisanship, and Pensions in the US States. Unpublished
manuscript.
Abott, C., & Singla, A. (2021). Helping or Hurting? The Efficacy of Municipal Bankruptcy. Public
Administration Review, 81(3), 428-445.
Alm, J., & Sjoquist, D. L. (2014). State government revenue recovery from the great recession.
State and Local Government Review, 46(3), 164-172.
Anderson, J. E., & Shimul, S. N. (2018). State and local property, income, and sales tax elasticity:
Estimates from dynamic heterogeneous panels. National Tax Journal, 71(3), 521-546.
Andonov, A., Hochberg, Y. V., & Rauh, J. D. (2018). Political representation and governance:
Evidence from the investment decisions of public pension funds. The Journal of Finance,
73(5), 2041-2086.
Ansell, C., Sørensen, E., & Torfing, J. (2020). The COVID-19 pandemic as a game changer for
public administration and leadership? The need for robust governance responses to
turbulent problems. Public Management Review, 1-12.
Ansell, C., & Trondal, J. (2018). Governing turbulence: An organizational-institutional agenda.
Perspectives on Public Management and Governance, 1(1), 43-57.
Anzia, S. F., & Moe, T. M. (2015). Public sector unions and the costs of government. the Journal
of Politics, 77(1), 114-127.
Anzia, S. F., & Moe, T. M. (2017). Polarization and policy: The politics of public‐sector pensions.
Legislative Studies Quarterly, 42(1), 33-62.
Anzia, S. F., & Moe, T. M. (2019). Interest groups on the inside: The governance of public pension
funds. Perspectives on Politics, 17(4), 1059-1078.
Aubry, J.-P., Chen, A., & Munnell, A. H. (2017). A first look at alternative investments and public
pensions. State and Local Pension Plans, 55.
Aubry, J.-P., & Crawford, C. V. (2017). State and local pension reform since the financial crisis.
Center for Retirement Research at Boston College: State and Local Pension Plans, 54.
Barbera, C., Jones, M., Korac, S., Saliterer, I., & Steccolini, I. (2021). Local government strategies
in the face of shocks and crises: the role of anticipatory capacities and financial
vulnerability. International Review of Administrative Sciences, 87(1), 154-170.
Bartle, J. R. (1996). Coping with cutbacks: City response to aid cuts in New York State. State &
Local Government Review, 38-48.
102
Berkes, F. (2007). Understanding uncertainty and reducing vulnerability: lessons from resilience
thinking. Natural hazards, 41(2), 283-295.
Berne, R., & Stiefel, L. (1993). Cutback budgeting: The long‐term consequences. Journal of Policy
Analysis and Management, 12(4), 664-684.
Bjørnholt, B., Bækgaard, M., & Houlberg, K. (2016). Does fiscal austerity affect political decision-
makers’ use and perception of performance information? Public Performance &
Management Review, 39(3), 560-580.
Boin, A., Kofman, C., Kuilman, J., Kuipers, S., & van Witteloostuijn, A. (2017). Does
organizational adaptation really matter? How mission change affects the survival of US
federal independent agencies, 1933–2011. Governance, 30(4), 663-686.
Boin, A., & Lodge, M. (2016). Designing resilient institutions for transboundary crisis
management: A time for public administration. Public Administration, 94(2), 289-298.
Boin, A., & Van Eeten, M. J. (2013). The resilient organization. Public Management Review,
15(3), 429-445.
Bonsall IV, S. B., Comprix, J., & Muller III, K. A. (2019). State pension accounting estimates and
strong public unions. Contemporary Accounting Research, 36(3), 1299-1336.
Boyne, G. A., & Meier, K. J. (2009). Environmental turbulence, organizational stability, and public
service performance. Administration & Society, 40(8), 799-824.
Bozeman, B. (1981). Organization design in the public bureaucracy. The American Review of
Public Administration, 15(2), 107-118.
Bozeman, B. (2010). Hard lessons from hard times: Reconsidering and reorienting the “managing
decline” literature. Public Administration Review, 70(4), 557-563.
Bozeman, B., & Slusher, E. A. (1979). Scarcity and environmental stress in public organizations:
A conjectural essay. Administration & Society, 11(3), 335-355.
Brien, S. T., Eger III, R. J., & Matkin, D. S. (2021). The Timing of Managerial Responses to Fiscal
Stress. Public Administration Review, 81(3), 414-427.
Brooks, J. (2019). Board on the job: public-pension governance in the United States (US) states.
Journal of Public Policy, 39(1), 1-34.
Brunner, E. J., & Ju, A. (2019). State collective bargaining laws and public-sector pay. ILR Review,
72(2), 480-508.
Calabrese, T. (2010). Public Pensions, Public Budgets, and the Risks of Pension Obligation
Bonds.Working Paper
103
Carroll, D. A. (2005). Are state governments prepared for fiscal crises? A look at revenue
diversification during the 1990s. Public Finance Review, 33(5), 603-633.
Cellini, S. R., Ferreira, F., & Rothstein, J. (2010). The value of school facility investments:
Evidence from a dynamic regression discontinuity design. The Quarterly Journal of
Economics, 125(1), 215-261.
Cepiku, D., Mussari, R., & Giordano, F. (2016). Local governments managing austerity:
Approaches, determinants and impact. Public Administration, 94(1), 223-243.
Chaney, B. A., Copley, P. A., & Stone, M. S. (2002). The effect of fiscal stress and balanced
budget requirements on the funding and measurement of state pension obligations. Journal
of Accounting and Public Policy, 21(4-5), 287-313.
Chen, G., Kriz, K., & Ebdon, C. (2015). The effect of board composition on public sector pension
funding. Journal of Public Budgeting, Accounting & Financial Management.
Chernick, H. (1998). Fiscal capacity in New York: The city versus the region. National Tax
Journal, 51(3), 531-540.
Chernick, H., Copeland, D., & Reschovsky, A. (2020). The fiscal effects of the COVID-19
pandemic on cities: An initial assessment. National Tax Journal, 73(3), 699-732.
Chernick, H., Langley, A., & Reschovsky, A. (2011). The impact of the Great Recession and the
housing crisis on the financing of America's largest cities. Regional Science and Urban
Economics, 41(4), 372-381.
Chernick, H., & Reschovsky, A. (2017). The fiscal condition of US cities: Revenues, expenditures,
and the “Great Recession”. Journal of Urban Affairs, 39(4), 488-505.
Clemens, J., & Veuger, S. (2020). Implications of the Covid-19 pandemic for state government tax
revenues (0898-2937), Working Paper.
Coggburn, J. D., & Kearney, R. C. (2010). Trouble keeping promises? An analysis of underfunding
in state retiree benefits. Public Administration Review, 70(1), 97-108.
Collins, B. K. G., B. G. (2008). Taken for Granted? Managing for Social Equity in Grant Programs.
Public Administration Review, 68(6), 1128-1141.
Comfort, L. K., Sungu, Y., Johnson, D., & Dunn, M. (2001). Complex systems in crisis:
Anticipation and resilience in dynamic environments. Journal of contingencies and crisis
management, 9(3), 144-158.
Crowley, G. R., & Beaulier, S. A. (2018). Public-sector unions and government policy:
Reexamining the effects of political contributions and collective bargaining rights. Public
Finance Review, 46(3), 454-485.
104
Dipoppa, G., & Grossman, G. (2020). The effect of election proximity on government
responsiveness and citizens’ participation: Evidence from English local elections. Comparative
Political Studies, 53(14), 2183-2212.
Dippel, C., & Sauers, Z. (2019). Does increased union power cause pension underfunding in the
public sector, Working Paper.
DiSalvo, D., & Kucik, J. (2018). Unions, parties, and the politics of state government legacy cost.
Policy Studies Journal, 46(3), 573-597.
Downs, G. W., & Rocke, D. M. (1984). Theories of budgetary decisionmaking and revenue
decline. Policy Sciences, 16(4), 329-347.
Duchek, S. (2020). Organizational resilience: a capability-based conceptualization. Business
Research, 13(1), 215-246.
Duit, A. (2016). Resilience thinking: Lessons for public administration. Public Administration,
94(2), 364-380.
Dye, R. F., & Gordon, T. (2012). Pension legacy costs and local government finances. Land Lines,
24(4).
Eaton, T. V., & Nofsinger, J. R. (2004). The effect of financial constraints and political pressure
on the management of public pension plans. Journal of Accounting and Public Policy,
23(3), 161-189.
Fisher, R. C., & Wassmer, R. W. (2015). An analysis of state–local government capital expenditure
during the 2000s. Public Budgeting & Finance, 35(1), 3-28.
Frandsen, B. R. (2016). The effects of collective bargaining rights on public employee
compensation: Evidence from teachers, firefighters, and police. ILR Review, 69(1), 84-112.
Frandsen, B. R., & Webb, M. (2017). Public employee pensions and collective bargaining rights:
Evidence from state and local government finances. Hutchins Center Working Papers.
Freeman, R. B., & Han, E. (2012). The war against public sector collective bargaining in the US.
Journal of Industrial Relations, 54(3), 386-408.
Freeman, R. B., & Valletta, R. G. (1987). The Effect of Public Sector Labor Laws on Collective
Bargaining, Wages, and Employment. NBER Working Paper(w2284).
Glaeser, E. L., & Ponzetto, G. A. (2014). Shrouded costs of government: The political economy
of state and local public pensions. Journal of Public Economics, 116, 89-105.
Glover, L., & Granberg, M. (2021). The Politics of Maladaptation. Climate, 9(5), 69.
Goldstein, R., Sances, M. W., & You, H. Y. (2020). Exploitative revenues, law enforcement, and
the quality of government service. Urban Affairs Review, 56(1), 5-31.
105
Gordon, T., Dadayan, L., & Rueben, K. (2020). State and local government finances in the
COVID-19 era. National Tax Journal, 73(3), 733-758.
Gore, A. K. (2015). Do governments hide resources from unions? The influence of public sector
unions on financial reporting choices. The Influence of Public Sector Unions on Financial
Reporting Choices (November 4, 2015).
Gorina, E. (2018). City revenue structure and unfunded pension liabilities. State and Local
Government Review, 50(3), 189-202.
Gorina, E., & Hoang, T. (2020). Pension Reforms and Public Sector Turnover. Journal of Public
Administration Research and Theory, 30(1), 96-112.
Gorina, E., Maher, C., & Park, S. (2019). Toward a theory of fiscal slack. Public Budgeting &
Finance, 39(4), 48-74.
Gramlich, E. M. (1976). The New York City fiscal crisis: what happened and what is to be done?
The American Economic Review, 66(2), 415-429.
Green, D., & Loualiche, E. (2021). State and local government employment in the COVID-19
crisis. Journal of Public Economics, 193, 104321.
Greenhalgh, L., & McKersie, R. B. (1980). Cost-effectiveness of alternative strategies for cut-back
management. Public Administration Review, 575-584.
Guo, H., & Neshkova, M. (2018). Fiscal severity and the choice of budget gap closing strategies.
Public Money & Management, 38(4), 305-314.
Han, Y., & Goetz, S. J. (2015). The economic resilience of US counties during the great recession.
Review of Regional Studies, 45(2), 131-149.
Hanak, E. (2009). State infrastructure spending and the federal stimulus package. National Tax
Journal, 62(3), 573-583.
Hendrick, R. (2002). Revenue diversification: Fiscal illusion or flexible financial management.
Public Budgeting & Finance, 22(4), 52-72.
Hendrick, R. (2006). The role of slack in local government finances. Public Budgeting & Finance,
26(1), 14-46.
Hendrick, R., & Degnan, R. P. (2020). In the Shadow of State Government: Changes in Municipal
Spending After Two Recessions. The American Review of Public Administration, 50(2),
161-175.
Hoang, T., & Goodman, D. (2018). Public pensions and collective bargaining rights: Evidence
from state and local governments. Public Administration Review, 78(5), 772-784.
106
Hoene, C., & Pagano, M. (2003). States decrease their aid to cities. American City & County,
118(11), 6-6.
Holling, C. S. (1973). Resilience and stability of ecological systems. Annual review of ecology and
systematics, 4(1), 1-23.
Holling, C. S. (2001). Understanding the complexity of economic, ecological, and social systems.
Ecosystems, 4(5), 390-405.
Hou, Y., & Moynihan, D. P. (2008). The case for countercyclical fiscal capacity. Journal of Public
Administration Research and Theory, 18(1), 139-159.
Hrebiniak, L. G., & Joyce, W. F. (1985). Organizational adaptation: Strategic choice and
environmental determinism. Administrative science quarterly, 336-349.
Hsin, P.-L., Mitchell, O. S., Schieber, S., & Shoven, J. (1997). Managing public-sector pensions.
Public policy toward pensions, 36(8), 247.
Jimenez, B. S. (2014). Smart cuts?: Strategic planning, performance management and budget
cutting in US cities during the great recession. Journal of Public Budgeting, Accounting &
Financial Management.
Jimenez, B. S. (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), 161-177.
Jimenez, B. S. (2021). Applying the Loss-Conflict Model of Fiscal Retrenchment: Understanding
City Expenditure and Revenue Responses to a Budget Crisis. Public Performance &
Management Review, 1-29.
Joassart-Marcelli, P. M., Musso, J. A., & Wolch, J. R. (2005). Fiscal consequences of concentrated
poverty in a metropolitan region. Annals of the Association of American Geographers,
95(2), 336-356.
Jordan, M. M., Yan, W., & Hooshmand, S. (2017). The role of State revenue structure in the
occurrence and magnitude of negative revenue variance. The American Review of Public
Administration, 47(4), 469-478.
Justice, J. B., & Yang, K. (2018). Local-Government Responses to the 2008 US Fiscal Crisis:
Strategic, Merely Predictable, or Lifted from the Garbage Can? Public Finance &
Management, 18(1).
Kelley, D. G. (2014). The political economy of unfunded public pension liabilities. Public Choice,
158(1-2), 21-38.
Kilgour, J. G. (2013). Public pension reform in California. Compensation & Benefits Review,
45(6), 350-356.
107
Kim, C., & Park, S. (2021). When Push Comes to Shove: The Effect of Economic Crisis on the
Spending of Government Savings. International Journal of Public Administration, 1-14.
Kim, J., Maher, C. S., & Lee, J. (2018). Performance information use and severe cutback decisions
during a period of fiscal crisis. Public Money & Management, 38(4), 289-296.
Kim, Y. (2019). Limits of property taxes and charges: City revenue structures after the great
recession. Urban Affairs Review, 55(1), 185-209.
Kim, Y., & Chen, G. (2020). Cutback management and path dependency: evidence from the two
recent recessions. Local Government Studies, 46(2), 278-305.
Klasa, S., Maxwell, W. F., & Ortiz-Molina, H. (2009). The strategic use of corporate cash holdings
in collective bargaining with labor unions. Journal of financial economics, 92(3), 421-442.
Klein, P. G., Mahoney, J. T., McGahan, A. M., & Pitelis, C. N. (2019). Organizational governance
adaptation: Who is in, who is out, and who gets what. Academy of Management Review,
44(1), 6-27.
Kogan, V., Lavertu, S., & Peskowitz, Z. (2017). Direct democracy and administrative disruption.
Journal of Public Administration Research and Theory, 27(3), 381-399.
Kuipers, B. S., Higgs, M., Kickert, W., Tummers, L., Grandia, J., & Van Der Voet, J. (2014). The
Management of Change in Public Organizations: A Literature Review. Public
Administration, 92(1), 1-20.
Ladd, H. F., & Bradbury, K. L. (1988). City taxes and property tax bases. National Tax Journal,
41(4), 503-523.
Lavertu, S. (2016). We all need help:“Big data” and the mismeasure of public administration.
Public Administration Review, 76(6), 864-872.
Lawrence, P. R., & Lorsch, J. W. (1967). Differentiation and integration in complex organizations.
Administrative science quarterly, 1-47.
Lee, S., & Chen, G. (2021). Understanding financial resilience from a resource-based view:
Evidence from US state governments. Public Management Review, 1-24.
Levine, C. H. (1978). Organizational decline and cutback management. Public Administration
Review, 38(4), 316-325.
Levine, C. H., Rubin, I. S., & Wolohojian, G. G. (1981). Resource scarcity and the reform model:
The management of retrenchment in Cincinnati and Oakland. Public Administration
Review, 619-628.
Lichten, E. (1980). The fiscal crisis of New York City and the development of austerity. Insurgent
Sociologist, 9(2-3), 75-92.
108
Liu, C., Mikesell, J., & Moldogaziev, T. T. (2021). Public Corruption and Pension Underfunding
in the American States. The American Review of Public Administration,
0275074021992891.
Lutz, B., & Sheiner, L. (2014). The fiscal stress arising from state and local retiree health
obligations. Journal of Health Economics, 38, 130-146.
Maher, C. S., & Deller, S. C. (2007). Municipal responses to fiscal stress. Intl Journal of Public
Administration, 30(12-14), 1549-1572.
Makowsky, M. D., Stratmann, T., & Tabarrok, A. (2019). To serve and collect: the fiscal and racial
determinants of law enforcement. The Journal of Legal Studies, 48(1), 189-216.
Markowitz, H. M. (1976). Markowitz revisited. Financial Analysts Journal, 32(5), 47-52.
Marlowe, J. (2005). Fiscal slack and counter‐cyclical expenditure stabilization: A first look at the
local level. Public Budgeting & Finance, 25(3), 48-72.
Martell, C. R., Kioko, S. N., & Moldogaziev, T. (2013). Impact of unfunded pension obligations
on credit quality of state governments. Public Budgeting & Finance, 33(3), 24-54.
Matkin, D. S., & Krivosheyev, A. Y. (2013). Recognizing and responding to retirement
obligations: Other postemployment benefits in Florida cities and counties. The American
Review of Public Administration, 43(5), 558-580.
Matsusaka, J. G. (2018). Public policy and the initiative and referendum: a survey with some new
evidence. Public Choice, 174(1), 107-143.
McCubbins, C. H., & McCubbins, M. D. (2009). Proposition 13 and the California Fiscal Shell
Game. USC CLEO Research Paper(C10-16), 10-19.
McDonald III, B. D. (2018). Local governance and the issue of fiscal health. State and Local
Government Review, 50(1), 46-55.
Miller, R. A., Klouda, N., & Fisk, J. M. (2021). Concrete evidence: Infrastructure challenges and
the COVID-19 pandemic. Public Works Management & Policy, 26(1), 19-25.
Moynihan, D. P. H., Daniel P. (2012). Responsiveness to Reform Values: The Influence of the
Environment on Performance Information Use. Public Administration Review, 72(51), 95-
105.
Munnell, A. H., & Aubry, J.-P. (2017). An Overview of the State and Local Government
Pension/OPEB Landscape. The Journal of Retirement, 5(1), 117-137.
Munnell, A. H., Aubry, J.-P., & Cafarelli, M. (2015). How did state/local plans become
underfunded. Center for Retirement Research at Boston College, 42.
109
Munnell, A. H., Aubry, J.-P., Hurwitz, J., & Quinby, L. (2011). Unions and public pension
benefits. Center for State and Local Excellence.
Munnell, A. H., Haverstick, K., Aubry, J.-P., & Sass, S. A. (2008). The miracle of funding by state
and local pension plans. Citeseer.
Munnell, A. H., Hou, W., & Sanzenbacher, G. T. (2018). Trends in retirement security by
race/ethnicity. Center for Retirement Research at Boston College: Chestnut Hill, MA.
Musgrave, R. A. (1971). Economics of fiscal federalism. Nebraska Journal of Economics and
Business, 3-13.
Nelson, K. L. (2012). Municipal choices during a recession: Bounded rationality and innovation.
State and Local Government Review, 44(1_suppl), 44S-63S.
Novy‐Marx, R., & Rauh, J. (2011). Public pension promises: how big are they and what are they
worth? The Journal of Finance, 66(4), 1211-1249.
O'Toole, L. J., & Meier, K. J. (2010). In Defense of Bureaucracy:Public managerial capacity, slack
and the dampening of environmental shocks. Public Management Review, 12(3), 341-361.
Pacewicz, J., & N Robinson, J. (2020). Pocketbook policing: How race shapes municipal reliance
on punitive fines and fees in the Chicago suburbs. Socio-Economic Review.
Pagano, M. A., Hoene, C. W., & McFarland, C. (2012). City fiscal conditions in 2012: National
league of cities research brief on America’s cities. Washington DC: National League of
Cities.
Peng, J. (2008). State public pension management over the business cycle. Journal of Public
Budgeting, Accounting & Financial Management.
Peterson, P. E. (1981). City limits. University of Chicago Press.
Peterson, P. E., & Rom, M. C. (2010). Welfare magnets: A new case for a national standard.
Brookings Institution Press.
Quinby, L. D., & Sanzenbacher, G. T. (2020). Do Pensions Matter for Recruiting State and Local
Workers? State and Local Government Review, 52(1), 6-17.
Raman, K. K., & Wilson, E. R. (1990). The debt equivalence of unfunded government pension
obligations. Journal of Accounting and Public Policy, 9(1), 37-56.
Raudla, R., Douglas, J. W., Savi, R., & Randma-Liiv, T. (2017). Fiscal crisis and expenditure cuts:
The influence of public management practices on cutback strategies in Europe. The
American Review of Public Administration, 47(3), 376-394.
[Record #747 is using a reference type undefined in this output style.]
110
Rauh, J. D. (2009). Are State Public Pensions Sustainable? Train Wreck: A Conference on
America’s Looming Fiscal Crisis,” sponsored by the Urban-Brookings Tax Policy Center
and the USC-Caltech Center for the Study of Law and Politics.(January 15, 2010),
Rauh, J. D. (2017). Hidden debt, hidden deficits: 2017 edition. Hoover Institution, 32.
Reschovsky, A. (1980). An evaluation of metropolitan area tax base sharing. National Tax Journal,
33(1), 55-66.
Riccucci, N. M. (2007). The changing face of public employee unionism. Review of Public
Personnel Administration, 27(1), 71-78.
Ross, J. M., & Mughan, S. (2016). The Effect of Fiscal Illusion on Public Sector Financial
Management. Public Finance Review, 46(4), 635-664.
Schmidt, J., & Groeneveld, S. (2021). Setting sail in a storm: leadership in times of cutbacks.
Public Management Review, 23(1), 112-134.
Scorsone, E. A., & Plerhoples, C. (2010). Fiscal Stress and Cutback Management Amongst State
and Local Governments. State and Local Government Review, 42(2), 176-187.
Shoag, D., Tuttle, C., & Veuger, S. (2019). Rules Versus Home Rule—Local Government
Responses to Negative Revenue Shocks. National Tax Journal, 72(3), 543-574.
Simon, H. A. (1955). A behavioral model of rational choice. The Quarterly Journal of Economics,
69(1), 99-118.
Singh, J. V., Tucker, D. J., & House, R. J. (1986). Organizational legitimacy and the liability of
newness. Administrative science quarterly, 171-193.
Singla, A., Kirschner, C., & Stone, S. B. (2020). Race, representation, and revenue: Reliance on
fines and forfeitures in city governments. Urban Affairs Review, 56(4), 1132-1167.
Skidmore, M., & Scorsone, E. (2011). Causes and consequences of fiscal stress in Michigan cities.
Regional Science and Urban Economics, 41(4), 360-371.
Splinter, D. (2017). State pension contributions and fiscal stress. Journal of Pension Economics &
Finance, 16(1), 65.
St. Clair, T. (2013). The impact of budget stabilization funds on state pension contributions. Public
Budgeting & Finance, 33(3), 55-74.
Stalebrink, O. J. (2014). Public pension funds and assumed rates of return: an empirical
examination of public sector defined benefit pension plans. The American Review of Public
Administration, 44(1), 92-111.
111
Stalebrink, O. J., & Donatella, P. (2021). Public pension governance and opportunistic accounting
choice: A politico-economic approach. The American Review of Public Administration,
51(3), 227-245.
Steindel, C. (2020). Public pension shortfalls and state economic growth: a preliminary
examination. Business Economics, 55(3), 138-149.
Su, M. (2019). Understanding the accumulation of local government savings: A dynamic analysis.
International Journal of Public Administration, 42(11), 893-903.
Su, M. (2020). Discretion in Traffic Stops: The Influence of Budget Cuts on Traffic Citations.
Public Administration Review.
Sundén, A. E., & Munnell, A. H. (1999). Investment practices of state and local pension funds:
implications for social security reform. Boston College Center for Retirement Research
Working Paper(1999-01).
Suyderhoud, J. P. (1994). State-local revenue diversification, balance, and fiscal performance.
Public Finance Quarterly, 22(2), 168-194.
Tang, S. Y., Callahan, R. F., & Pisano, M. (2014). Using common‐pool resource principles to
design local government fiscal sustainability. Public Administration Review, 74(6), 791-
803.
Thom, M. (2017). The drivers of public sector pension reform across the US states. The American
Review of Public Administration, 47(4), 431-442.
Thom, M., & Randazzo, A. (2015). Underfunding annual pension contributions: examining the
factors behind an ongoing fiscal phenomenon. State and Local Government Review, 47(1),
35-44.
Thompson, J. D. (1967). Organizations in Action: Social Science Bases of Administrative Theory.
1967. Transaction pub, 192.
Triest, R. K., & Zhao, B. (2014). The role of economic, fiscal, and financial shocks in the evolution
of public sector pension funding. Proceedings. Annual Conference on Taxation and
Minutes of the Annual Meeting of the National Tax Association,
Valletta, R. G. (1993). Union effects on municipal employment and wages: a longitudinal
approach. Journal of Labor Economics, 11(3), 545-574.
Wang, S., & Pagano, M. A. (2017). Cities and fiscal federalism in the trump era: A discussion.
State and Local Government Review, 49(3), 184-198.
Wang, W. (2015). The great recession and the use of fund balances in North Carolina counties. In
Local government budget stabilization (pp. 17-32). Springer.
112
Wang, W., & Hou, Y. (2012). Do local governments save and spend across budget cycles?
Evidence from North Carolina. The American Review of Public Administration, 42(2), 152-
169.
Warner, M. E., Aldag, A. M., & Kim, Y. (2020). Pragmatic Municipalism: U.S. Local Government
Responses to Fiscal Stress. Public Administration Review.
Weaver, T. (2017). Urban crisis: The genealogy of a concept. Urban studies, 54(9), 2039-2055.
Wei, W. (2020). Exploring local government fiscal slack in a political-budgetary-managerial
framework. Public Management Review, 1-23.
Weller, C. E., & Wenger, J. B. (2009). Prudent investors: the asset allocation of public pension
plans. Journal of Pension Economics & Finance, 8(4), 501-525.
West, J. P., & Davis, C. (1988). Administrative values and cutback politics in American local
government. Public Personnel Management, 17(2), 207-222.
Wolman, H., & Peterson, G. (1980). State and local government strategies for responding to fiscal
pressure. Tul. L. Rev., 55, 773.
Yang, T. S., & Mitchell, O. S. (2005). Public pension governance, funding, and performance: a
longitudinal appraisal.
Zax, J. S. (1989). Employment and local public sector unions. Industrial Relations: A Journal of
Economy and Society, 28(1), 21-31.
Zhang, F., & Maroulis, S. (2021). Experience is not enough: A dynamic explanation of the limited
adaptation to extreme weather events in public organizations. Global Environmental
Change, 70, 102358.
Zhang, F., Welch, E. W., & Miao, Q. (2018). Public organization adaptation to extreme events:
Mediating role of risk perception. Journal of Public Administration Research and Theory,
28(3), 371-387.
113
APPENDICES
I Appendix to Chapter 2
1.1 Full Model for Spillovers to Service Categories
Full Model for Spillovers from Unfunded Pensions to Service Categories
Public Safety Economic
Development
Redistribution
Pension Performance
Asset Shock 0.015
***
(0.002)
0.013
***
(0.002)
0.015
***
(0.003)
0.014
***
(0.003)
0.011
*
(0.004)
0.004
(0.004)
Pension Liabilities (UAAL) -0.0018
(0.002)
-0.016
***
(0.002)
0.0001
(0.003)
-0.016
***
(0.003)
0.0037
(0.003)
-0.004
(0.003)
Asset Shock ×UAAL 0.0024
***
(0.0002)
0.0022
***
(0.0003)
0.0019
***
(0.0003)
Institutional Factors
Charter city 0.028
***
(0.006)
0.028
**
(0.011)
0.053
***
(0.013)
Control Variables
Population -0.013
***
( (0.003)
-0.002
(0.005)
-0.026
***
(0.006)
Percent Registered
Democrat
0.12
***
(0.023)
0.18
***
(0.04)
0.17
*
(0.07)
Hispanic -0.099
***
(0.015)
-0.13
***
(0.02)
-0.15
***
(0.03)
African American -0.058
(0.068)
-0.31
**
(0.11)
-0.19
(0.12)
Unemployment -0.31
**
(0.11)
-0.32
*
(0.15)
0.32
(0.22)
Median household income 0.036
**
(0.012)
0.065
***
(0.02)
0.011
(0.02)
Housing tenure -0.14
***
(0.032)
-0.15
**
(0.04)
-0.18
**
(0.07)
Constant 1.77
***
(0.009)
1.56
***
(0.12)
1.63
***
(0.01)
0.98
***
(0.18)
1.77
***
(0.24)
Observations 15,539 15,539 15,535 15,535 15,523 15,523
R2 0.09 0.43 0.16 0.37 0.23 0.31
Note: Coefficients are log points for [t+1]. All models include fiscal year and pension reform fixed effects. Standard errors are clustered
at the plan sponsor level. Population and median household income are log transformed.
***
p<0.001,
**
p<0.01,
*
p<0.05
114
1.2 Full Model for Spillovers from Overfunded vs. Underfunded Pensions
Full Model for Spillovers from Overfunded and Underfunded Pensions
Public Services Public Safety Economic
Development
Redistribution
Over-
funded
Under-
funded
Over-
funded
Under-
funded
Over-
funded
Under-
funded
Over-
funded
Under-
funded
Pension Performance
Asset Shock 0.016
***
(0.002)
0.013
***
(0.002)
0.017
***
(0.002)
0.014
***
(0.002)
0.016
***
(0.003)
0.016
***
(0.003)
0.019
***
(0.004)
0.0071
(0.004)
Unfunded Liabilities -0.014
***
(0.002)
-0.016
***
(0.002)
-0.017
***
(0.003)
-0.0059
(0.004)
Asset Shock ×
Unfunded Liability
0.0018
***
(0.0002)
0.0022
***
(0.0002)
0.0021
***
(0.0003)
0.0017
***
(0.0003)
Institutional Factors
Charter city 0.039
***
(0.008)
0.027
***
(0.006)
0.035
***
(0.007)
0.027
***
(0.006)
0.043
***
(0.011)
0.028
**
(0.01)
0.074
***
(0.018
0.051
***
(0.014)
Control Variables
Population -0.011
**
(0.003)
-0.012
***
(0.003)
-0.011
***
(0.003)
-0.014
***
(0.003)
0.0041
(0.005)
-0.0022
(0.005)
-0.028
***
(0.007)
-0.025
***
(0.006)
Percent registered
Democrat
0.11
**
(0.03)
0.11
***
(0.02)
0.15
***
(0.03)
0.12
***
(0.02)
0.17
***
(0.05)
0.19
***
(0.04)
0.16
*
(0.08)
0.19
**
(0.07)
Hispanic -
0.096
***
(0.019)
-0.096
***
(0.014)
-0.1
***
(0.02)
-0.1
***
(0.01)
-0.16
***
(0.02)
-0.14
***
(0.03)
-0.15
***
(0.04)
-0.16
***
(0.03)
African American -0.14
*
(0.05)
-0.11
*
(0.05)
-0.062
(0.06)
-0.054
(0.06)
-0.35
***
(0.08)
-0.28
**
(0.11)
-0.16
(0.12)
-0.2
(0.12)
Unemployment -0.08
(0.09)
-0.077
(0.09)
-0.43
***
(0.11)
-0.29
**
(0.11)
-0.092
(0.15)
-0.39
*
(0.16)
0.36
(0.24)
0.22
(0.23)
Median household
income
0.013
(0.01)
0.029
**
(0.01)
0.025
*
(0.012)
0.038
**
(0.012)
0.063
***
(0.02)
0.064
***
(0.02)
-0.014
(0.02)
0.009
(0.02)
Housing tenure -0.089
**
(0.02)
-0.12
***
(0.02)
-0.11
***
(0.03)
-0.14
***
(0.03)
-0.16
**
(0.04)
-0.15
**
(0.04)
-0.16
*
(0.08)
-0.19
**
(0.07)
Constant 1.86
***
(0.11)
1.76
***
(0.1)
1.61
***
(0.13)
1.54
***
(0.13)
0.94
***
(0.19)
1.01
***
(0.18)
2.03
***
(0.29)
1.77
***
(0.2)
Observations 6,635 15,336 6,622 13,695 6,621 13,692 6,620 13,680
R2 0.31 0.42 0.41 0.43 0.25 0.38 0.15 0.32
Note: Coefficients are log points for [t+1]. Subsample of overfunded consists of those plans in which assets > liabilities. Subsample
of underfunded consists of those plans in which assets < liabilities. All models include fiscal year and pension reform fixed effects.
Standard errors are clustered at the plan sponsor level. Population and median household income are log transformed.
***
p<0.001,
**
p<0.01,
*
p<0.05
115
1. 3 Full Model for Spillovers to Pension Generosity
Full Model for Spillovers to Pension Generosity
Shock
Model
Robustness Shock +
Unfunded Model
Full Model
Pension Performance
Asset Shock 0.47
***
(0.02)
0.40
***
(0.013)
-0.054
*
(0.02)
-0.009
(0.02)
Unfunded Pensions -0.88
***
(0.03)
-0.80
***
(0.03)
Asset Shock × Unfunded
Pensions
0.077
***
(0.002)
0.068
***
(0.002)
Institutional Factors
Home Rule
0.0087
(0.019)
-0.025
(0.016)
Control Variables
Population 0.15
***
(0.01)
0.096
***
(0.009)
Percent Registered
Democrat
0.42
***
(0.08)
0.19
**
(0.07)
Percent Hispanic -0.33
***
(0.05)
-0.18
***
(0.04)
Percent African American -0.55
**
(0.17)
-0.36
**
(0.13)
Unemployment Rate -1.35
***
(0.37)
-0.79
**
(0.3)
Median Household Income 0.054
(0.03)
0.041
(0.02)
Housing Tenure -0.33
***
(0.09)
-0.22
**
(0.07)
Constant 6.77
***
(0.2)
5.66
***
(0.41)
12.7
***
(0.2)
11.11
***
(0.4)
Observations 22,136 22,136 16,959 16,959
R2 0.56 0.63 0.72 0.74
Note: Coefficients are log points for [t+1]. Asset shocks and unfunded pensions are per-member, inflation adjusted to
2018 constant dollars and log transformed. Population and median household income are log transformed. All models
include fiscal year and pension reform fixed effects. Standard errors are clustered at city level.
***
p<0.001,
**
p<0.01,
*
p<0.05.
116
1.4 Full Model for Mediating Effect of Unions on Pension Generosity
Mediating Effect of Unions on Pension Benefits
Shock
Model
Robustness Shock +
Unfunded Model
Robustness
Pension Performance
Asset Shock -0.044
*
(0.02)
0.012
(0.021)
-0.046
*
(0.02)
0.003
(0.02)
Pension Liabilities -0.66
***
(0.03)
-0.53
***
(0.03)
-0.74
***
(0.04)
-0.61
***
(0.04)
Asset Shock ×
Unfunded Liabilities
0.062
***
(0.003)
0.049
***
(0.003)
0.068
***
(0.003)
0.055
***
(0.003)
Public Employee Unions
Firefighter Union 0.19
***
(0.032)
0.16
***
(0.03)
0.89
***
(0.21)
0.82
***
(0.15)
Police Union 0.27
***
(0.03)
0.24
***
(0.02)
1.27
***
(0.13)
1.225***
(0.13)
General Employee Union -0.29
***
(0.02)
-0.35
***
(0.02)
-0.82
***
(0.09)
-0.66
***
(0.08)
Unfunded Pensions ×
Firefighter Union -0.075
***
(0.02)
-0.068
***
(0.01)
Police Union -0.11
***
(0.01)
-0.11
***
(0.01)
General Employee Union 0.062
***
(0.009)
0.038
***
(0.008)
Institutional Factors
Charter city 0.0012
(0.015)
-0.002
(0.01)
Control Variables
Population 0.11
***
(0.009)
0.097
***
(0.008)
Percent registered Democrat 0.20
**
(0.07)
0.19
**
(0.06)
Hispanic -0.17
***
(0.04)
-0.15
***
(0.04)
African American -0.26
(0.16)
-0.29
(0.15)
Unemployment -0.76
*
(0.29)
-0.56
*
(0.27)
Median household income 0.085
**
(0.03)
0.091
**
(0.027)
Housing tenure -0.12
(0.081)
-0.12
(0.07)
Constant 12.26
***
(0.23)
9.87
***
(0.41)
12.39
***
(0.27)
10.01
***
(0.42)
Observations 16,959 16,959 16,959 16,959
R2 0.79 0.83 0.82 0.85
Note: Coefficients are log points for [t+1]. All models include fiscal year and pension reform fixed
effects. Standard errors are clustered at the sponsor level. Population and median household income are
log transformed.
***
p<0.001,
**
p<0.01,
*
p<0.05
117
1.5 Robustness Tests: Spillovers to public services and pension generosity
Persistence Effects of Spillovers to Public Services and Pension Generosity
Public Services Pension Generosity
[t+2] [t+3] [t+4] [t+2] [t+3] [t+4]
Pension Performance
Asset Shock 0.011
***
(0.002)
0.011
***
(0.002)
0.0097
***
(0.002)
-0.056
***
(0.02)
-0.11
***
(0.02)
-0.15
***
(0.02)
Pension Liabilities -0.013
***
(0.002)
-0.013
***
(0.002)
-0.013
***
(0.002)
-0.82
***
(0.02)
-0.83
***
(0.03)
-0.84
***
(0.03)
Asset Shock ×
Unfunded Liability
0.0021
***
(0.0002)
0.0022
***
(0.0002)
0.0024
***
(0.0002)
0.071
***
(0.002)
0.074
***
(0.002)
0.076
***
(0.003)
Institutional Factors
Charter city 0.027
***
(0.006)
0.027
***
(0.006)
0.027
***
(0.006)
-0.028
(0.02)
-0.031
(0.02)
-0.032
(0.02)
Control Variables
Population -0.012
***
(0.003)
-0.012
***
(0.003)
-0.013
***
(0.003)
0.095
***
(0.009)
0.091
***
(0.009)
0.079
***
(0.009)
Percent registered
Democrat
0.11
***
(0.03)
0.10
***
(0.03)
0.098
***
(0.03)
0.21
**
(0.07)
0.21
**
(0.07)
0.18
*
(0.08)
Hispanic -0.094
***
(0.02)
-0.091
***
(0.02)
-0.086
***
(0.02)
-0.18
***
(0.04)
-0.17
***
(0.04)
-0.16
***
(0.05)
African American -0.12
*
(0.05)
-0.11
*
(0.05)
-0.11
*
(0.05)
-0.38
**
(0.14)
-0.38
**
(0.14)
-0.34
*
(0.14)
Unemployment -0.065
(0.09)
-0.075
(0.09)
-0.09
(0.09)
-0.74
*
(0.31)
-0.71
*
(0.32)
-0.71
*
(0.32)
Median household
income
0.026
*
(0.01)
0.024
*
(0.01)
0.021
*
(0.01)
0.058
(0.03)
0.065
*
(0.03)
0.071
*
(0.03)
Housing tenure -0.11
***
(0.03)
-0.10
***
(0.03)
-0.097
***
(0.03)
-0.23
**
(0.07)
-0.23
**
(0.07)
-0.24
**
(0.07)
Constant 1.77
***
(0.1)
1.79
***
(0.1)
1.83
***
(0.1)
11.38
***
(0.4)
11.77
***
(0.4)
12.22
***
(0.4)
Observations 15,531 13,918 12,306 15,201 13,512 11,850
R2 0.42 0.42 0.42 0.74 0.74 0.75
Note: Coefficients are in log points for [t+1]. All models include fiscal year and pension reform
fixed effects. Standard errors are clustered at sponsor level. Population and median household
income are log transformed.
***
p<0.001,
**
p<0.01,
*
p<0.05
118
II Appendix to Chapter 3
2.1 Diagnostics for Group Stability: Expanded Covariates
Group Stability: City Characteristics
Failed
Referenda
[Mean (SD)]
Approved
Referenda
[Mean (SD)]
Difference
[Failed -
Approved]
t-stat
(p-value)
Total Registered
0.48 (0.13) 0.48 (0.12) 0.0008 -0.12 (0.9)
Registered Democrats
0.21 (0.07) 0.21 (0.07) -0.0005 0.14 (0.8)
Population
122,624
(405,802)
173,967
482,370
-51,343
2.17 (0.02)
Percent White
0.66 (0.17) 0.65 (0.15) 0.003 -0.34 (0.6)
Percent Black
0.04 (0.06) 0.04 (0.06) -0.002 0.58 (0.5)
Percent Hispanic
0.31 (0.22) 0.31 (0.21) -0.004 0.34 (0.7)
Percent aged 18 years
and under
0.25 (0.06) 0.24 (0.05) 0.004 -1.43 (0.2)
Percent aged 18 to 64 0.61 (0.05) 0.62 (0.05) -0.005 1.91 (0.05)
Percent aged 65 years
and over
0.12 (0.06)
0.12 (0.05)
0.001
-0.36 (0.7)
Percent over 25 with no
high school
0.06 (0.04)
0.06 (0.05)
0.0002
-0.07 (0.9)
Percent over 25 with
high school degree
0.38 (0.08)
0.38 (0.08)
0.001
-0.32 (0.7)
Percent over 25 with
Bachelors’ education
0.2 (0.14)
0.2 (0.13)
-0.003
0.45 (0.6)
Unemployment rate 0.03 (0.01) 0.03 (0.01) -0.0001 1.09 (0.2)
Median household
income
75,397
(30,475)
75,386
(29,553)
10.86
-0.006 (0.9)
Median home value 488,632
(306,514)
492,204
(293,153)
-3,571
0.22 (0.8)
Housing tenure: percent
owner-occupied units
0.19 (0.13)
0.2 (0.15)
-0.01
1.4 (0.2)
Note: Covariate stability tests are estimated on a robust, bias-corrected, optimal bandwidth identified using the
rdbandwidth selection algorithm. Group means are change in treatment to year prior [t-1 to t]. Short-term solvency
is the general fund unrestricted balance as a share of expenditure. Budgetary solvency is the general fund assets
net of liabilities as a share of expenditure.
119
2.2 Full Model of Service Adaptation
Direct and Mediated Effects of Negative Revenue Shocks on Service Adaptation
Mediated Effects of Fiscal Adaptation
Negative Revenue
Shock
Professional
Management
Political
Pressure
Pension
Stress
Public
Services
Public
Safety
Public
Services
Public
Safety
Public
Services
Public
Safety
Public
Services
Public
Safety
(1) (2)
(3) (4)
(5) (6)
(7) (8)
4 years prior -0.013
(0.02)
-0.027
*
(0.01)
0.028
(0.02)
0.012
(0.01)
0.031
(0.02)
0.054
**
(0.02)
-0.015
(0.014)
-0.0006
(0.005)
3 years prior -0.01
(0.02)
-0.034
**
(0.01)
0.038
(0.02)
0.006
(0.01)
-0.0004
(0.02)
0.029
(0.022)
-0.013
(0.009)
-0.0022
(0.009)
2 years prior -0.0002
(0.02)
-0.014
(0.01)
0.02
(0.03)
0.025
(0.02)
-0.024
(0.02)
0.0047
(0.02)
-0.007
(0.009)
-0.002
(0.005)
1 year prior -0.009
(0.01)
-0.017
(0.009)
0.034
*
(0.02)
0.021
(0.013)
-0.0014
(0.02)
0.012
(0.01)
0.0057
(0.008)
0.0004
(0.004)
Base year - - - - - - - -
1 year later -0.078
(0.01)
-0.012
(0.01)
-0.027
(0.02)
0.014
(0.01)
0.036
*
(0.02)
0.0014
(0.01)
0.01
(0.01)
0.0068
(0.004)
2 years later -0.034
*
(0.02)
-0.025
*
(0.01)
-0.065
**
(0.02)
0.034
(0.02)
0.061
*
(0.02)
0.023
(0.02)
0.0036
(0.01)
0.0007
(0.006)
3 years later -0.045
**
(0.02)
-0.044
***
(0.01)
-0.023
(0.02)
0.013
(0.02)
0.047
*
(0.02)
0.018
(0.01)
0.014
(0.02)
0.005
(0.006)
4 years later -0.03
†
(0.02)
-0.03
*
(0.01)
0.031
(0.02)
0.022
(0.01)
0.027
(0.02)
0.044
*
(0.02)
0.0055
(0.01)
0.01
(0.005)
5 years later -0.004
(0.02)
-0.035
*
(0.01)
0.006
(0.02)
0.0067
(0.02)
0.032
(0.02)
0.078
***
(0.02)
0.021
(0.01)
0.007
(0.008)
Constant -0.015
*
(0.005)
-0.012
**
(0.004)
-0.012
(0.008)
-0.0053
(0.008)
-0.017
*
(0.006)
-0.017
***
(0.005)
-0.019
(0.02)
-0.017
(0.02)
Observations 23,587 23,605 23,500 23,520 23,587 23,605 10,734 10,758
R2 0.21 0.19 0.22 0.2 0.22 0.19 0.26 0.16
Note: Coefficients are in log points. Empirical models include fiscal year and election year fixed effects. Standard errors
are clustered at the city level.
***
p < 0.001
**
p < 0.01
*
p < 0.05
120
2.3 Full Model of Sourcing Adaptation
Direct and Mediated Effects of Negative Revenue Shocks on Sourcing Adaptation
Mediated Effects of Fiscal Adaptation
Negative Revenue
Shocks
Professional
Management
Political
Pressure
Pension
Stress
Fees Fines Fees Fines Fees Fines Fees Fines
(1) (2)
(3) (4)
(5) (6)
(7) (8)
4 years prior -0.014
(0.03)
-0.068
(0.03)
-0.005
(0.04)
0.15
(0.1)
-0.063
(0.03)
0.016
(0.05)
-0.025
(0.02)
-0.012
(0.02)
3 years prior -0.028
(0.03)
-0.099
*
(0.03)
0.016
(0.03)
0.12
(0.08)
-0.048
(0.03)
0.025
(0.04)
-0.025
(0.02)
0.014
(0.03)
2 years prior -0.01
(0.02)
-0.04
(0.03)
0.04
(0.03)
0.07
(0.06)
-0.067
*
(0.03)
0.0024
(0.04)
-0.004
(0.02)
0.014
(0.02)
1 year prior -0.003
(0.02)
-0.028
(0.03)
0.052
(0.04)
0.032
(0.02)
-0.02
(0.02)
0.016
(0.03)
0.008
(0.01)
0.028
(0.02)
Base Year - - - - - - - -
1 year later 0.025
(0.02)
-0.075
*
(0.03)
-0.051
(0.04)
-0.041
(0.04)
-0.016
(0.02)
0.044
(0.03)
0.003
(0.01)
0.023
(0.02)
2 years later -0.031
(0.02)
-0.11
**
(0.03)
-0.07
(0.04)
0.038
(0.07)
0.031
(0.03)
0.096
*
(0.04)
-0.003
(0.02)
0.033
*
(0.02)
3 years later 0.0037
(0.02)
-0.095
*
(0.04)
-0.044
(0.04)
-0.02
(0.06)
-0.005
(0.03)
0.091
(0.05)
-0.008
(0.02)
0.047
*
(0.02)
4 years later 0.006
(0.03)
-0.13
*
(0.05)
-0.019
(0.05)
-0.067
(0.07)
0.003
(0.03)
0.17
**
(0.06)
0.006
(0.02)
0.072
*
(0.03)
5 years later -0.021
(0.03)
-0.092
(0.05)
-0.013
(0.04)
-0.116
(0.06)
0.022
(0.03)
0.12
(0.06)
0.003
(0.02)
0.022
(0.02)
Constant -0.02
*
(0.01)
-0.03
**
(0.01)
-0.015
(0.01)
-0.078
(0.04)
-0.022
†
(0.01)
-0.035
**
(0.01)
0.012
(0.05)
-0.14
**
(0.05)
Observations 23,608 23,338 23,521 23,251 23,608 23,338 10,758 10,660
R2 0.46 0.15 0.46 0.15 0.46 0.15 0.59 0.09
Note: Coefficients are in log points. Empirical models include fiscal year and election year fixed effects. Standard errors
are clustered at the city level.
***
p < 0.001
**
p < 0.01
*
p < 0.05
121
2.4 Full Model of Solvency Adaptation
Direct and Mediated Effects of Negative Revenue Shocks on Solvency Adaptation
Mediated Effects of Fiscal Adaptation
Negative Revenue
Shock
Professional
Management
Political
Pressure
Pension
Stress
Short-
Term
Long-
Term
Short-
Term
Long-
Term
Short-
Term
Long-
Term
Short-
Term
Long-
Term
(1) (2)
(3) (4)
(5) (6)
(7) (8)
4 years prior 0.02
(0.01)
-0.0053
(0.01)
0.0077
(0.02)
-0.0022
(0.02)
-0.012
(0.017)
-0.013
(0.02)
-0.004
(0.008)
0.0031
(0.008)
3 years prior 0.0095
(0.01)
-0.0084
(0.01)
0.018
(0.02)
-0.00087
(0.02)
-0.022
(0.015)
-0.015
(0.015)
-0.014
(0.009)
-0.0072
(0.009)
2 years prior 0.0094
(0.01)
-0.0027
(0.01)
0.01
(0.02)
-0.00044
(0.022)
-0.011
(0.014)
-0.0038
(0.014)
-0.0062
(0.007)
-0.0018
(0.006)
1 year prior 0.012
(0.008)
-0.0074
(0.01)
0.013
(0.017)
-0.0039
(0.01)
-0.012
(0.01)
-0.00021
(0.0095)
-0.0046
(0.005)
0.00073
(0.0067)
Base Year - - - - - - - -
1 year later 0.0022
(0.01)
0.008
(0.01)
0.0069
(0.01)
0.019
†
(0.01)
-0.0065
(0.01)
-0.023
*
(0.01)
0.0002
(0.006)
-0.0017
(0.006)
2 years later -0.016
(0.01)
-0.007
(0.01)
-0.002
(0.01)
0.024
†
(0.01)
-0.011
(0.01)
-0.025
*
(0.01)
0.0059
(0.006)
-0.002
(0.006)
3 years later -0.013
(0.01)
-0.002
(0.01)
0.021
(0.02)
0.028
†
(0.01)
-0.02
(0.02)
-0.03
*
(0.01)
0.0064
(0.008)
0.0016
(0.008)
4 years later -0.012
(0.01)
-0.004
(0.01)
0.017
(0.02)
0.006
(0.01)
-0.007
(0.02)
-0.017
(0.01)
0.0024
(0.008)
0.0062
(0.007)
5 years later -0.007
(0.01)
-0.01
(0.02)
0.005
(0.02)
-0.001
(0.02)
-0.013
(0.02)
-0.02
(0.02)
-0.004
(0.01)
0.013
(0.0095)
Constant 0.084
***
(0.01)
0.16
***
(0.02)
0.09
***
(0.01)
0.17
***
(0.02)
0.092
***
(0.02)
0.17
***
(0.02)
0.066
*
(0.03)
0.17
***
(0.037)
Observations 23,466 23,551 23,379 23,464 23,466 23,551 10,636 10,724
R2 0.033 0.096 0.035 0.099 0.034 0.098 0.046 0.15
Note: Estimated models include fiscal year and election year fixed effects. Standard errors are clustered at the city level. Short-
term is short-term solvency measured as unrestricted balance in the general fund as a share of expenditure. Long-term solvency
is general fund assets less liabilities as a share of expenditure.
***
p < 0.001
**
p < 0.01
*
p < 0.05
122
2.5 Robustness Tests for Direct Effect of Negative Revenue Shocks
Robustness Test 1: Limited Bandwidth Models
Service Adaptation Sourcing Adaptation Solvency Adaptation
Public
Services
Public
Safety
Service
Fees
Regressive
Fines
Short-Term
Solvency
Long-Term
Solvency
1 year later
0.017
(0.02)
-0.0016
(0.01)
0.04
(0.03)
-0.053
(0.03)
-0.008
(0.01)
-0.013
(0.013)
2 year later -0.028
(0.02)
-0.0094
(0.01)
0.006
(0.04)
-0.13
**
(0.04)
-0.009
(0.01)
-0.014
(0.01)
3 years later -0.02
(0.02)
-0.044
*
(0.02)
0.0022
(0.04)
-0.13
*
(0.05)
-0.01
(0.02)
-0.025
(0.01)
4 years later -0.008
(0.02)
-0.026
(0.01)
0.027
(0.04)
-0.13
*
(0.06)
-0.01
(0.02)
-0.032
(0.02)
5 years later -0.0042
(0.02)
-0.046
(0.02)
0.089
(0.05)
-0.12
(0.06)
-0.004
(0.02)
-0.029
(0.02)
Constant -0.0057
(0.008)
-0.013
*
(0.006)
-0.032
(0.01)
-0.0047
(0.01)
0.07
***
(0.02)
0.13
***
(0.02)
N 15,594 15,612 15,401 15,614 15,530 15,567
Robustness Test 2: Placebo [Non-Tax Referenda]
Public
Services
Public
Safety
Service
Fees
Regressive
Fines
Short-Term
Solvency
Long-Term
Solvency
1 year later 0.003
(0.01)
-0.009
(0.01)
-0.02
(0.03)
0.09
*
(0.04)
0.008
(0.01)
-0.005
(0.01)
2 year later -0.02
(0.02)
-0.004
(0.01)
-0.03
(0.03)
0.04
(0.05)
0.002
(0.01)
0.01
(0.01)
3 years later -0.002
(0.02)
-0.008
(0.01)
-0.05
(0.03)
0.03
(0.05)
0.008
(0.02)
0.009
(0.01)
4 years later -0.003
(0.02)
-0.004
(0.02)
-0.04
(0.04)
0.07
(0.05)
0.005
(0.02)
0.003
(0.01)
5 years later -0.032
(0.02)
0.005
(0.02)
-0.05
(0.04)
0.04
(0.05)
0.018
(0.01)
0.02
(0.02)
Constant -0.0087
(0.01)
-0.004
(0.004)
0.019
(0.01)
-0.029
*
(0.01)
0.072
***
(0.01)
0.13
***
(0.02)
N 19,150
19,147 19,003 19,157 19,071 19,133
Note: Limited Bandwidth models use robust bias-corrected bandwidth selected using the rdrobust routine.
***
p < 0.001
**
p < 0.01
*
p < 0.05
123
III Appendix to Chapter 4
3.1 Tests of Baseline Proportional Hazard Assumptions
The proportional hazards assumption test employs the postestimation command phtest.
chi2 df p>chi2
Fiscal Recovery
Expenditure: Dotcom Recession 3.13 13 0.99
Revenue: Dotcom Recession 2.82 13 0.99
Expenditure: Great Recession 9.01 13 0.77
Revenue: Great Recession 10.15 13 0.68
Fiscal Renewal
Balanced Growth: Dotcom Recession 0.55 13 1.0
Balanced Growth: Great Recession 3.42 13 0.99
Note: Estimates are from global tests of the proportional hazard assumption.
3.2 Non-Parametric Kaplan-Meier Survival Estimates
0 .25 .5 .75 1
0 2 4 6
analysis time
95% CI Survivor function
Kaplan-Meier survival estimate
Hazard of Fiscal Renewal: Great Recession
0 .25 .5 .75 1
0 5 10 15
analysis time
95% CI Survivor function
Kaplan-Meier survival estimate
Hazard of Fiscal Renewal: Dotcom Recession
0 .25 .5 .75 1
0 2 4 6 8 10
analysis time
95% CI Survivor function
Kaplan-Meier survival estimate
Hazard of Revenue Recovery: Great Recession
0 .25 .5 .75 1
0 5 10 15
analysis time
95% CI Survivor function
Kaplan-Meier survival estimate
Hazard of Revenue Recovery: Dotcom Recession
0 .25 .5 .75 1
0 2 4 6 8 10
analysis time
95% CI Survivor function
Kaplan-Meier survival estimate
Hazard of Service Recovery: Great Recession
0 .25 .5 .75 1
0 5 10 15
analysis time
95% CI Survivor function
Kaplan-Meier survival estimate
Hazard of Service Recovery: Dotcom Recession
Abstract (if available)
Abstract
This dissertation presents three chapters on fiscal decision-making in local government. I draw on interdisciplinary perspectives from local public finance, urban economics, and political economy to propose new theoretical frameworks, and empirically evaluate the causal impact of fiscal crises on urban economic development. The dissertation is anchored by the conceptual framework of mitigation-adaptation-resilience, each chapter examines a distinct form of fiscal crisis facing American cities and the ways in which cities mitigate, adapt, and demonstrate resilience to fiscal crises.
In the first chapter, I investigate the crisis of public pension underfunding. In particular, the paper is a causal evaluation of the extent to which pension underfunding spills over to service provision and pension generosity and the fiscal circumstances under which spillovers (if any) are mitigated. Findings indicate significant spillovers to both services and pension generosity, redistributive crowd-out indicates pension underfunding impacts social equity and spillovers to pension generosity are less likely to be explained by union bargaining power. Although spillovers are mitigated by market performance, they persist over the long run. The second paper is a causal assessment of the mechanisms through which local governments adapt to negative revenue shocks. I exploit a natural experiment in which passage or failure of fiscal referenda cause local revenue shocks. The paper employs a dynamic regression discontinuity framework to estimate the Average Treatment Effect (ATE) of negative revenue shocks on spending, revenue substitution and fiscal solvency. I find that although the direct ATE is limited to strategic decline in fiscal spending, when mediated by political economic factors such as political pressure and fiscal stress, fiscal adaptation occurs through revenue substitution including increased reliance on regressive revenue sources as well as decline in fiscal solvency. The paper discusses integration of adaptive management strategies as essential to ensuring social equity in local public finance. The third paper, co-authored with Juliet Ann Musso and Matthew M. Young, is a theoretical and empirical examination of fiscal resilience to disruptions such as those caused by economic recessions. We introduce a new conceptual framework of fiscal resilience and tease out its determinants. The paper quantifies fiscal resilience, uses data from previous two recessions and estimates cox proportional hazard models to identify structural and strategic factors that enable fiscal resilience. We find that while strategic approaches such as revenue diversification and local countercyclical fiscal policies contribute to resilience, local fiscal structures have a stronger bearing on fiscal resilience over the short- and long-term. The paper recommends utilizing a resilience frame to assess fiscal preparedness to recessionary pressures.
Linked assets
University of Southern California Dissertations and Theses
Conceptually similar
PDF
Essays on fiscal outcomes of cities in California
PDF
Spatial and temporal expenditure-pricing equity of rail transit fare policies
PDF
Diverging pathways to citizenship and immigrant integration in the U.S.
PDF
Three essays on health, aging, and retirement
PDF
Urban consumer amenities and their accessibility
PDF
Essays on congestion, agglomeration, and urban spatial structure
PDF
Three essays on the political economy of federal spending
PDF
Technological innovation in public organizations
PDF
Financial crises and trade policy in developing countries
PDF
On planning, place, and race: how trauma imaginaries reveal communal trauma and impact collective well-being
PDF
A view from below: the development and role of organizational social capital in neighborhood regeneration in Los Angeles
PDF
Structure, agency, and the Kuznets Curve: observations and implications for sustainability planning in U.S. cities
PDF
Essays on the economics of cities
PDF
Essays on development economics
PDF
Curating gastronomy: restaurants and social media in the cultural economy
PDF
Essays on political economy and corruption
PDF
The role of leadership and collaboration as a catalyst for regional economic development: a grounded theory study
PDF
Resilient and equitable urbanism by design: insights from the collaborative process to reimagine the SF Bay Area
PDF
Essays in opioid use and abuse
PDF
Essays in climate change adaptation: role of market power in incentivizing adaptation behavior
Asset Metadata
Creator
Rao, Manita
(author)
Core Title
Essays on mitigation, adaptation, and resilience to urban fiscal crises
School
School of Policy, Planning and Development
Degree
Doctor of Philosophy
Degree Program
Public Policy and Management
Degree Conferral Date
2022-08
Publication Date
07/22/2022
Defense Date
04/22/2022
Publisher
University of Southern California
(original),
University of Southern California. Libraries
(digital)
Tag
OAI-PMH Harvest,Political Economy,public finance,urban economics
Format
application/pdf
(imt)
Language
English
Contributor
Electronically uploaded by the author
(provenance)
Advisor
Musso, Juliet Ann (
committee chair
), Matsusaka, John G. (
committee member
), Phlillips, Mark D. (
committee member
)
Creator Email
manitara@usc.edu,manitarao@gmail.com
Permanent Link (DOI)
https://doi.org/10.25549/usctheses-oUC111373708
Unique identifier
UC111373708
Legacy Identifier
etd-RaoManita-10900
Document Type
Dissertation
Format
application/pdf (imt)
Rights
Rao, Manita
Type
texts
Source
20220722-usctheses-batch-960
(batch),
University of Southern California
(contributing entity),
University of Southern California Dissertations and Theses
(collection)
Access Conditions
The author retains rights to his/her dissertation, thesis or other graduate work according to U.S. copyright law. Electronic access is being provided by the USC Libraries in agreement with the author, as the original true and official version of the work, but does not grant the reader permission to use the work if the desired use is covered by copyright. It is the author, as rights holder, who must provide use permission if such use is covered by copyright. The original signature page accompanying the original submission of the work to the USC Libraries is retained by the USC Libraries and a copy of it may be obtained by authorized requesters contacting the repository e-mail address given.
Repository Name
University of Southern California Digital Library
Repository Location
USC Digital Library, University of Southern California, University Park Campus MC 2810, 3434 South Grand Avenue, 2nd Floor, Los Angeles, California 90089-2810, USA
Repository Email
cisadmin@lib.usc.edu
Tags
public finance
urban economics