University of Southern California Dissertations and Theses

The risks and rewards of city management: how city managers evaluate the nature of the job and compensation

(USC Thesis Other)

The risks and rewards of city management: how city managers evaluate the nature of the job and compensation

PDF

Download a page range

Download transcript

Copy asset link

Request this asset

Transcript (if available)

Content i

THE RISKS AND REWARDS OF CITY MANAGEMENT:
HOW CITY MANAGERS EVALUATE THE NATURE OF THE JOB AND
COMPENSATION

by

Jennifer M. Connolly

______________________________________________________________

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 PHILOSPHY
(PUBLIC POLICY AND MANAGEMENT)

August 2014

Copyright 2014 Jennifer M. Connolly

ii

DEDICATION

To my mentor, Tony Bertelli, who encouraged me to think about the big questions, carefully
guided me along the path to becoming an academic, and who patiently helped me through the
long phase of dissertation writing. Had I listened to Tony’s advice to not live near the beach
while in graduate school, perhaps this dissertation would have been completed a bit sooner.

iii

LIST OF TABLES

Table 1. Sample Selection Analysis – Probit Model ………...……………………...…..77
Table 2. Summary Statistics………………………. ………...……………………...…..91
Table 3. Tobit Model of Severance Protection ………...…………………...……..…...115
Table 4. Tobit Model of Severance Protection w/ Personal Characteristics ..….......…..117
Table 5. Measurement of Electoral Competitiveness ………...…………...………..….130
Table 6. Probit Model of Political Protection from Risk ………...……………......…...134
Table 7. Probit Model of Political Protection from Risk w/ Personal Characteristics…136

iv

LIST OF FIGURES

Figure 1. Determining the Degree of Risk of a City Manager Position…...………...…..58
Figure 2. Distribution of Data Based on Severance Protection…………………….…..110
Figure 3. Predicted Severance for Four Hypothetical Managers ……….....…………...131
Figure 4. Dispersion of Data Based on Year……………………………....…….....…..147
Figure 5. Box Plots of Deficit Spending by Pre or Post-2008…………...……….....….147

v

ABSTRACT

While numerous public administration studies examine principal-agent problems, only a
narrow set of work investigates this relationship in the local context, and most of this work
focuses on ex post municipal level outcomes. In this study, I advance the literature by linking
municipal conditions to the ex ante contractual specification of principal-agent relationships
between council and manager. I develop a novel theory of risk shifting between principal and
agent at the local level and predict that political constraints and financial conditions drive the risk
shifting behavior of managers and councils.
I create a unique municipal-level data source containing economic, financial and political
conditions in cities and towns throughout the state of California as well as information on the
risk protection provisions granted to the city managers serving in these cities via their
employment contracts. Using this data, I empirically test the predictions of the risk-shifting
theory. The results of the empirical analysis offer substantial support for the theory of risk-
shifting between council and manager and suggests that city managers recognize potential career
risk when entering into a new employment arrangement and they seek to protect themselves and
their professional reputation through risk-shifting. As the career risk associated with a given
manager position increases, the manager shifts that risk back onto the council through either
severance protection or political protection. These results suggest that municipal economic,
financial, and political conditions do impact the degree of career risk facing an incoming city
manager, and as such, the manager seeks to shift that risk back onto the city council through
insurance against and protection from termination. This study advances our understanding of
principal-agent problems within local government and presents implications for future research.
vi

TABLE OF CONTENTS
DEDICATION …………………………………………………………………………….…...... ii
LIST OF TABLES ……………………………………………………………………….....….. iii
LIST OF FIGURES …………………………………………………………….…………...….. iv
ABSTRACT ………………………………………………………………………………...……v
CHAPTER 1: INTODUCTION …………………………………………………...…………….. 1
The Risky Nature of City Management …………………………………...………….... 11
How the Project Proceeds …………….…………………………………...………….... 17
CHAPTER 2: THE GROWTH AND NATURE OF CITY MANAGEMENT……………….... 19
The Historical Development of City Management …………………………...………... 20
The Duties and Responsibilities of City Managers …………………………...………....25
CHAPTER 3: LOCAL GOVERNMENT ADMINISTRATION…………….……………….... 30
Council-Manager v. Mayor-Council Cities…………………………………...………... 32
City Manager Career Concerns and Turnover ……………...………………...………... 35
Politics and Administration in Local Government ………...………………...…….…... 41
CHAPTER 4: THEORY AND PREDICTIONS……………………….…….……………….... 45
Agency Theory in Local Public Management ……………...………………...………... 49
Incentive Alignment in Private Sector Organizations ………...…………………...…... 55
Predictions.…..………………………………………………………………...………...57
Testable Hypotheses……………………………………………...…………...………... 61
CHAPTER 5: CITY MANAGER EMPLOYMENT CONTRACTS……….…….…….…….... 67
Risk-Shifting Between Principal and Agent..………..……………………...…...……... 68
California City Manager Employment Contracts.…………..………………...………... 73
Roles and Responsibilities……………….....………..……………………...…...……... 78
Performance Expectations……………………....…………..………………....………...79
Risk Protection………….……………….....………..……………………...…...……... 83
CHAPTER 6: EMPIRICAL ANALYSIS……………………...…...……….…….…….……....90
Measuring Risk Protection……………………...…………..………………...………... 90
Measuring Municipal Condition…………………..………..………………...…….…... 93
Political Constraints……...……………………...…………..………………...………... 97
Control Variables………...……………………...…………..………………...………..101
vii

Methods…………..……...……………………...…………..………………...……….. 109
Post-2008 Sample………...…………..………...…………..………………....………..111
Results…………….……...……………………...…………..………………...………..114
The Impact of Economic Condition on Severance Protection.…………………....……117
The Impact of Financial Condition on Severance Protection.………….................……120
The Impact of Political Conditions on Severance Protection.………….....………....…122
A Substantive Example of Predicted Risk Shifting.………..………......................……129
The Impact of Municipal Conditions on Political Protection.…………...…………..…132
CHAPTER 7: CONCLUSION…………...………………………...……….…….…….…...... 142
APPENDIX A………………..…………...………………………...……….…….…….…...... 147
BIBLIOGRAPHY………………………...………………………...……….…….…….…...... 148

1

CHAPTER 1: INTRODUCTION

In July of 2012, San Bernardino, a city of approximately 213,000 people just 65 miles
east of Los Angeles, filed for bankruptcy protection, the third city in California to do so in a one-
month period. In their bankruptcy filing, San Bernardino estimated that it owed more than $1
billion in liabilities to more than 10,000 creditors. Pension obligations alone, which had doubled
since 2006, were more than the city’s total annual revenues and one of the primary causes of the
financial trouble (Lovett 2013). At the time of the bankruptcy filing, San Bernardino owed more
than $50 million on a pension bond and $195 million to the California Public Employee
Retirement System (CalPERS) in unfunded pension obligations. The city also owed $61 million
in unfunded retiree healthcare and $40 million of workers compensation payments. In addition to
their pension obligations, the city’s annual expenditures had exceeded annual revenues. For
example, in Fiscal Year 2010-2011, the city had expenditures of approximately $255.6 million
but brought in only $248 million in revenues, resulting in a $7.6 million budget deficit, the city’s
second deficit of more than $7 million in the last three years. Having been hit particularly hard
by the economic downturn, falling property values, and unemployment approaching 20 percent,
San Bernardino recognized that it could no longer meet its obligations and filed for bankruptcy
(Hagen 2012).
As part of the bankruptcy filing, the San Bernardino city council adopted an emergency
three-month fiscal plan to suspend all debt payments, freeze hiring, and cease payments into the
retiree health plan. Amidst massive spending cuts, city officials also reduced the police force by
80 officers and froze the hiring of new officers. In the wake of this decision, San Bernardino,
having been previously recognized as a national example for the dramatic reduction of gang
2

violence
1
, saw its crime rate rise substantially (Lovett 2013, Willon and Sewell 2012). In 2012,
the number of murders committed in San Bernardino increased by 50 percent over the previous
year, and frustrated citizens descended upon city council meetings to voice their concerns
(Lovett 2013). Residents asked about the increase in crime and what could be done to address
slower police response rates. The most candid answer came from City Attorney Jim Penman who
told city residents, “lock your doors and load your guns” (Hagen 2012). Penman defended his
remarks the following day, stating, “the problems [in San Bernardino] are serious” (Hagen
2012). While it may appear that San Bernardino was caught in a perfect storm of increasing
debt and falling revenues as a result of economic conditions, it was a combination of not only
fiscal problems but also political turmoil within city hall that contributed to instability in San
Bernardino and the ultimate decision to file for bankruptcy.
Various circumstances can bring a local government to the point of financial insolvency,
but given the nature of local government finance, it often takes only one or two changes in
municipal circumstance for fiscal health to be significantly impacted. For example, the
mismanagement of debt, such as over-borrowing based on inaccurate expenditure projections, or
a dramatic decrease in revenue as a result of a weakened economy can each severely harm the
financial health of a municipality. Local governments are required by law to balance their
annual budgets, meaning that a fragile or shrinking tax base, the primary source of revenue for
municipalities, can spell disaster for local finances within just a few budget cycles. However,
political conditions, which play an important role in the council’s attitude toward the means of

1
In early 2009, President Obama and Vice President Biden invited San Bernardino Mayor Pat Morris to
the White House to discuss emerging strategies for the control of gang violence based on the
overwhelming success of Morris’ crime-fighting strategy in the city, which was credited with reducing
the incidence of violent crime throughout the city (Morris 2009).
3

addressing financial problems, also impact the stability of a municipality. As Palmdale City
Manager David Childs said, “The financial turmoil in San Bernardino, while in many ways a
product of its own politics, does illustrate the devastating effect the economic downturn has had
on cities and the basic services they provide… I can really sympathize with them being on the
brink. One or two bad things can put a city over the edge” (Willon and Sewell 2012). In the
aftermath of the recession, San Bernardino was not the only municipality to experience
unprecedented increases in unemployment and steadily decreasing per capita income, making it
difficult to raise enough revenue to continue to pay creditors. However, San Bernardino was
unique because of the combination of fiscal stress stemming from economic conditions and a
turbulent political environment in which membership on the council changed frequently and
council members were largely unwilling to pass politically unpopular policies that would have
addressed the budget deficit (Willon and Sewell 2012).
The political turmoil in San Bernardino can be traced to at least 2009, when, a little more
than three years before filing for bankruptcy, the San Bernardino city council unanimously voted
to hire City Manager Charles McNeely, former city administrator in Reno, Nevada, to oversee
the day-to-day operations of the city (Edwards 2009a, b, c). When McNeely started his position
on June 1
st
of 2009, San Bernardino was considering the reduction of staffing, wages, and service
provision in the face of falling tax revenues (Edwards 2009a, b, c). The decline in tax revenue
was linked to falling property tax values and a high citywide unemployment rate of 17.9
percent
2
. According to Mayor Pat Morris, McNeely was brought in to bring an air of professional

2
Based on US Bureau of Labor Statistics, San Bernardino’s unemployment rate reached 19.6 percent in
2010, the highest unemployment rate in the city on record, based on data first available in 1990. When
McNeely took office in June of 2009, the unemployment rate stood at 17.9 percent and had increased in
each of the twelve months prior.
4

management to the city whose elected officials had been unable to reach consensus about how to
address their budget deficit (Hagen 2012). McNeely immediately examined the municipal budget
and presented financial projections to the council. He described to council the significance of the
declining revenue base, resulting from increasing unemployment and falling property values and
warned them that if they maintained the status quo with regard to fiscal policy not only was
bankruptcy possible in the next several years, it was, in his opinion, the most likely scenario
(Willon and Sewell 2012). However, rather than implementing any of McNeely’s suggested
policy proposals, the council accused him of using the threat of bankruptcy as a scare tactic
designed to persuade the council to raise taxes or outsource police and/or fire protection (Willon
and Sewell 2012). The council repeatedly voted against such proposals to reduce expenditures
after vocal objections from the electorate and public safety workers (Edwards 2009, Simondes
2009).
McNeely was outspoken about the negative impact the political climate in the city had on
his efforts to improve fiscal health, stating that disagreements between council members and an
unwillingness of the council to impose any policy measures that might be at all unpopular with
voters meant very little action was taken to address the budget deficit (Willon and Sewell 2012).
In fact, despite McNeely and his staff presenting the council with numerous options to balance
the budget, the council continuously failed to pass tax increases and other proposals to raise
revenue, citing their desire to appease voters who were already facing financial hurdles because
of the recession (Willon and Sewell 2012). However, the council’s unwillingness to take action
to address the budget deficit left McNeely feeling as though his policy recommendations,
grounded in his professional training and experience in city management, were largely ignored
for political reasons (Willon and Sewell 2012).
5

By most accounts, the political climate in San Bernardino from 2009 to 2012 was
competitive at best and divisive and counterproductive at worst (Willon and Sewell 2012,
Edwards 2009, Simondes 2009). The city held two council elections in 2009, one special election
in March
3
, and one regularly scheduled election for three of the seven city council seats in
November. In the March special election for a single council seat all four candidates were
newcomers to local politics
4
and the primary issues of salience during the campaign were public
safety spending and economic development (Edwards 2009). Pharmaceutical Sales
Representative Fred Shorett won the special election on a platform focused on “streamlining
public services and looking for efficiencies” (Edwards 2009). During the campaign period
leading up to the regular November election, the council contentiously debated several proposals
designed to increase revenues and/or decrease expenditures (Simondes 2009). During a May
2009 council meeting at which the councilors discussed a reduction in the municipal fire
department, local residents left the meeting “completely disenchanted and even disgusted with
the machinations that were brought forth by certain members of the city council and the city
attorney… [who] proceeded to use this problem as a campaign tool” (Simondes 2009). Not
surprisingly, the regular council election in November was competitive; only one of the three
incumbent council members successfully won re-election and the closest margin of victory in the
race was less than 1 percent (San Bernardino Sun 2009). The particularly competitive 2009
council elections could explain the refusal of city council members to make politically unpopular

3
The special election was held to select a replacement for former council member Neil Derry who
resigned in December of 2008 to join the San Bernardino County Board of Supervisors leaving a vacant
seat on the San Bernardino City Council.
4
The candidates included a University IT Manager, an Online Mattress Merchant, a Self-Employed Sales
Contractor, and a Pharmaceutical Representative (Edwards 2009).
6

policy decisions in the years leading up to the bankruptcy filing, despite the city manager’s
repeated warnings and recommendations to take action (Willon and Sewell 2012).
The political battle over how to address a growing budget deficit continued in San
Bernardino for the next several years. During a 2010 debate over the municipal budget, McNeely
reprimanded the council for their shortsighted approach to budgeting, lamenting that for 13 of
the previous 15 years, the city had begun the annual budgeting process with a deficit and
appeared likely to continue on the same path (Hagen 2013a). In a contentious 2010 budget
meeting, McNeely told the council, “my focus… is to constantly remind all of us that we need to
get out of that one-year crisis mode and come to a five-year plan” (Hagen 2013a). McNeely’s
pleas went largely ignored as the council modified his proposed budget to keep open a popular
community center and eliminate cuts to the fire department; in response, City Attorney Jim
Penman refused to sign the budget because it wasn’t balanced (Hagen 2013a). The council
passed the budget despite McNeely and Penman’s objections, and Penman responded, “Mr.
McNeely, they’ve just voted the city into bankruptcy, sir. I hoped you would exercise your
influence. This is not a threat. This is not hysterics. This vote is insanity” (Hagen 2013a).
McNeely, Penman, and the city council sustained their heated disagreements over how to
balance the budget for the next couple of years, during which time San Bernardino’s fiscal health
continued to decline (Hagen 2013b).
After serving three years, McNeely resigned from his post as city manager in 2012, and
in a resignation letter sent to city employees, blamed the turbulent politics within the city and a
“toxic and horrendous” work environment for his departure, referring to the difficulty of advising
and working with a council who frequently disregarded his policy recommendations for political
reasons (Hagen 2012). After his resignation, McNeely told the Los Angeles Times, “the politics
7

of that place are just impossible to deal with” (Willon and Sewell 2012). McNeely was not the
only senior official in the municipality to express concerns over political divisiveness and the
role of politics on council decision making. In the wake of McNeely’s resignation, Mayor Pat
Morris, who himself faced multiple challengers in the November 2009 election, spoke out about
the recent loss of numerous senior employees, including the Police Chief, Assistant City
Manager, and Public Works Director, directly citing their resignations as the result of divisive
city politics (Hagen 2012).
With a vacuum of professional leadership in the city, San Bernardino brought in interim
city manager Andrea Travis-Miller who quickly discovered the dire reality of the city’s financial
situation, which despite McNeely’s warnings had never been adequately addressed. By the time
Travis-Miller took office in 2012, bankruptcy was imminent, just as McNeely had predicted
several years earlier. Travis-Miller guided the city through the bankruptcy process, but in
January of 2013, less than a year after she took the position, Travis-Miller announced her
resignation, along with Finance Director Jason Simpson. City officials lamented the loss of both
Travis-Miller and Simpson, and worried about how their resignation would impact the city’s
bankruptcy proceedings (Hagen 2013c). With the departure of multiple senior officials, San
Bernardino was left without a professional city manager or finance director to oversee the city’s
bankruptcy and to answer questions from the court and creditors. Less than six months into their
Chapter 9 bankruptcy filing, amidst falling revenues, massive budget cuts and public safety costs
spiraling to more than 70 percent of the general fund budget, San Bernardino’s problems were
far from over, and the city was faced with recruiting and hiring a new city manager.
After a month long search, the city council voted unanimously to hire Allan Parker to
lead the insolvent city through its bankruptcy proceedings and to oversee the day-to-day
8

administration of the municipality. However, even Parker’s hiring did not come without
controversy. At the time of his hiring, Parker had 24 years of experience as a city manager, but
he had previously filed for personal bankruptcy twice (Vuong 2013). Concerned citizens once
again questioned the council’s logic, but city council member Virginia Marquez spoke out
defending the council’s decision to hire Parker. Marquez said, “As a governing body, we felt that
we had the best person for the job. It is not like we had hundreds of applicants knocking on our
door. Nobody wants to come to our city. They know we have no stability. I’m grateful Mr.
Parker will even work for our city” (Vuong 2013).
A confluence of events in San Bernardino including high unemployment, falling
revenues, increasing crime, and spiraling debt all precipitated the city’s decision to file for
bankruptcy. But the financial conditions alone were not entirely responsible for the city’s spiral
into financial insolvency. The tumultuous political environment in the city strained the
relationship between council and manager and thwarted the manager’s attempts to lead the city
toward fiscal health. City managers are a class of professionals tasked with upholding
professional norms and improving financial conditions while working as an agent of the
politically elected council. As such, both fiscal stress and political instability can individually
and together increase the degree of career risk associated with a certain position and can
dramatically increase the probability that the city manager will fail to meet council expectations
or fail to improve municipal conditions.
The story of San Bernardino’s decline exemplifies that significant role that poor financial
conditions and divisive city politics can play in fostering an unappealing and risky work
environment for professional public managers. The career risk city managers face is directly
linked to their likelihood of success on the job; success in both upholding professional standards
9

and meeting the expectations of council. However, as the experience of Charles McNeely in San
Bernardino demonstrates, success in one area can mean failure in the other. McNeely’s attempts
to balance the budget, which were largely in accordance with professional norms, failed to
satisfy a city council more concerned with pleasing voters than with reducing the deficit. We
should expect any city manager taking on a new position to survey the municipal environment
for potential sources of career risk and to seek to shift that risk back onto the council.
As Charles McNeely experienced in San Bernardino, improving municipal financial
condition often requires making difficult and politically unpopular policy choices, and this is a
clear source of career risk for a professional who is hired and fired by the elected city council.
But when working for a city council that is concerned more with reelection than with financial
solvency, attempting to make the changes necessary to improve financial condition can be
extremely risky professionally. Making decisions that are unpopular with voters always presents
professional challenges for a city manager, but making those decisions within a tumultuous
political climate and under the supervision of a city council whose membership is likely to
change with any given municipal election presents even more career risk. As Steve Pinkerton,
former city manager of Stockton (one of the three California cities that filed for bankruptcy in
2012)
5
explained,
“I once had to cut 30% of a city’s workforce and reduce the salaries of the rest to address
a drop in revenue, and that wasn’t a pleasant experience to go through. Ever since the

5
Stockton, a city of nearly 300,000 residents just outside of San Francisco, filed for bankruptcy in July of
2012 after it was unable to close a $26 million budget deficit. In the case of Stockton, the weak economy
and high unemployment of 15.4 percent at its peak, were ruinous to the city’s finances. In 2011, the city
had general fund revenues of $180 million but general fund debt of nearly $250 million. After Moody’s
downgraded the city’s credit rating to Caa3, it projected that given the city’s current fiscal health,
bondholder losses would be greater than 20 percent.

10

great recession, if you’re doing your job as a city manager, you aren’t going to be well
loved by your employees or by the entire community” (Pinkerton 2014).
6

Financial or economic instability in a municipality places the professional city manager in a
difficult and risky position. Managing a city in a way that improves financial condition while
continuing to provide public services can be challenging in and of itself, but doing so while
operating within a political environment characterized by competitive council elections and/or
divisive council politics only increases the likelihood that council will not support the manager
when his proposals are politically unpopular or will ignore his policy proposals altogether.
Serving at the will of the council, the manager is an agent of the council, and indirectly of the
people. But serving as a professionally trained city manager, the manager also has an ethical and
professional duty to manage the city with certain positive outcomes in mind, and the means to
those outcomes may not be politically popular with citizens or with the council. While a city
with financial instability presents some career risk for a manager, a city with both financial
instability and a high degree of political constraint from the council generates an even higher
degree of career risk.
When considering taking on a position within a municipality, a professional city manager
must evaluate and consider this risk and attempt to protect himself and his professional
reputation from that risk. While there are numerous ways that a manager might seek to protect
himself from career risk after he has started in his new position, it is equally important to think
about how the city manager protects himself from career risk ex ante, before the first day on the

6
Five current California city managers were interviewed during the course of this project. I selected and
contacted 22 current California city managers based on their employment in cities of various degrees of
political and financial instability. Based on availability, I was able to schedule phone or in-person
interviews with five managers. Quotes from personal communications with these managers are included
throughout this dissertation. The managers interviewed throughout this project are not intended to serve
as a representative sample of managers.
11

job. As institutions influence the environment in which city managers operate, it is logical to
expect the political and financial structures in a municipality to shape the way city managers
assess the career risk associated with any given position. The career risk resulting from financial
and political instability impacts the way city managers think about taking on a new position and
the ways in which they protect themselves from that risk. Ex ante, city managers should seek to
shift career risk back onto the council, and they do this primarily through the terms of the
employment contract they negotiate with the council. Given the high value city managers must
place on maintaining their professional reputation and mobility, understanding the sources of
career risk for city managers and the ways in which they seek to protect themselves by shifting
some degree of that career risk back onto the city council ex ante has important implications for
the study of local governance and the institution of the city manager.

The Risky Nature of City Management
City managers, appointed public officials responsible for overseeing the administration of
the municipality and the implementation of the municipal budget, are a relatively mobile class of
professional public managers. The average city manager serves approximately six to seven years
per city (Watson and Hassett 2004) often changing positions numerous times throughout their
professional careers in pursuit of positions in larger organizations or with better pay (Feiock and
Stream 1998, DeHoog and Whitaker 1990, DeSantis and Renner 1994). A city manager, as the
highest serving official in the organization, cannot “move up” within their current organization;
the only opportunity for professional advancement is to “move on” to another organization.
Feiock and others (Feiock and Clingermayer 1997, Feiock 2004, Zhang and Feiock 2007) have
presented evidence that city managers see their performance in their current position as a means
12

to advancing in future positions. Francis, Feiock, and Kassekert (2009) have suggested that city
managers “pursue innovation and cost efficiencies seeking to be residual claimants on local
government performance” as a means of advancing their careers (6). In addition to considering
their performance as a means of advancing their professional careers, city managers must also
consider the institutional design of the position as a source of career advancement or risk. Almost
all city managers serve at the pleasure of the city council, an elected body whose membership
could change in any given election and who could decide to terminate the manager with a simple
majority vote. As such, city managers, with a vested interest in protecting their professional
reputation, must consider the nature of the city council as a source of career risk when taking on
any new position. Municipal fiscal condition and political condition both present a distinct
source of career risk for city managers. As Rodney Gould, the city manager of Santa Monica and
ICMA West Coast Regional Vice President, succinctly states, “Risk-averse people should not be
in the profession of city management” (Gould 2014).
Cities that are politically unstable and in poor financial condition, such as San
Bernardino, present the highest degree of career risk to city managers, but also offer the
opportunity for high reward. On the one hand, if the manager is successful in turning the city
around and improving its condition, the rewards, in terms of their professional reputation and
future employment opportunities, could be abundant. However, if the manager is unsuccessful in
improving the condition of an unstable city, the results could be disastrous as they face
termination by the council, and the potential for developing a reputation of inadequate
management skills or an inability to work with council.
The story of San Bernardino’s growing financial instability and increasing political
constraints illuminates a puzzle of interest to scholars of both political science and public
13

administration: How do elected councils and appointed managers seek to align their interests ex
ante? When an elected city council delegates managerial responsibility to a city manager this
creates a principal-agent problem in which the manager serves at the council’s discretion, yet the
council and the manager have different incentives. The most obvious source of this divergence
in incentives is the institutional nature of their positions. City managers are relatively well
compensated (based on the sample used in the analysis portion of this project, 95% of city
managers in the state of California earn $100,000 or more per year in salary). City managers
approach their positions as full-time professional management positions with a concern for
maintaining the professional standards of public management instilled through their graduate
education and membership in professional organizations such as the International City/County
Management Association (ICMA). Beyond upholding professional standards and ethics, city
managers seek to protect and improve their professional reputations.
On the other hand, city council members are popularly elected and are often community
members with no professional training in the administration of local government. In most small
and medium size cities, service on the council is treated as a part-time commitment, at least
based on the compensation provided to council members. In general, council service receives
modest compensation at best but requires a significant time commitment (Svara 2003). For
example, according to a study by Svara (2003), approximately 68 percent of cities in the United
States paid council members an annual salary of less than $10,000. This level of annual
compensation is even more common in council-manager cities, 76 percent of which paid council
members less than $10,000 annually (in 2001 dollars). However, the weekly time commitment
for council members resembles that of a part-time or full-time position. Council members in
small, medium, and large cities spent an average of 20, 25, and 42 hours respectively on council
14

related activities each week. This is a stark contrast to the mean salary of city managers, which in
California exceeds $175,000 per year (2012 dollars). Further exacerbating the different
backgrounds of council members and city managers is the divide in professional training
between council members and city managers. While a substantial majority of city managers hold
at least one graduate level degree
7
, only 40% of city council members hold a graduate level
degree (Svara 2003).
In general, city managers are well compensated, well educated, and are professionally
trained in financial management, public works, economic development, and the professional
ethics of public management. City managers have gained power and autonomy as reform ideas
and professionalization have become more widespread (Ammons and Newell 1989, Nalbandian
1989) but this has not been countered by a similar professionalization of city councils (Protasel
1988). Most council members do not come from the same educational background as city
managers, and they are meagerly compensated for their time on the council, which often requires
a significant time commitment. Most city council members are not professionally trained public
managers, but rather are community members with a variety of independent professions who
choose to serve on the council as a form of service to their communities or as a means to
influence local policy, but not as a result of any professional training in local government
management. This dynamic was evident in the case of San Bernardino where council members
frequently disagreed with the recommendations of the city manager and where many candidates
for council were new to local government.

7
According to data collected for the empirical analysis portion of this project, approximately 81 percent
of city managers in California hold a graduate level degree. See Chapter 7 for a full discussion of the
methods used to collect this data.
15

City council members must be popularly elected to their positions, and in California,
municipal elections for council are nonpartisan. So, while council members are typically not
trained in local government management and face electoral pressure from municipal voters, city
managers are professional public administrators trained in local government management and
motivated by their own career aspirations to maintain professional norms and to produce
outcomes that not only please their current council but that would also be seen as desirable by
other city councils as well. San Bernardino’s city manager was tasked with attempting to
improve the financial and economic stability of the community (with his success at this task
directly impacting his professional reputation), while the city council faced electoral pressure
from citizens concerned about the quantity and quality of municipal service provision.
While San Bernardino, one of only a handful of municipalities to file for bankruptcy in
the wake of the recession, faced a financial and political crisis more severe in nature than most,
the same dynamic of misaligned incentives and divergent professional standing exists between
city managers and councils in many council-manager cities throughout California and the United
States. Having to reduce staff and municipal salaries or cut back on municipal service provision
to balance the budget creates tension between city councils concerned with reelection and city
managers concerned with maintaining professional standards and improving the condition of the
municipality. In light of these divergent incentives, managers should seek to shift the career risk
associated with taking on a position in an unstable municipality back onto the elected members
of the city council. Local government managers are in the unique position of serving not one, but
typically five (or more) elected bosses who can change with any given election, and council
divisions or intense competition for council seats are sources of career risk for managers.
However, it is important to note that with regard to the shifting of risk ex ante, the council
16

functions as one legislative body who makes decisions based on majority vote, but nonetheless
must come to one agreed upon outcome in their negotiation with the manager. As such, when
thinking about the relationship between council and manager, while competition between
individual council members can create career risk, the manager must negotiate with the council
as a single body who acts as one after majority vote.
To address the important question of how municipal conditions impact the shifting of
career risk ex ante, I develop a theory of risk distribution between city manager and city council,
which predicts the manner in which managers shift risk back on to the council based on
economic, financial and political stability in the municipality. I introduce a new dataset drawn
from the employment contracts of California city managers and empirically examine the degree
of risk protection contracted for between city managers and municipal councils. This analysis
adds to our understanding of the institution of the city manager within the context of local
government. While other scholars have thoroughly examined the growth of the council-manager
form of government (Renner and DeSantis 1998), city manager career mobility and turnover
(McCabe et al 2008, Feiock and Stream 2002), the impact of the council-manager form on
service delivery and municipal expenditures (Benton 2002, Feiock and Tavarres 2002), and the
role of city managers in municipal policy making (Banovetz 1995, Nalbandian 1999, Morgan
and Watson 1992), none have theorized about or empirically tested the relationship between
municipal condition and risk shifting between council and manager. The manner in which risk is
shifted between manager and council can have wide reaching implications for the way council
and manager interact and the way in which local policies are implemented. By developing a
theory of risk distribution between manager and council and empirically testing the hypotheses
17

produced from this theory, I advance our understanding of the institution of the city manager and
the role of municipal condition on the structure of the relationship between manager and council.
While much of the scholarly work in public administration focuses on principal-agent
problems, a much narrower set of work investigates the principal-agent relationship between city
council and city manager, and most of this work focuses on municipal level outcomes after the
city manager takes office. Perhaps the largest gap in the literature on the relationship between
council and manager is the lack of theoretical or empirical work focused on the original decision
between the two parties to enter into a principal-agent relationship with the principal (city
council) delegating discretion and seeking to align their incentives with the agent (city manager).
Little research focuses on the employment contract between city manager and city council and
how the terms of that contract structure the relationship between manager and council and
protect the manager from the inherent career risk that varies from city to city. My goal in this
study is to expand on this work by linking municipal conditions to the contractual specification
of principal-agent relationships within local government management: what are the implications
of municipal instability on city manager terms of employment and protection from risk?

How The Project Proceeds
From here, this study proceeds as follows. In the next chapter, I discuss the growth and
nature of city management and the duties and responsibilities of city managers. In Chapter Three,
I examine common local government institutions, the importance of city managers, the major
differences between council-manager and mayor-council cities, the relationship between politics
and administration in local government, and finally city manager career concerns and turnover.
In Chapter Four, I lay the theoretical foundations upon which this project is built and begin a
18

discussion of agency theory and its application to local government management. I then turn to
the consideration of risk shifting between city managers and city councils. I conclude Chapter
Four by carefully outlining my theory of risk shifting in local government management and the
hypotheses I seek to test. In Chapter Five, I describe the original data collection process used to
collect and analyze city manager employment contracts from the state of California and map out
the typical features of these contracts. In Chapter Six, I discuss the data and methods of the
empirical portion of the project and the results of the empirical analysis. I conclude with Chapter
Seven, in which I discuss the implications of the results on our understanding of local
government management and map out directions for future research.

19

CHAPTER TWO: THE GROWTH AND NATURE OF CITY MANAGEMENT
The council-manager form of municipal government is a relatively new institution in
American local government, first introduced in the city of Staunton, Virginia in 1908 and slowly
but steadily adopted by numerous municipalities since (Stillman 1974). In order to understand
the risk shifting relationship between council and manager it is important to examine the
historical development of the council-manager form of government and the historical impetus
which shaped the design of the profession of city management. The council-manager form of
government was largely developed in response to what many saw as shortcomings in other forms
of municipal government, shortcomings that were the result of the elected executive model used
in most U.S cities (DeSantis and Renner 2002; Feiock, Jeong, and Kim 2003). Out of this
movement against elected executives and toward the professional management of municipal
service provision grew a commonly devised and generally accepted set of core duties and
responsibilities of city managers, which has been more formally codified by professional
associations (such as ICMA) and adopted by most municipalities employing a city manager and
formalized in their municipal codes (Hayes and Chang 1990, ICMA 2007). This chapter focuses
on following the historical development of the profession of city management from its early 20
th

century beginnings, through a rise in popularity in the mid-twentieth century and ending with an
eventual surpassing of the mayor-council form as the most popular form of municipal
government among American cities
8
. This chapter concludes with a discussion of the core duties

8
According to a survey and report issued by ICMA (2006), the council-manager form of government
surpassed 50 percent usage in US cities between 1996 and 2006, while Renner and DeSantis (1998) argue
that the council-manager form of government had become the most common form of local government in
the United States by 1996.
20

and responsibilities of today’s city managers that have developed along with the growth and
development of the profession.

The Historical Development of City Management
In order to understand the nature of risk shifting between a city manager and city council
and its relationship to a municipality’s fiscal, economic and political condition, a short
discussion of the historical development of the council-manager form of government and the
typical roles and expectations of city managers is warranted. Since their founding, many cities
and towns in the United States have operated under a strong-mayor or mayor-council form of
government – an institutional structure in which all policy making power and implementation
authority is vested in popularly elected officials. The growth in popularity of the council-
manager form of government was largely a result of the progressive movement away from
political leadership at the local level and toward more professional executive decision-making
and policy implementation. As Feiock, Jeong, and Kim (2003), write, “the Progressive ideology
of the separation of politics and administration, institutionalized in the council-manager plan,
allows administrators and elected officers to more easily resist opportunistic behavior” (616).
The shift from politically elected municipal executives to professionally appointed
managers is one of the most notable historical changes in municipal government in the last half-
century. In the 1960s and 1970s, one of the primary arguments made in support of a move to the
council-manager form of government was that council-manager cities are more efficient than
mayor-council cities because professional managers, seeking to capture efficiency gains and
bolster their career mobility are more likely to minimize the cost of producing municipal services
and are better able to negotiate personnel matters with a focus on holding down labor costs
21

(Booms 1966). Over the last five decades, local governments have trended toward increasing
professionalization in their management (Frederickson, Johnson and Wood 2003). The growth in
popularity of New Public Management (NPM) has shifted the values, organizational structure,
and approach to decision making in many local governments, especially those that have shifted
from a political management (mayor-council form) to professional management (council-
manager form). The movement towards the NPM model shifts the values of the municipality as
an organization from accountability and representation to cost-effectiveness and responsiveness
to citizens who are thought of more as customers than as voters. Additionally NPM promotes a
shift in organizational structure from pluralism to competitive and firm-like with a focus on
rational-scientific decision-making.
As a result of the trend toward professionalization, the belief that professional managers
will produce more efficient outcomes, and the popularity of new public management, many
municipalities adopted the council-manager form of government and now turn to the market for
professional city managers to hire top professionals to serve as the highest-ranking executive in
the municipal organization. In fact, the growth in the popularity of the council-manager form of
government has led some local government scholars to question whether Masters of Public
Administration programs, which train many such managers, will be able to keep up with demand
for professionals trained in municipal management (Gabris, Davis, Nelson 2012).
This movement toward professional local government management has been especially
pronounced in California, a state known for its progressive roots, where more than three-fourths
of all cities now operate under the council-manager form of government. Given this high rate of
adoption, a dichotomous focus on the differences between council-manager cities and mayor-
council cities becomes less relevant. Rather, scholars need to develop a deeper understanding of
22

the institutional structure and implications of council-manager government and the differences
between council-manager governments themselves. There is an open question in public
administration about the structure of the institution of the city manager and the resulting
incentives, which govern the behavior of the manager. Not all municipalities or city managers
are made equal, and it is important for scholars to move beyond research which seeks to
understand differences between mayors and managers, but rather seeks to understand differences
between managers themselves; differences that are the result of municipal condition and the ex
ante shifting of risk between manager and council.
There are numerous reasons why a municipality would want to hire an unelected,
professional manager to oversee the administration of the city. The position of the city manager
was originally seen as an institutional fix to widespread political corruption and inefficiency in
local government (DeSantis and Renner 2002). The key benefits of the council-manager form of
government are that the presence of both a council and an unelected manager “centralizes
supervisory and administrative responsibility in one individual, allowing the individual’s
expertise and knowledge of administrative activities to be developed while vesting all power in
an elected governing body to promote representative democracy” (DeSantis and Renner 2002).
In other words, the council-manager form of government seeks to hold government officials
accountable to the people, maintaining the core tenets of democracy, while also vesting a high
degree of rulemaking authority and implementation authority within a professionally trained
manager focused more on the efficient and effective provision of services than on pleasing
political factions.
In general, there has been a trend over the last 50 years away from the mayor-council
form of government and towards the council-manager form which comprised approximately
23

48% of municipalities in 1981 and approximately 60% of municipalities in 1996 (Renner and
DeSantis 1998). The council-manager form is especially prevalent in municipalities with a
population of fewer than 100,000 people (ICMA 2006). Even those cities that do not operate
with an appointed city manager answering to an elected city council are likely to have some sort
of chief appointed official or hybrid form of government, relying on an appointed administrator
for policy implementation. In a recent ICMA Municipal Form of Government Survey (2006),
85% of municipalities reported having a chief appointed official, defined as a city manager, chief
executive officer, city administrator, chief administrative officer, town administrator, village
manager, or employee of a similar title. While professional trained and appointed local
government managers in these various roles do not necessarily have the same authority or
responsibilities as a city manager, the increase in their existence suggests a larger move toward
including some sort of appointed public official in the high ranks of local government, even
when a different form of government is used.
9
While some cities use alternative forms of
government such as the commission structure or the town-meeting format, scholars agree that the
mayor-council and council-manager forms of government account for the large majority of cities
in the United States (DeSantis and Renner 2002, ICMA 1996).
With local governments increasingly adopting the council-manager form of government
or employing a professional in a similar capacity to the classic city manager, developing our
scholarly understanding of this institutional structure is increasingly crucial. For many years after

9
Another trend in local government is the adoption of a hybrid form of government placing both an
elected executive and an appointed executive in a position of shared managerial responsibility. This
arrangement is a relatively new development in local government and often varies in its implementation
from city to city. For example, Los Angeles operates under a strong-mayor form of government but due to
its large size also employs a Chief Administrative Officer who performs many of the same duties as a
typical city manager, but reports directly to the Mayor and the council. Due to the varied institutional
structures of cities employing administrative officers in a hybrid form of government, they are excluded
from the discussion of city management in the remainder of this analysis.
24

the initial growth in the popularity of the council-manager form of government, scholarly
research sought to examine the purported benefits of the council-manager form over the mayor-
council form, particularly in terms of local government efficiency in service provision. However,
as local governments continue to evolve toward a structure that vests power to manage municipal
finances, human resources, and policy advisement in the hands of the city manager, scholarly
research that focuses on the dichotomous distinction between the mayor-council and council-
manager form of government becomes less and less relevant. While the work of Feiock and
colleagues has deepened our scholarly understanding of how form of government (council-
manager versus mayor-council) impacts the formulation and implementation of local policy
(Feiock and Kim 2001; Feiock, Jeong, and Kim 2003; Lubell, Feiock, and De La Cruz 2009,
Feiock 2004) their work focuses on a dichotomous distinction between an elected local executive
and an appointed one, often presenting evidence that the council-manager form of government is
associated with different outcomes than the mayor-council. A great deal of work has advanced
our understanding of how managers impact municipal service delivery (Clingermayer and Feiock
1997), the relationship between managers and their staff (Nalbandian 2004), the nature of city
manager career paths (Feiock and Stream 1998, Watson and Hassett 2004) and how manager
turnover impacts local politics (McCabe et al 2008). Feiock, Jeong, and Kim (2003) have argued
that as a result of the separation of politics and administration seen in the council-manager plan,
managers face different incentives than mayors and this impacts the policies they implement
with regard to economic development. However, the trend over the last several decades of “a
sizeable portion of previously mayor-council cities adopting ‘city-manager-like’ positions”
(Meek, Schildt, & Witt, 2002, 146) has made a move toward a research agenda focused on
variations between individual city managers and the institutional design of city management
25

imperative. While the empirical research examining differences in local outcomes based on form
of government acknowledges the institutional relationship between councils and managers and
their divergent incentives, it does not seek to understand differences between managers
themselves but rather compares managers to mayors. While their contributions to our scholarly
understanding of local government are noteworthy, no local government scholars have looked at
the institutional risk shifting between council and manager and no systematic work has examined
the employment contracts of city managers. The contribution of this project is to advance the
study of city management by focusing on the manager as the unit of analysis and seeking to
understand the shifting of risk between manager and council as a function of municipal
condition.

The Duties and Responsibilities of City Managers
In order to refocus the local government research agenda on the institutional structure of
council-manager government, it is important to first understand the role of a typical city manager
in the municipal organization. City managers are responsible for the day-to-day administrative
operations of the locality and supervise all municipal departments in carrying out the policies
laid out by the city council. In many ways, city managers can be thought of as the public sector
equivalent of the CEO of a corporation. The typical duties of a city manager include:
 Oversee and direct the administrative operations of the municipality, including the
provision of services
 Prepare a budget for the council’s consideration
 Recruit, hire, terminate, and supervise government staff
26

 Serve as council’s chief policy advisor, particularly with regard to financial matters,
economic development, public works, labor issues, and service provision.
 Interpret and implement council policy
One of the most important responsibilities of a city manager is to oversee the budget proposal
and management process. The city manager is almost always responsible for preparing and
proposing the annual city budget. After the city or town council passes a budget, the city
manager is responsible for monitoring and implementing the city budget. The city manager’s role
in the budget process is especially important when one considers that most city council members
are members of the community serving in part-time positions, often without any education or
experience in financial management. When discussing the unique relationship between manager
and council, Steve Pinkerton said, “When you become city manager, you suddenly have five
bosses, none of whom are professional in your field, and that is quite an adjustment because you
are always having to make sure that no more than one or two people are mad at you. The nature
of the job is that you can’t make everyone happy” (Pinkerton 2014).
As the previous analogy between local government managers and Chief Executive
Officers suggests, the institutional structure of a council-manager city mimics that of private
business in several ways. The voters act as stockholders, the council as a board of directors, and
the city manager as a chief executive officer (Hayes and Chang 1990). It is widely accepted that
city councils seek to hire highly trained, professional managers on the basis of some combination
of education, experience, skill and ability (ICMA 2007). City managers are expected to perform
in an executive capacity over the entire municipality, which includes the management of
municipal finances. Just like a private sector CEO or manager, city managers are tasked with
maintaining the fiscal health and integrity of their municipality.
27

This discussion of the typical duties and professional responsibilities of city managers is
not intended to suggest that city councils do not also desire positive governance outcomes in
their municipalities. I argue that both city managers and city council members desire good
governance. However, they desire good governance, distinctively defined. The city council is an
elected body, and as such faces classic electoral pressures and has a vested interest in re-election.
On the other hand, city managers are appointed professionals trained to prioritize certain core
tenets of public serve, such as equity, effectiveness, efficiency, and fairness. City managers,
especially those trained in Masters of Public Administration programs, are trained to abide by a
professional code of ethics and to protect their professional reputation and maintain their career
mobility. Both city managers and city council members desire good governance within their
municipality, the puzzle for scholars of political science and public management arises in trying
to sort out how the council (principal) and the manager (agent) seek to align their divergent
incentives to achieve their shared goal of good governance.
I argue that the employment contract negotiated for between manager and council before
the manager takes office is the key institution defining the relationship between manager and
council and shifting risk between the two. My analytical strategy to understand how managers
and councils align their incentives ex ante is to carefully examine the employment contract and
the way in which the manager shifts risk back onto the elected council through protection from
termination, whether in the form of financial protection or political protection. I argue that a
manager considering taking on a new position in a municipality will survey the economic,
financial, and political environment for key sources of career risk, and in the presence of those
risk factors will seek to shift that risk back onto the council.
28

Previous work has shown that a turbulent municipal environment negatively impacts city
manager tenure and that the presence of an employment agreement plays an intervening role in
terms of how municipal conditions impact tenure (Feicok and Stream 1998). This work suggests,
and I assert, that city managers recognize potential sources of career risk when considering a new
position. In other words, they understand that if a municipality is unstable, whether financially or
politically, the likelihood of improving the city’s condition and/or meeting council expectations
is lowered, and thus career risk increases. A manager faces higher career risk when a city is
experiencing fiscal stress because the manager will be tasked with improving the financial
condition of the city, a more difficult task than maintaining the status quo in a city that is not
experiencing fiscal stress. In a related vein, the manager faces higher career risk when the
economic condition of the city is declining, as declining economic conditions can lead to falling
revenues and increased expenditures. Finally, a manager faces higher career risk when the
political environment of a municipality is competitive and/or tumultuous. If the membership of
the council changes often or if council members and voters frequently disagree about the
formulation of policy, the likelihood of the manager being able to meet council expectations, all
else equal, is lower than in a city with steady or unchanging membership on the council and a
relatively homogenous group of voters with consistent policy preferences. Under each of these
conditions of municipal instability, the manager faces a higher likelihood of failing to meet
council expectations and being terminated or desiring to voluntarily exit his position.
I argue that a manager recognizes municipal conditions which increase the career risk he
assumes when taking on a position, and seeks to shift some of that career risk back onto the
politically elected council through his employment contract. The manager, in the presence of
higher career risk, will negotiate with the council for some degree of protection from
29

termination, whether in the form of severance protection or protection from termination before or
after a municipal election. These terms of the contract define the stake of each party (manager
and council) in the principal-agent relationship between them. Severance protection serves a
particularly important role in shifting risk between council and manager, because as severance
protection increases, the costs that the council would incur to terminate the manager increase,
which may decrease the likelihood of the council terminating the manager, all else equal. Not
only does severance protection impose a defined cost on the council to terminate the manager but
it offers the manager insurance against termination; severance pay provides the manager with
continued income affording him both the time and resources to seek employment elsewhere. The
higher the severance pay, the more time the manager may take after a termination to secure
another position. This time can be particularly valuable to managers who are terminated for
failing to improve municipal condition during their tenure or as a result of political
disagreements with the council, both of which can tarnish the manager’s reputation. My
attempts to understanding how municipal conditions impact risk shifting between council and
manager through the terms of the employment contracts in this analysis adds to the current
knowledge regarding the institutional relationship between council and manager, which has
implications for the formulation and implementation of all types of local policy.

30

CHAPTER THREE: LOCAL GOVERNMENT ADMINISTRATION
While it is a common topic of scholarly inquiry, local government can be an ambiguous
term, as there are multiple forms of local governments. General purpose local governments, the
most familiar type, perform a wide range of functions and include towns, counties, and
municipalities. In most cases, all of the local governments within a region were not planned at
the same time or with the same purpose in mind. Local governments grew and developed in
response to citizen demands for services and governmental functions that weren’t already
provided by the state or federal government (Bowman and Kearney 2008). Because of this
pattern of development, local governments have become the level of government responsible for
many of the day-to-day services provided to citizens (Levin and Tadelis 2010). As evidenced in
the story of San Bernardino, the responsibilities of local governments to citizens are vast and
significant (eg. Public safety protection) and a decrease in revenues can lead to contentious
debates amongst citizens and local officials about how to provide services, at what level, and at
what cost. Municipalities have historically been the primary provider of key services and thus,
the primary unit of local government studied by scholars. General purpose municipalities tend to
have greater decision-making authority and discretion, and they also provide a wider array of
services to their citizens than do most counties or other types of local governments (Bowman and
Kearney 2008).
Although the International City/County Management Association recognizes five
functional forms of local government
10
, the large majority (about 84% according to the ICMA)
of general purpose local governments tend to operate under one of two government structures:

10
The five functional forms are council-manager, mayor-council, commission, town meeting, and
representative. Due to their low incidence in cities of more than 50,000 people, the commission, town
meeting, and representative form of government are much less frequently the target of scholarly inquiry.
31

the mayor-council form or the council-manager form (Renner and DeSantis 1998, Frederickson
and Johnson 2001). Many of the largest municipalities in the country, like Los Angeles or New
York City, operate under a strong mayor-council form of government characterized by relatively
powerful council members who are popularly elected to represent smaller districts and a strong
mayor elected through citywide elections (Renner 1988). These cities sometimes employ a city
manager or a chief administrative officer who serves as a policy advisor with regard to budgetary
issues, personnel issues, or other highly technical policy concerns, but this person typically
answers directly to both the council and the strong-mayor. This institutional relationship is such
that a chief administrative officer reporting to a strong-mayor and a council typically has much
less authority than a city manager with the broader authority and discretion over the day-to-day
administration seen in council-manager cities.
Most smaller municipalities (under 100,000 in population) operate under substantially
different government structures than America’s largest cities. Smaller cities are more likely to
use the council-manager form of government characterized by an elected city council who
chooses an un-elected, professionally-trained city manager who is responsible for many of the
same duties as a mayor but is chosen under different conditions and answers to the council rather
than the electorate.
One key thing to consider when examining governing structure at the local level, is that it
is up to local voters to decide which basic legal form of government they want, and in many
cities, voters have taken steps to alter their form of government by adding or removing individual
positions without actually changing the legal form of government. For instance, in 2006, San
Diego changed from a council-manager form of government and added a strong mayor and a
chief operating officer who reports to the mayor, while Oakland amended their council-manager
32

form of government and added a strong mayor position but maintained the position of city-
manager, creating a sort of hybrid form of government that combines structures from both the
traditional council-manager and mayor-council form (Bowman and Kearney 2011). Many local
governments are moving toward a hybrid form of government by adding new institutions to their
government structure without removing old ones. For instance, cities tend to change their
institutional structures incrementally, with council-manager cities adapting many of the
characteristics traditionally associated with strong mayor cities and vice versa (Frederickson and
Johnson 2001). The trend over the last several decades has been for “a sizeable portion of
previously mayor-council cities to adopt ‘city-manager-like’ positions” (Meek, Schildt, & Witt,
2002, 146). With the growth in popularity of the council-manager form of local government and
the general trend towards creating a separation between politics and administration at the local
level, many scholars have examined the impact of municipal form on local outcomes.

Council-Manager v. Mayor-Council Cities
A city manager, as the highest serving official in a managerial capacity in the
municipality, is important to the overall administration of local government. The ability of a
local government to provide citizens with efficient and effective services largely depends on the
health of their budget and the relative wealth of their tax base, and city managers are tasked with
advising the council on the formulation of the budget and other fiscal policies and with
implementing the budget once it is passed. The city manager is typically hired based on some
combination of his experience and education, both of which serve to train the manager in the
specific techniques important to managing municipal finances: budgeting, accounting,
33

performance measurement, auditing, capital planning, debt administration, and cash
management.
Proponents of local government reform often argued that institutional reform would lead
to lower expenditures and more efficient provision of services, particularly if local governments
moved toward the adoption of professional management practices borrowed from the business
community. A full discussion of the virtues of council-manager government with regard to
outcomes is beyond the scope of this project, but some of the empirical work in the area offers
important guidance on the structure of the principal-agent relationship in government so a short
review is warranted. The argument in favor of a move toward council-manager government has
typically centered on the idea that with a professional manager at the helm of the organization, a
manager facing incentives to reduce costs while maintaining service quality, council-manager
municipalities will realize efficiency gains over mayor-council governments. However, after
several decades of research on the topic, the empirical evidence of this is far from conclusive.
Numerous scholars have found that a change in political structure does result in increases in
service delivery and increases in expenditures on services (Benton 2002, Schneider and Park
1989, Feiock and Tavarres 2002). On the other hand there is a similar body of evidence
suggesting that local form of government has no significant effect on municipal expenditures
(Morgan and Kickham 1999, Hayes and Chang 1990). Finally, empirical work on administrative
capacity has suggested that administrative capacity and local decision maker professionalism (or
their skills and experience) do play a role in the initiation of new policies at the local level
(Donahue, Selden, Ingraham 2000, Honadle 1981).
While numerous factors influence the behavior and decision-making of a city manager,
those factors can shape behavior only so much as institutional arrangements allow. Institutional
34

arrangements define the range of acceptable choices available to a manager and define the
manager’s incentives to make given decisions (Hambrick and Finkelstein 1987). As such, it isn’t
sufficient to simply study the general form of government. A majority of municipalities in the
United States now employ a city manager, yet no empirical research has been conducted to
examine micro-level institutional design and the rules under which local public managers
operate. The city council doesn’t have the power to decide whether or not they would like their
municipality to employ a professional, unelected city administrator, but they do have the ability
to not only choose that administrator, but also to write the contract which governs that
administrator’s behavior. As such, the employment contract agreed to by the city council and the
city manager is a key piece of the institutional design of municipalities, but has yet to be
systematically studied by scholars of local government or public management. While the
previously described literature related to how institutional structure impacts municipal
expenditures and local policies has deepened our understanding of the role form of government
plays in shaping policy outcomes, it has not sought to examine the makeup of those institutional
structures themselves. In other words, while this literature illuminates the impact of operating
under the council-manager form on local outcomes, it has not shed light on the employment
contract, the institution which defines the incentive structure under which these ever important
city managers operate, or how municipal conditions shape the incentive structure created by the
employment contract.

35

City Manager Career Concerns and Turnover
Previous work on the compensation of local government managers is limited. While some
scholars have examined manager compensation and tenure, there is no systematic work on ex-
ante protections or terms of employment for city managers. There is evidence that city manager
salary is a function of productivity, difficulty of the job, and professionalism (Goldstein and
Ehrenberg 1976).
Considering the growing importance of city managers to not only the administration of
municipal government but also the formulation of municipal policy, turnover among managers
has become an increasingly studied topic among scholars (eg. McCabe et al. 2008, Feldman and
Khademian 2002, DeHoog and Whitaker 1990, Feiock and Stream 1998, among others).
Scholars suggest that leadership turnover in organizations can have negative repercussions, as a
change in municipal leadership disrupts operations and the hierarchical structures and
relationships within the organization (Feldman and Khademian 2002, McCabe et al. 2008). With
the important role that city managers play in determining both the substance of policies and the
way in which they are implemented (Svara 1999), a change in management can significantly
disturb the everyday operations of the city (Feldman and Khademian 2002). Turnover has
important implications for municipal management as a “change from one city manager to
another has significant policy implications if different managers bring different preferences,
skills, and backgrounds to the job” (McCabe et al. 2008, 381).
The previous literature has shed light on the reasons for city manager turnover,
suggesting that political conflict and change on the city council increase the likelihood of
manager turnover (McCabe et al. 2008), that a desire for professional advancement is equally as
likely to lead to manager turnover as political conflict (DeHoog and Whitaker 1990) and that
36

ideological conflicts significantly shorten manager tenure (Feiock and Stream 1998). In the case
that manager performance is the impetus for termination, research shows that fiscal policies can
modestly impact tenure, the logic being that “in communities with strong fiscal capacity, low
taxes, and low debt, the performance of managers may be noncontroversial… while smaller,
wealthier, more homogenous communities may be more politically stable and experience less
turnover among managers” (McCabe et al. 2008, 381). The literature has also examined the
implications of city manager turnover, suggesting that manager turnover disrupts urban
governance relationships between key public officials (Feldman and Khademian 2002). There is
also a line of research into the career concerns and career paths of city managers which presents
evidence that most city managers serve in cities other than those they grew up in (DeSantis and
Newell 1996), the majority of city manager exits are due to involuntary terminations (Kammerer
et al. 1962), and that mobility plays a key role in city manager career paths because they most
often proceed from one organization to another rather than through the ranks of a single
organization (Barber 1988).
Turnover among city managers has become an especially interesting topic of inquiry for
scholars of public administration and local government in part due to the unique structure of
career development in local government management (DeSantis and Newell 1996, McCabe et al.
2008). In one recent study, the average tenure of city managers in Florida was estimated to be
only 3.4 years per city (Feiock and Stream 1998) while others have shown that only 5 percent of
acting city managers have been in their current position for more than 20 years (Watson and
Hassett 2003). Most city managers are highly educated, typically with a graduate education, and
well-paid for their efforts. However, given that the city manager is the highest managerial
position within the municipal organization, career advancement within the profession requires
37

mobility. Nearly 80 percent of city managers serve in a community different from the one in
which they grew up or originally resided (DeSantis and Newell 1996) and given that the city
manager post is the highest position in the municipal organization, professional advancement
requires a move to a new organization.
There are two common career paths for city managers. The first often proceeds from
“appointment as an assistant city manager to manager of a small town, and finally to a city
management post in a larger or more prestigious community” (Barber 1988) with each step
requiring a change in municipal organization. The second “typical” career path is that in which a
manager starts his or her career working in one city and moves up the organizational hierarchy
from one position to another. This career path is more often seen in large cities that are known
for “growing their own” managers through positions available within the city (Watson and
Hassett 2004).
While turnover has become a regular target of empirical inquiry, the study of city
manager turnover has focused almost exclusively on the factors that lead to turnover and the
negative impact turnover has on municipal performance and outcomes. In terms of the causes of
turnover, scholars have found that turnover is more often due to termination by council than
voluntary exit on the part of the city manager (Kammerer et al. 1962), and that in both the case
of termination and voluntary exit, political conflict on the city council is associated with higher
rates of manager turnover (DeHoog and Whitaker 1990, and Kaatz 1996). Specifically, city
manager tenure decreases significantly as their responsiveness to local politicians decreases
(Kammerer et al. 1962, Loveridge 1971). City managers serve at the will of the council, and as
such, political disagreements and political controversy can put politicians and managers at odds,
increasing the likelihood of city manager termination or voluntary resignation (Kaatz 1996,
38

DeHoog and Whitaker 1990). McCabe et al (2008) have empirically shown that an increase in
council turnover is significantly associated with an increase in city manager turnover. In short,
they wrote, “turnover on city councils may lead to turnover in city managers” (McCabe et al
2008, 381).
Additionally, a city’s fiscal condition can impact manager turnover. Scholars have
posited that in municipalities with low debt, strong fiscal capacity, and political stability,
manager performance is often noncontroversial, leading to lower rates of turnover and longer
tenure among managers (Feiock and Stream 2002). A community’s fiscal capacity has a
significant impact on manager turnover, with administrators serving in more affluent cities
having lower rates of turnover (Watson and Hassett 2003). One explanation for the role of fiscal
condition in city manager turnover is the growing use of municipal economic performance as a
proxy for a city manager’s skill and accomplishments in office. Municipalities have increasingly
delegated the task of managing economic development to the city manager (Banovetz 1995) and
ICMA surveys confirm that city managers see managing economic development as one of the
most important issues they face on the job (Renner 2001). The responsibility over economic
development can bring benefits to city managers who can credibly claim credit for economic
growth in their municipality. Successful oversight of economic growth can positively impact the
professional reputation of a manager and lead to new, more lucrative job opportunities in other,
larger cities (Stein 1990, Feiock and Stream 2002). However, just as managers can professionally
reap the benefits of success in economic development, they can also be blamed for economic
distress and experience a negative impact on their professional reputation. A manager currently
working in a city experiencing economic growth is more likely to be recruited to new, more
lucrative jobs, while a manager in a city experiencing fiscal distress is more likely to be
39

terminated by the council (McCabe et al. 2008). In their study of the causes of city manager
turnover, McCabe et al (2008) found that the significant determinants of city manager turnover
are council turnover, economic change, and per-capita income.
The mechanism suggested by the results of these empirical studies is that managers who
enter positions where the likelihood of meeting council expectations is lower due to unstable
municipal conditions are more likely to be terminated when they are unable to improve
municipal conditions or please the council with their performance. This mechanism is at the
heart of the theory of risk-shifting I put forth: poor municipal conditions, such as fiscal distress
or frequent council turnover, increase the likelihood that a manager will not meet council
expectations and will be terminated or choose to voluntarily exit. This change in the likelihood of
failure on the job is the single most significant source of career risk facing all managers
considering taking on a new position in a municipal organization, and as such they should seek
to protect themselves from this risk by shifting it back onto the politically elected council
through the terms of the employment contract.
This theory of risk-shifting between council and manager is unique in that the study of
employment contracts has rarely been applied to municipal level management, with the
exception of a number of studies focused on the tenure and turnover of local government
managers (Feiock and Stream 1998). One study of county managers found that political conflict
within the county and other organizational characteristics lead to an increased likelihood in the
board of supervisors dismissing the county administrator (Tekniepe and Stream 2012). There is
a long line of research on municipal governance which suggests that municipal conditions are the
underlying impetus for the operation of both push and pull factors which influence city manager
employment decisions (Kammerer et al. 1962, Feiock and Stream 2002, Barber 1988, DeHoog
40

and Whitaker 1990). Push factors are those that force a manager from his current post or lead the
manager to voluntarily leave, while pull factors are those that attract a manager to a new
opportunity. Scholars have primarily focused on the reasons for managerial turnover. While this
line of research has led to a more nuanced understanding of the ex post decision to terminate
employment, no empirical work has examined the ex ante protection from risk which structures
the principal-agent relationship and has implications for the interaction between the council and
the manager and the degree of autonomy given to (or perceived by) the manager. I expect that
the same factors that influence termination influence the degree to which city managers shift
career risk back onto city councils in contract negotiations.
The process of hiring a new city manager is time consuming and costly. Many
municipalities hire public sector recruiting firms to manage their search and the process can take
several months to complete. Thus, when the city council does select a candidate to hire as the
city manager, the retention of that manager is a concern. In fact, bureaucratic turnover is an
important concern of most principals seeking to hire public managers. Hirschman (1970) argued
that in the presence of a decline in the quality of an organization employees choose to exit, voice
their concerns, or show loyalty to the organization. Part of the information asymmetry between
the principal and the agent is that the principal can’t fully know the agent’s intent to leave the
organization, but alternatively, the agent can’t fully know the nature of the municipal condition
and how those (sometimes unobservable) conditions affect the difficulty of the job and the
likelihood the manager will be able to meet council expectations.
However, the manager can survey the environment of the municipality and can seek to
protect himself from the risk of termination by gaining an understanding of the current affairs of
the city. Public sector executive recruiters and city managers alike have expressed the
41

importance of gaining as full an understanding as possible of the condition of a municipality
before taking on a position (Roberts 2013). Rod Gould, city manager of Santa Monica, said that
before accepting a city manager position, “any job candidate must do their research by reviewing
recent budgets, the last financial report, the current service levels, and recent press accounts that
mention the city council” because without doing that background research, “you’re taking a big
risk” (Gould 2014).

Politics and Administration in Local Government
The council-manager form of government originally grew in popularity based on the idea of
the competent, neutral city manager, and for many years scholars viewed city managers as
politically neutral technocrats with a limited role in the policy-making process. It was long
thought by scholars of local government that the council should not get involved in
administration, the manager should have no involvement in shaping policies, and the manager
should occupy the role of a “neutral expert who efficiently and effectively carries out the policies
of the council” (Svara 1998, 52). This belief stemmed from the idea of the politics-administration
dichotomy, first proposed by Woodrow Wilson (1887) who, in his seminal essay “The Study of
Administration,” writes
“The field of administration is a field of business. It is removed from the hurry and strife
of politics… Administration lies outside the proper sphere of politics. Administrative questions
are not political questions. Although politics sets the tasks for administration, it should not be
suffered to manipulate its offices” (18).

While Wilson is often credited as the first to advance the idea of the separation of politics
from administration, the model took hold among scholars of local government during the 1930s.
As Bertelli and Lynn (2006) describe the reform movements that began in the nation’s large
42

cities, “So-called orthodox ideas – scientific management, separation of politics and
administration, neutral competence, unity of command – far from being products of abstract
theorizing, were elements of reform agendas intended to empower government to meet the
challenges of a growing, industrializing, urbanizing, and diversifying society” (17). Early
adopters of the council-manager plan embraced the ideas of independent-minded reformers who
promoted the separation of politics from administration through a non-partisan but strong
executive (Bertelli and Lynn 2006). Founded in 1914, the City Manager’s Association (now
ICMA) facilitated the development for a professional identity for those individuals serving the
role of professional local government manager. The City Manager’s Association encouraged
municipalities to clearly define within their city charters or municipal codes the duties and
powers of the city manager, a public official who they argued should be free from political
influence (Stone, Price and Stone 1940). The first city managers were experienced professionals
from various fields relevant to city government such as planning, architecture, and engineering,
the idea being that these non-partisan experts could apply their knowledge and experience to the
administration of local government. In the beginnings of the council-manager form of
government, a president of the National Municipal League said, “It is not the politicians, not
even the people at large who have initiated the great modern improvements in city governments,
but experts, sanitary and civil engineers, architects and landscape gardeners, bacteriologists,
physicians, educators and philanthropists” (quoted in Martin 1977, 128). Along with the growth
in the adoption of the council-manager municipal plan came increased demand for professionally
trained public managers, and a number of universities developed academic programs for city
managers (Stillman 1974, Bertelli and Lynn 2006). In the decades since, despite the early beliefs
43

in the virtues of the politics-administration dichotomy
11
, scholars have begun to question the
ability of local government officials to truly separate politics from administration and have
suggested that the politics administration dichotomy may be more symbolic than realistic in
nature (Svara 1998, Stillman 1974).
The true separation of politics and administration in local government, while a primary goal
of the early progressive reformers, may be a goal that practitioners can work toward but never
fully achieve. Given the nature of the principal-agent relationship between the council and the
manager, and the electoral pressure placed on city councilors, the city manager is never
completely insulated from political concerns or debates. Empirical research has suggested that
city managers are integral to municipal policy-making and rarely immune to political pressure
(Nalbandian 1999, 2000; Svara 1999, 1998). As Svara (1999) explains, “It [the politics
administration dichotomy] is commonly viewed as the traditional statement of how officials
should relate… [but] recognizing the interdependent relationship between elected officials and
administrators – as did the framers of the council-manager form – leads to a model of
complementarity rather than dichotomy” (51). As Svara suggests, city managers and elected
council members have developed relationships characterized by interdependence rather than
independence. City manager are now a fundamental part of the municipal policy making process,
often proposing policies, making policy recommendations to the council, and facilitating the
democratic communication between citizens and council members (Nalbandian 1999, Martin
1990, Morgan and Watson 1992). In fact, some scholars have argued that city managers are

11
For an at-length historical analysis of the rise of the American administrative model and the emergence
of the field of public administration see Bertelli and Lynn 2006.
44

actually expected to play an integral role in the formation of municipal policy, particularly with
regard to economic development (Banovetz 1995, Wright 1969, McCabe et al. 2008).
While most of today’s city managers are professionally trained in core competencies integral
to the administration of local government, in practice, the institutional nature of the position does
not incentivize managers to behave as purely apolitical public servants. Managers serve at the
will of the politically elected council and the continuation of their employment is constantly
linked to their ability to meet council expectations. Even city managers that can only be
terminated by a unanimous vote of the council are not immune to political pressure. City
managers are always one city council vote away from employment termination, a dynamic that
incentivizes them to act in accordance with the preferences of the council, whether those
preferences are homogeneous and how closely they are tied to electoral pressure varies from
organization to organization. However, the core mechanism at work in local government
management today is that city managers must balance two sets of incentives which do not
necessarily align: the incentive to perform their duties in accordance with the professional norms
and ethics on which they were trained and the incentive to behave in accordance with council
preferences.

45

CHAPTER FOUR: THEORY AND PREDICTIONS
At its core, agency theory examines how principals attempt to control the behavior of
agents to keep their actions in line with the desires and goals of the principal (Jensen and
Meckling 1976). An important type of agency relationship is a contract between one principal
and one agent, under which the principal hires a manager (the agent) to perform some service on
his behalf, thus necessitating the delegation of decision-making authority to the manager. The
principal-agent problem arises when the self-interest of the two parties do not align. If both the
principal and the agent seek to serve their own self-interest, and if both have their own unique set
of preferences, the agent will not always act in the best interests of the principal.
In a general sense, the heart of the principal-agent problem is that managers delegated
responsibility from principals cannot be expected to give the same amount of attention to detail
and care of action as the principal himself. As Adam Smith wrote in 1776,
“The directors of such companies, however, being the managers rather of other people’s
money than of their own, it cannot well be expected, that they should watch over it with
the same anxious vigilance with which the partners in a private copartnery frequently
watch over their own. Like the stewards of the rich man, they are apt to consider attention
to small matters as not for their master’s honour, and very easily give themselves a
dispensation from having it. Negligence and profusion, therefore, must always prevail,
more or less, in the management of the affairs of such a company.”

The principal-agent problem is pervasive in organizations requiring the delegation of authority.
There are two primary ways through which a principal can seek to limit deviation from his
preferred line of behavior on the part of the manager. The first is to offer incentives for the agent
encouraging him to behave in a way that aligns with the principal’s interest. The second is to
engage in monitoring designed to decrease the information asymmetry between the two parties
and to limit activities on the part of the agent that do not align with the principal’s interest.
46

The primary mechanism organizations use to address the principal-agent problem
described by Smith is the clear delineation of individual rights which determine how costs and
rewards are distributed among the participants in the organization. According to Jensen and
Meckling (1976), “Since the specification of rights is generally affected through contracting,
individual behavior in organizations, including the behavior of managers, will depend on the
nature of these contracts” (4). While much of the literature on private firms has explored the
nature of such contracts and the structure they create between principal and agent, this
relationship has not been thoroughly examined in the public sector context.
The principal-agent problem itself is quite general, existing in all organizations within
which a principal seeks to induce an agent to behave in a way that maximizes the principal’s
utility. As Aggarwal and Samwick (1998) write, “The standard principal-agent model posits an
economic tradeoff between inducing the correct amount of unobservable effort by the agent and
minimizing the amount of risk she is required to bear” (2). The principal-agent problem is
pervasive in private firms, corporations, universities, non-profit organizations, and public sector
organizations. However, the contractual relationships in each of these settings vary significantly,
requiring the development of theories specific to each type of organization. Just as the principal-
agent problem is a central concern of private organizations, democratic governments must also
consider the nature of the relationship between the principal and the agent in government. Not
only do specific theories relating to the nature of the principal-agent relationship need to be
developed for both public and private sector organizations, but even within the public sector,
scholars should be careful to generalize too broadly, as varying levels and types of government
organizations operate under vastly different institutional structures.
47

In the public sector, applications and examinations of the principal-agent problem center
around how democratic governments can structure incentives so that managers delegated
authority and discretion to make decisions regarding policy formulation and implementation will
make those decisions in a manner consistent with the interests of the principal. While the
institutional nature of a principal-agent relationship in public sector governance may vary, public
managers never serve a single principal, for the necessity for accountability to the electorate
persists, regardless how indirect or tempered the link between the manager and the public may
be. As Bertelli (2012) writes, “In a democracy, administration must be constructed in such a
way that it serves the will of the people through their elected representatives. The connection
between the public and any given administrative task may be quite distant, but it must be in
place” (3). While the city manager serves the council directly, the nature of the electoral
connection is such that the manager serves the people through the city council, a body that is
elected by the people to serve in their interest. Any attempt to examine the principal-agent
relationship between council and manager must, due to the very nature of the principal-agent
relationship in public sector governance, include consideration of the electorate’s preferences.
Understanding the political process is key to understanding the principal-agent
relationship within government organizations. In the principal-agent setting at the federal level of
government, Niskanen (1975), focusing on the U.S. Congress and federal bureaucracy, finds that
principals (members of Congress) as rational actors seek to maximize their chances of reelection
while agents (bureaucrats) as rational actors seek to maximize their wages, benefits, and perks.
Since the work of Niskanen in the 1970s, agency theory has developed into the predominant
theoretical framework for understanding bureaucratic politics in which public managers are
tasked with the formulation and implementation of public policy by principals in various levels
48

of government. However, much of the work on agency theory in political science seeks to
explain the behavior of principals, with regard to the way they use monitoring, incentives, and
procedural control to shape the behavior of bureaucrats (Bendor, Taylor and Van Gaalen 1987,
Epstein and O’Halloran 1994, Huber and Shipan 2002, Lupia and McCubbins 1994, McCubbins
and Schwartz 1987). However, there exists less work in the stream of public administration
research to examine and understand the behavior of managers, who are also political actors, in
relation to principals. Much of the work on agency theory in political science treats agents as
“automatons or kleptocrats,” (Carpenter 2006, 42) rather than as independent political actors. As
Daniel Carpenter (2006) observed, in order to better understand the behavior of agents, principal-
agent theories should “be more psychologically realistic, incorporate complexity… [and]
maintain a degree of goal pursuit and rationality” (42). As Carpenter suggests, a fully developed
theory of the principal-agent relationship must focus as much on the agent as an independent
rational actor as on the principal. I am responding to this call by developing a theory of ex ante
risk-shifting between council (principal) and manager (agent) which posits that municipal
conditions which create increased career risk for city managers drive the nature of risk-shifting
seen in city manager employment contracts negotiated with council.
One of the core assumptions of agency theory is that principals and agents are rational
actors with divergent preferences. The presence of information asymmetry between principal and
agent can impact the degree to and manner in which principals and agents seek to align their
interests. Two types of information asymmetry are analytically important to the relationship
between council and manager: adverse selection and moral hazard. Adverse selection affects the
theory of risk shifting between council and manager because I assume that the city manager
(agent) has more information regarding his own skill or qualifications and his own preferences
49

than does the council (principal). However, moral hazard is perhaps most important to the
principal-agent problem that exists between council and manager as the council (principal) must
address the potential that the manager (agent), who does not directly bear the consequence of his
actions (via popular election), will take a risk in his capacity as city manager that he may not
normally take were he held directly accountable to the voters as is the council. Given these
information asymmetries, agency theory, as developed in the political science literature,
hypothesizes that principals will seek to monitor, constrain, and incentivize the behavior of
agents so as to reduce the risk of adverse selection and moral hazard. While monitoring is an
important part of the relationship between council and manager, it primarily takes the form of
informal and non-contractual periodic observance and supervision of manager behavior.
12
In the
case of the city manager and council, I focus on the shifting of risk between council and
manager, the result of which serves to formally constrain and incentivize the behavior of the
manager. As the degree of risk shifted from the manager onto the council increases, the council’s
strength of constraint over the manager’s behavior decreases.

Agency Theory in Local Public Management
The relevant work in the public administration literature has shown that the policy
choices of public managers are tempered by their desire to serve the public interest (Perry 1997),
their autonomy (Lipsky 1980, Carpenter and Krause 2012), or their policy preferences (O’Leary
2013). Policy implementation at the local level involves making specific choices about how to

12
Chapter Five presents an in-depth discussion on the lack of clearly defined monitoring procedures
within city manager employment contracts. Council has incentives to avoid contractually committing to
specific performance goals and typically includes only vague reference to annual performance
evaluations. Interviews with acting city managers suggest that most council monitoring is informal and
non-contractual.
50

interpret and put into practice the policies laid out by the city council in the municipal codes,
statutes, and annual budgets. While the city manager is typically hired for knowledge or
expertise in areas related to city management, the city council seeks to create institutions and
incentives which align the interests of the manager with their own, thus diminishing the
principal-agent problem.
There are several ways in which elected officials, as part of a city council, can seek to
influence decision making by the unelected city manager, and these various control mechanisms
are most explicitly laid out in the employment contract of the city manager. As these contracts
create the incentive structure within which the city manager operates, they are an important piece
of the theoretical understanding of the relationship between principal and manager, yet they have
received little attention from scholars of public management.
Much of the previous work on the principal-agent problem has focused on the alignment
of incentives and linking of executive compensation to performance outcomes. There exists little
empirical evidence of pay for performance or incentive based pay structures used for public
managers at the local level. The limited empirical work examining the link of incentive based
local public manager compensation and performance outcomes is based on a small scale study of
school superintendent salaries and another study of tax assessor performance uniformity
(Santerre 1990, Santerre and Bates 1996). While these studies do at least speak to the use of
performance based pay at the local level, neither approximates the principal-agent relationship
between city council and city manager or offers evidence of the existence of or effectiveness of
incentive based pay for city managers.
Perhaps pay for performance schemes in city management have not been empirically
studied because this type of compensation and incentive structure is rarely used in local
51

government management contracts. This necessitates a different approach to understanding the
principal-agent relationship between city council and city manager. Bowlin (1997) examines the
impact of various factors on the structure of public manager employment contracts. He argues
that taxpayers or the general public, as the principal, are primarily interested in the efficient
operation of the government while the public manager (agent) is primarily interested in
maximizing both his financial compensation and leisure. The problem with this principal-agent
model, and many others focused on federal bureaucratic managers, is that it is based on the
assumption that the government manager has a primary incentive to “increase the size of the
budget they administer with only limited incentives to minimize costs” (205). A model that
assumes a manager is concerned with their salary, benefits and professional reputation, but also
assumes that a manager seeks to maximize the budget with little incentive to minimize costs
simply ignores the nature of the profession of local government management. A city manager
concerned with his professional reputation and ability to secure future employment as a manager
with other organizations faces clear incentives not to maximize costs, but to maximize
efficiency. Additionally, these models are based on the assumption that the principal (the
taxpayers) are primarily interested in the efficient operation of government. However, in council-
manager local government, constituents elect the council to act as their voice, but for the
manager, the taxpayer is not the principal, but rather the council is the principal and the council
faces many political pressures beyond just that of efficiency.
There is no empirical evidence explaining why city manager compensation is rarely
linked to performance outcomes, but this phenomenon has been theoretically attributed to the
incentives of the city council (principal) and the difficulty of measuring the performance of a city
manager (agent). One common argument made in support of the council-manager form of
52

government is that city managers will bring efficiency gains to the municipality by virtue of their
professional training and further distance from the electorate. This implies that managers have a
credible commitment to the goal of efficiency. In other words, in order for this mechanism to
work, city managers must face clear incentives to increase efficiency in the municipality, and
even then the mechanism is tempered by city manager risk-protection which offers managers the
protection from termination so that they may seek efficiency gains when doing so goes against
the interests of council. While concerns over maintaining and improving their professional
reputation may incentivize city managers to increase efficiency for the betterment of their
professional reputation, there is no empirical evidence that councils formally incentivize
managers to increase efficiency. In their empirical study of the association between the council-
manager form of government and efficiency gains in local government, Hayes and Chang (1990)
find no empirical support for the hypothesis that the council-manager form of government is
more efficient than the mayor-council form of government. While their study does not
empirically address manager contracts, they attribute the lack of efficiency gains by city
managers to the possibility that manager “compensation is not tied to efficiency levels” (176).
With regard to their decision-making, city managers must balance two types of behavior:
behavior undertaken to improve municipal condition and behavior undertaken to meet the
expectations of the city council. While these two types of behavior can be one in the same, they
can also be mutually exclusive, and it is in the circumstance of the mutual exclusivity of these
two types of behavior that career risk develops for city managers. Thinking back to the story of
San Bernardino, City Manager McNeely behaved in a way so as to improve the long-term
financial condition of the city, regularly warning the council that without drastic steps to reduce
the deficit the city could face bankruptcy. However, McNeely’s interest in improving municipal
53

condition was not shared by the council whose interests stemmed from maintaining high levels
of the provision of services that were deemed popular with the electorate. While improving
municipal financial condition may be in the interest of the manager with regard to bolstering his
professional reputation as a competent manager of municipal finance, the actions necessary to
achieve efficiency gains may go directly against his interest in pleasing the city council with
regard to his desire to maintain his employment.
This dynamic between a manager’s interest in improved municipal condition and a
council’s interest in reelection offers an explanation for the presence of vague performance
objectives assigned to the manager by council. While the manager always maintains an interest
in improved municipal performance, the council’s interest in the improved municipal
performance changes as the preferences of the electorate oscillate. The city council wants to
commit to a well-run government, and the city manager will provide that if (a) allowed to by
politicians, and if (b) the risks to fiscal health are not too high. However, the city manager faces
the risks of blame-shifting by politicians and seeks to protect himself through protection from
termination before or after elections. In addition, the manager faces the risk of personal hardship
and harm to his professional reputation should he be terminated by the council and seeks to
protect himself through financial severance protection.
The council may have many goals with regard to municipal policy, including efficiency
gains and improved municipal condition, but those goals only remain so long as political support
for them remains. Council members responsive to voters may change their stance on municipal
policy so as to continue to serve the will of the people. One theory as to why manager
compensation is not linked to performance objectives is that the council has numerous,
fluctuating goals and leaves performance expectations in the contract vague so as not to tie their
54

own hands to any one particular outcome. To contractually define clear performance goals for
the manager would be for the council to effectively tie their own hands by formally committing
to that goal. The council is responsive to the median voter, for whom it is difficult to fully
ascertain any efficiency gains or productivity gains made on the part of the city manager.
Without pressure from the median voter for efficiency gains, the council’s expectations of the
manager, and the council’s decisions regarding performance review of the manager may not be
based on efficiency gains at all (Deno and Mehay 1987). This offers an explanation as to why the
performance expectations and performance review section of city manager employment contract
are typically vague and absent of any quantitatively measurable performance goals or outcomes.
If a council is primarily concerned with reelection and is responsive to the demands of voters,
then it is in their best interest to leave the performance expectations of a city manager open to
revision and review, during which time the demands of voters may change.
Finally, even if a council does desire efficiency gains on the part of the city manager, it
is difficult to objectively measure the performance of city managers with regards to efficiency as
the city manager has a myriad of responsibilities and is in charge of numerous departments, a
multitude of employees, and the provision of all services within the municipality. Local
governments function without any single one goal, they have various goals, which could include
efficiency, effectiveness, equitable service provision, or accountability, and each city council
determines their own set of goals based on the demands of the voters. However, these goals can
often conflict. For example, a goal of improving resident satisfaction with a certain service could
contradict a goal of reducing costs and improving efficiency with regard to that same service.
The reinventing government movement first espoused by Osborne and Gaebler (1992)
encourages principals to shift their interactions with agents so as to steer rather than row, to
55

empower rather than to serve, and to transform rule-driven organizations, but even goals such as
these which were embraced by many as a beneficial reform agenda can directly conflict with
other common goals of government such as accountability and equity. Without the single
overarching motive for profit-maximization found in the private sector, there is no one outcome
on which city manager performance can be objectively evaluated and as such council allows
themselves more freedom to evaluate the manager as they see fit at any given moment in time by
purposefully omitting any measurable performance objectives from the employment contract.

Incentive Alignment in Private Sector Organizations
Research on issues related to executive compensation in the private sector is abundant
(Abowd and Ashenfelter 1981, Aggarwal and Samwick 1998, Rau and Xu 2012, Lys and Sletten
2006) and offers some generalizeable empirical evidence that can guide the development of a
theory about the ex ante relationship between council and manager within the administrative
agency relationship at issue in this project. The theory in private sector research, based on the
competitive market model, is that there is a positive association between dismissal risk and pay
such that in order to compensate for job risk, pay should be higher in more constrained sectors
with a higher degree of job risk and increased costs for dismissed employees (Abowd and
Ashenfelter 1981). Scholars of private sector management characterize this relationship such that
an executive’s pay-performance sensitivity increases as instability in the firm decreases
(Aggarwal and Samwick 1998). Evidence from the business, economics, and accounting
literature suggests that firms with higher volatility or higher distress risk (typically those with
higher stock price volatility or high turnover on the board of directors) are more likely to contract
for larger severance packages with executives, in this case the CEO of a private firm (Rau and
56

Xu 2013). Additionally, younger executives, those recruited from outside the organization, those
coming from more secure positions, and younger managers are more likely to be given larger
severance packages because of their increased career risk (Rau and Xu 2013, Lys and Sletten
2006).
Rau and Xu (2013) write, “the optimal executive contract, consisting of a mixture of cash
and bonus payouts, equity and option-based incentives, and severance pay should be designed as
a whole to attract talented executives and incentivize them to exert effort on behalf of
shareholders” (2). While equity and option-based incentives are not relevant to the local
government context, city manager employment contracts do consist of a mixture of cash payouts
and severance pay. While there has been popular outrage about the dollar value of some
executive severance payouts in both the public and private sector given to executives whose
performance was well below expectations (See Brown 2014, Luckerson 2014), this public
outrage entirely mischaracterizes the purpose of severance agreements altogether. As Rau and
Xu (2013) imply, severance packages are designed to entice talented managers to accept a given
position, particularly in an unstable organization, by offering financial protection should they fail
to succeed. In other words, severance is best thought of as insurance that protects managers
willing to take on positions with a high degree of inherent career risk.
The logic behind the mechanism of severance protection is relatively simple and extends
to the incentives facing city managers. As the stability of a firm or organization decreases, the
degree of protection offered via severance increases as the manager seeks to protect himself from
termination since it is less likely he will be able to meet board expectations given the instability
(Lys and Sletten 2006). Despite popular arguments to the contrary, severance is not a reward for
unsatisfactory behavior. Severance is designed to protect professional executives from the career
57

risk associated with taking on a position in an unstable environment. The stability associated
with a position can be considered in terms of both organizational stability (for example, the
cohesion of the board of directors or turnover on the city council) and financial stability (for
example, changes in quarterly earnings reports or budget deficits). Additional private sector
scholarship suggests that CEOs recruited from outside the organization tend to have larger
severance packages as they have a higher risk of not meshing with the culture of the organization
(Lys and Sletten 2006). Younger executives with less firmly established professional reputations
tend to have larger severance packages than older, more experienced managers because in the
event of termination, a younger executive’s reputation is more easily tarnished and their future
career prospects diminished (Lys and Sletten 2006). The overall findings of the literature on
private sector relationships between boards of directors and private executives suggest that the
more volatile a company is, both organizationally and financially, the more risk protection that is
negotiated with the incoming executive.

Predictions
Based on the core tenets of agency theory, the nature of the principal-agent relationship,
and the unique institutional structure of local governments in the United States, I develop a core
set of predictions about how municipal factors influence the principal-agent relationship in local
government management and the shifting of risk between council and manager. These
predictions are captured by a two by two matrix, which clearly organizes the way in which
conditions within a municipal organization impact the nature of and difficulty of the job of city
manager. These predictions, seen in Figure 1, serve as the basis for the hypotheses tested in this
project.
58

Figure 1. Determining the Degree of Risk of a City Management Position

There is a fruitful line of research debating the relative importance of political
imperatives or economic constraints on the determination of local government policies and
decision-making (Tiebout 1956, Peterson 1981, Donovan and Neiman 1992, Goetz 1994). While
some scholars strongly contend that cities are largely ruled by economic constraints and a desire
to expand their economic base (Tiebout 1956, Peterson 1981), others argue that local
government policy making is largely driven by the preferences of the voters (Dahl 1961,
Donovan and Neiman 1992, Goetz 1994). In other words, while the one side contends that local
government decisions are based on the efficient distribution of goods and services, others argue
that “residents not only control the outcome of elections but they can also determine the direction
of policy” (Hajnal and Trounstine 2010, 1131).
I theorize that at the local level, both economic and financial factors and political or
organization factors constrain and influence local government decision-making. City managers
59

work in complex environments. Local economic and fiscal conditions are constantly changing
and the preferences and policy demands of politicians and citizens also change with time. As this
table shows, I predict that varying degrees of political and organizational constraints combined
with varying degrees of economic and financial condition work together to determine the relative
career risk associated with a given position as a city manager. Career risk, which in this context
can be thought of as the likelihood of the manager failing to meet council expectations and being
terminated by council, is a relative construct, and this table presents a simplified depiction of
career risk for illustrative purposes. Economic and financial condition can be thought of as the
overall stability of the municipality with regard to economic position and fiscal health. This
dimension of the municipal environment is made up of numerous factors related to the overall
health of the municipal economy, such as unemployment or economic growth, and the overall
fiscal stability of the community, such as deficit spending or a decrease in bond rating. An
unhealthy economic and financial condition in the municipality increases the difficulty of the job
of the manager who is tasked with managing the budget and the provision of all municipal
services. Perhaps nothing is more essential to a manager’s ability to improve the quality of life in
a community than his access to resources. Each city manager I interviewed mentioned fiscal
strain or decreasing revenues as one of the biggest challenges for a city manager and one of the
primary sources of career risk. Rod Gould, city manager of Santa Monica said the most
important thing he looked for when considering whether to accept a city manager position was
whether the municipality had “the will and the wallet” that would provide him with the “latitude
to manage with both productivity and quality of life in mind” (Gould 2014). The wallet Gould
refers to is the financial resources that make a city manager’s job easier. The will, on the other
60

hand, refers to a political and organizational culture, which gives the manager the discretion and
authority to make decisions that will improve the quality of life in the community.
The second dimension, the degree of political and organizational constraint, refers to the
degree to which council politics or electoral pressures restrict the behavior of the manager. When
considering a range of possible choices a manager could make regarding how to implement a
policy or how to administer services, certain characteristics of the political environment could
constrain the manager in the sense that certain decisions would be too far out of line with the
principal’s interest and would increase the likelihood of termination. The narrower this window
of acceptable behavior becomes, as a result of political constraint, the higher the likelihood that
the manager will fail to meet council expectations and face termination. For example, taking
municipal elections as one piece of the overall dimension of political and organizational
constraint, a system of competitive elections and an engaged citizenry places pressure on the
council for the passage of certain policies or the achievement of certain outcomes. This would
make it more difficult for a manager to make decisions that contradict the preferences of the
electorate, tempered through the council, than in a municipality in which the citizens are little
engaged in the political process and place little pressure on the council with regard to outcomes.
Rod Gould describes the factors he considered before accepting his current position, “ the
council had a reputation for being divided and the political involvement of the citizens was
legendary. The people in this town enjoy local politics and they expect a lot from government”
(Gould 2014). In other words, Gould recognized that the highly engaged citizenry and the
divided council would likely place pressure on him as manager to behave in a certain manner and
to meet their relatively high expectations. This is why I expect that in the presence of increasing
61

political constraints or worsened fiscal health a manager will seek to shift risk back onto the
council ex ante.
When considering both of these dimensions that contribute to career risk, political
constraint and fiscal health, I develop a set of predictions as to which municipal manager
positions present the greatest degree of career risk and which present the lowest degree of career
risk. A municipal organization with both a high degree of political and organizational constraint
and an unhealthy economic condition presents the highest career risk for a manager.
Alternatively, a municipal organization with a low degree of political and organizational
constraint and a relatively healthy economic and financial condition presents the lowest relative
career risk to a manager. Somewhere in the middle of this spectrum of career risk lie the
municipalities in which political constraints are high but economic and fiscal conditions are good
or in which political constraints are low but economic and fiscal conditions are poor. It bears
reiterating that this matrix is designed to illustrate the relative impact of these municipal level
factors on career risk for city managers, not an absolute impact. In other words, the matrix
simply suggests that a manager entering a position with a city in the top right quadrant faces
relatively more career risk than a manager entering a position with a city positioned in the lower
left quadrant.

Testable Hypotheses
Based on the previous discussion of the nature of the principal-agent problem in local
government management, the primary responsibilities of the city manager, the discussion of the
career concerns of city managers, and my predictions regarding how the two dimensions of
municipal collection determine the relative career risk for a city manager in a given municipality,
62

I present the following hypotheses. I seek to test each of these hypotheses to better understand
how municipal conditions, both in terms of economic and financial condition and in terms of
political constraints, impact city manager protection from termination risk.

Hypothesis 1: As municipal economic instability increases, city manager protection from
termination risk increases.

Hypothesis 2: As municipal financial instability increases, city manager protection from
termination risk increases.

As discussed in chapters one and two, a city manager essentially functions as the CEO of
the municipality and is expected to perform many roles and duties within the municipality. Some
of the most important responsibilities of the city manager include the preparation and
implementation of the budget, oversight of human resource management within the city, which
often includes negotiations with labor unions over pay and benefits, and overseeing the daily
administration of the municipality including service provision. In the role as the manager of
overall municipal fiscal health, most managers are tasked with overseeing economic
development efforts within the municipality. As the matrix above shows, worsening economic
and financial conditions can result in increased career risk for the manager who is the primary
employee of the city tasked with managing these two areas. As such, I predict that increases in
municipal economic instability and increases in municipal financial stability are both positively
related to protection from termination risk.
The economic and political condition of the municipality is only one of the dimensions
upon which the difficulty of the job and inherent career risk of a given position depend. The
other dimension, political and organizational constraints contribute to the organizational
63

structure and political climate in which the manager will be expected to operate on a day-to-day
basis. The council, accountable to the electorate, can place pressure on the manager to behave in
a certain way and can use the threat of termination to induce manager behavior. The unique part
of the principal-agent relationship between council and manager is that the council members
themselves are agents of the people, and as such, the preferences of the electorate are filtered
through the council to the manager. While economic and fiscal constraints are one source of
constraint on local decision-making, “political imperatives and institutional constraints strongly
influence local government decision making” (Hajnal and Trounstine 2010, 1130). As political
and organizational conditions become restrictive so as to constrain the behavior of the manager
to a relatively smaller window of behavioral alternatives, the risk of a given municipal manager
position increases. Considering the nature of the principal-agent relationship between manager
and council and the electoral incentives of the city council, I test the following hypotheses to
examine how political conditions within a municipality impact city manager protection from
termination risk.

Hypothesis 3: As the political cohesiveness of the electorate in a municipality increases, city
manager protection from termination risk decreases.

Hypothesis 3 seeks to test the impact of political cohesiveness among voters on city
manager risk protection. While many studies of national politics focus on voter partisanship,
California local government elections for city council are non-partisan (Kemp 1999, Northrup
1987). Thus, when thinking about political cohesiveness, I refer not to party identification among
local voters, but rather to the homogeneity of the electorate’s ideological preferences in
elections; the more cohesive they are the less diffuse their ideological preferences. For example,
64

in an election in which a large majority selects the winning alternative on the ballot and a small
minority selects the other alternative(s), the voters are considered relatively cohesive. On the
other hand, when considering a ballot with only two alternatives, as the majority of voters
selecting the winning alternative nears 50 percent, and the minority of voters selecting the losing
alternative also approaches 50 percent, the voters are considered relatively disjointed or divided.
The political cohesiveness of voters can have important implications in local elections. A more
politically cohesive electorate is more predictable in elections. On the contrary, the less
politically cohesive the electorate is, the less predictable they are in elections.
13

Hypothesis 4: As the political engagement of the electorate in a municipality increases, city
manager protection from termination risk increases.

The second hypothesis related to political condition depicts the relationship between an active
citizenry and the career risk facing a city manager. As the manager serves the council, who are
each accountable to the people through regular elections, an active citizenry that frequently goes
through the steps required to make local policy directly and to circumvent the authority of the
council can place more pressure on the council for certain outcomes or policies, and the council
will in turn place more pressure on the manager to behave within the confines of what the
council sees as desirable behavior. This is an especially important consideration in the state of
California where citizens have the ability to draft and place initiatives on the local ballot, which

13
It is worth noting that my concept of political cohesion is not a measure of polarization as typically
defined in the political science literature. Fiorina and Abrams (2008) define polarization as “the
simultaneous presence of opposing or conflicting principles, tendencies, or points of view… an intuitive
notion of polarization is as a bimodal distribution of observations” (566). My concept of political
cohesiveness is distinct from polarization in that rather than considering the bimodal distribution of
preferences, I refer to the homogeneity of preferences among voters. Political cohesiveness is a measure
of the like-mindedness of voters, not the ideological distance from one key group to another.
65

if submitted through the correct legal process and approved by a majority of voters becomes part
of the municipal code. A politically engaged citizenry, especially one apt to draft legislation and
place that legislation before voters to directly approve or deny, creates an additional layer of
political pressure which may be applied to the manager directly, rather than through the council.
An engaged citizenry that makes their preferences known directly to the manager can constrain
the manager’s ability to make decisions as the manager now must seek to balance the preferences
of both the political activists and the council, for any forward thinking manager knows that while
council members may change frequently, they are always elected by the people, so ignoring the
wishes of the people can impact the manager negatively in addition to the council. Additionally,
one of the manager’s key roles in the municipality is the implementation of policy, so if an
actively engaged citizenry makes policy, the manager is responsible for implementing that
policy. A manager serving in a city with a highly engaged electorate faces more constraints on
his behavior as he must consider the preferences of more groups and seek to make decisions so
as to address the concerns and preferences of the electorate directly and of the council.

Hypothesis 5: As the competitiveness of municipal elections increases, city manager protection
from termination risk increases.

The final hypothesis related to political constraint seeks to test the notion that the increased
likelihood of change in council membership places a constraint on the manager who must make
decisions with not only the preferences of the current council in mind, but also considering the
possibility that he could be serving a different set of principals in the wake of a municipal
election. City managers are already in the unique position of serving multiple bosses, each with
unique preferences, but a high likelihood of council turnover increases the career risk of the
66

manager. For, while the manager’s behavior may serve the interests of the current council, if the
make-up of the council changes, the manager’s behavior may no longer be seen as satisfactory
and the manager’s risk of termination increases. As such, I seek to test the hypothesis that
increasing competitiveness in municipal elections is positively associated with city manager
protection from risk.
Each of these five hypotheses seeks to test a key piece of my theory of risk shifting
between council and manager. Career risk stems from political constraint and from fiscal health,
and managers survey the municipal environment for these sources of career risk before agreeing
to take on a new position. The employment contract between manager and council is negotiated
for ex ante, with the two parties seeking to allocate each of these risks between the manager and
council (who represents the interests of the median voter). In the presences of a higher degree of
career risk, my theory predicts that a manager will seek to shift that risk back onto council.

67

CHAPTER FIVE: CITY MANAGER EMPLOYMENT CONTRACTS
A city manager’s employment contract, negotiated with council before the start of
employment, is the key institution which shapes the relationship between council and manager
and creates the environment within which the manager operates. The employment contract
clearly defines the terms of the agreement, the compensation and benefits to be provided to the
manager, the circumstances under which the manager can and cannot be terminated, and most
important for this study, the protections from career risk afforded the manager in the form of
severance pay or protection from termination before or after a municipal election. According to
Feiock and Stream (1998), “the International City/County Management Association (ICMA) has
taken an active role in encouraging member cities to provide employment agreements. The 1980
report of the ICMA Future Horizons Committee noted the unique needs of city managers in
relation to their job mobility and political vulnerability. The report advocated the use of
employment agreements” (122). Today, employment contracts between manager and council are
considered standard practice, and both ICMA and the California City Management Foundation
(CCMF) provide sample employment contracts to members of the profession.
14

Of particular importance to this study is an understanding of the form and content of city
manager employment contracts. In the empirical portion of this project, I analyze a unique hand-
collected sample of employment contracts from 332 California municipalities, the most
comprehensive attempt to systematically understand city manager employment contracts to date.
The sample includes municipalities of all sizes, from all regions of California, and of all
municipal conditions. In this chapter, I first outline the theoretical foundations of the shifting of

14
The CCMF sample employment contract includes an addendum written by an attorney explaining the
suggested provisions of the contract. The explanation of recommended severance provision
considerations accounts for more than three pages of this addendum.
68

risk between managers and council and then provide a detailed description of the collection of
the employment contracts and of their content.

Risk-Shifting Between Principal and Agent
Given that principals in local government do not seek to align their interests with those of
their agents through the use of incentives linked to performance metrics, it is necessary to
consider other incentive mechanisms introduced before the start of their relationships with
managers. Similar to private sector firms, municipalities in the market for a new city manager
seek to recruit top talent and to hire a highly qualified executive to manage these fiscal
challenges. However, the labor market for city management is complex, and just as
municipalities seek to recruit top management talent, managers carefully assess the condition and
stability of cities when considering employment and negotiating the terms of employment. What
remains to be seen is whether these conditions are reflected in the terms of the employment
contract related to compensation and severance packages. Any public manager considering
taking on a position as a city manager must consider the management challenges presented by
the stability of the municipality and the resulting likelihood that the manager will be able to meet
council expectations for management. These contracts should be of greater interest to scholars of
public management as they are a rich source of information about the relationship between the
city council and the city manager and offer key insights into how the conditions in a city shape
the degree of risk shifted to the city manager.
While cities want to recruit the best talent to manage their city, managers seek protection
from the negative consequences of risk-taking behavior. City managers considering employment
with any given municipality consider the current condition of the municipality, or the “stability
69

of the municipality,” before negotiating the terms of employment. Likewise, city council
members, in recruiting and negotiating with city managers must consider the stability of the
municipality in their effort to recruit the best managerial talent to supervise the operations of the
city. Furthermore, in their employment negotiations with a given city manager, the council must
balance the city’s desire to keep costs low with the need to reward top managerial talent with
both up-front compensation and ex-ante protection from the damage that failing to live up to
expectations would have on their career prospects. From the city council’s perspective, the more
protection from risk that is given to the city manager via the employment contract, the more
likely it is that the manager might act in a way that does not serve interests and desires of the
council.
A rational council seeks to minimize the amount of severance protection granted to a city
manager, as severance protection directly ties the hands of the council regarding their control
over the manager and increases the costs of a decision to terminate the manager. Severance
protection also provides managers with an insurance policy should they fail to meet council
expectations. On the one hand if the council terminates the manager, the manager has financial
protection from the subsequent harm to his professional reputation and his immediate loss of
earnings. On the other hand, a high level of severance may dissuade the council from proceeding
with termination, even if they so desire. As with any type of insurance, along with severance
protection comes the risk of moral hazard. The more the manager is protected from the risk of
termination or the more credible his commitment to steady administration actually is, the more
likely he is to make decisions or engage in behavior that the council themselves would not. The
manager knows that he himself does not bear the risk of those decisions for if they result in
negative outcomes and he is terminated as a result, he has a financial payout that offers him a
70

period of time of continued earnings to seek employment elsewhere. Agency theory clearly
predicts that the principal, in this case the city council, will seek to minimize such insurance or
protection from risk granted to the manager, while the manager will seek to maximize such
protection from risk. Thus, the negotiation of the employment contract between manager and
council is essential in defining the principal-agent relationship between the two parties.
I argue that the employment contract is one of the most important institutional constructs
in council-manager government because it shifts risk between manager and council in such a
way that can shape the implementation of every policy and the way in which the manager
approaches the provision of all government services. While the importance of the terms of the
contract are clear, there is no theoretical or empirical work explaining the factors that influence
the manner in which this risk is shifted ex ante between council and manager via the employment
contract. The more career risk is shifted from manager back on to council, the more freedom the
manager has to make decisions in line with his own professional standards and interests. Any
attempt to understand how the council seeks to align their incentives with managers’ or the
degree to which a manager is willing to make decisions based on their own professional training
and professional interests when those decisions contradict council interests must first seek to
understand the nature of ex ante risk-shifting between council and manager.
Incoming city managers seek some form of protection from the harm that failing to meet
council’s goals or requirements would have on their careers. In other words, when considering
employment, especially employment with a city facing instability, a city manager is risking
tarnishing his reputation and decreasing his future career prospects should he fail to meet
expectations in the current position. It is for this reason that city managers seek some degree of
insurance in the event that they are terminated. There are two key ways that city managers can
71

protect themselves from the risk of termination – political protection and severance protection.
Severance packages can be thought of as a form of financial insurance from the career harm that
would come from failing to meet expectations and from termination. The severance package
negotiated between the principal (the city) and the agent (city manager) reflects underlying
organizational conditions that may increase or decrease the city manager’s likelihood of success
in meeting council expectations.
When a council decides to hire a new manager, the two parties must first come to an
agreement on the manager’s employment contract. This contract typically provides for the duties
and expectations of the manager. In addition, the contract explicitly defines the terms of payment
and benefits, as well as protection from the risk of termination – typically in the form of a
severance package and protection from termination before or after a municipal election. While
the other terms of the contract lay out what the council considers the duties of the city manager
as the primary executive of the city, the severance package and protection from termination can
essentially be thought of as insurance for the city manager in the case of poor performance (Rau
and Xu 2013).
Severance agreements are contracted before turnover in city manager becomes imminent,
and, as laid out in the initial employment contract, provide for cash payments (typically lump
sum payments) should the council choose to involuntarily terminate the city manager without
cause. While these severance agreements do not apply when termination is “for cause,” the
scope of termination “for cause” is explicitly outlined in the contract making termination “for
cause” the more unlikely, and in fact, quite rare, outcome.
15
The other possible outcomes, either

15
Termination for cause is typically only used in the instance of illegal or unethical activity on the part of
the manager.
72

termination without cause or voluntary termination, are far more likely. In the event of
termination without cause, meaning that the city manager is terminated even though he or she is
capable and willing to continue performing his or her duties as city manager, the city manager is
entitled to the severance payment agreed to in his original or amended employment contract. It is
worth noting that in the event of voluntary exit on the part of the city manager, the city is no
longer obligated to payout on the severance package.
The key contribution of this study is the measurement and modeling of city manager
protection from employment risk. I compiled information about the contractual terms of
employment governing the principal-agent relationship that exists between each municipal
council and their appointed city manager. While city manager employment contracts are
available to the public by virtue of freedom of information statutes in the state of California, no
effort has been made to systematically compile the information and provisions contained in these
contracts into one easily accessible data set. As a part of this project, I set out to assemble such a
dataset; a dataset that would contain not only information about the contractual terms of
employment of city managers but would link that information with data relating to the
municipalities in which each manager served. By pulling data from each of these sources, I
created a unique municipal-level data source containing economic, financial and political
conditions in cities and towns throughout the state of California as well as information on the
risk protection provisions granted to the city managers serving in these cities via their
employment contracts.

73

California City Manager Employment Contracts
In order to compile this unique dataset including the severance protections for city
managers in California, I conducted an original data collection process, first submitting a public
records request for a copy of the acting city manager’s original employment contract from each
city in California that employs a professionally appointed town or city manager. While
California freedom of information statutes require each city manager’s employment contract be
made available to the public, there is no existing database which includes all of these contracts.
In order to collect the employment contracts of city managers in the state of California, I created
a comprehensive listing of all 482 cities and townships in the state of California and began by
determining the municipal form of government in each city or town. For each city that did
employ a city manager, I collected the contact information for the city manager, including their
name, phone number, and email address. This information was gathered from the individual city
government websites. A number of cities did not have a website or any contact information
available online and all such cities were excluded from the remaining data collection process.
16
I
collected contact information for 423 municipalities (approximately 88%) operating under the
council-manager form of government, with the help of a research assistant.
Additionally, any city that employs a city administrator, but operates under the mayor-
council form of government with most executive decision making authority vested with the
mayor was excluded from the remaining data collection process. City administrators operate
under a decidedly different principal-agent structure as they typically answer to the strong
Mayor, who serves as the executive of the municipality. While many city administrators perform

16
While this raises the potential for sample bias, the sample selection analysis presented in the next
chapter shows no evidence of a significant sample bias.
74

similar tasks as city managers, they do not operate under the same fundamental institutional
structure, necessitating their exclusion from the sample. The role of a city administrator cannot
be treated as simply an offshoot of council-manager government; city administrators and chief
administrative officers typically exist in mayor-council cities, an entirely different institutional
structure than council-manager cities. For this reason, city administrators and chief
administrative officers were excluded entirely.
While the contracts are considered public record, not all municipalities replied in a timely
manner and a number of municipalities that did respond to the request sent incorrect documents.
A number of cities sent a copy of the general municipal code (which applies to the municipal city
manager in general and is not negotiated for between manager and council) rather than the legal
employment contract specifically governing the terms of employment of the current city
manager. In an attempt to collect as many contracts as possible, I followed up with each of these
cities, first with an email request for the correct documents and then with a phone call to the city
clerk. If the city clerk was unavailable, a voicemail message was left whenever possible. After
following up with these cities, some did send the original employment contract, others claimed
that their city manager did not operate under an employment contract, and still others never
responded.
Another source of missing data was noncompliance with my request for access to the
original contracts. The City Clerk offices of numerous municipalities replied to the freedom of
information request by sending only employment contract amendments, which include only
information relating to changes to the terms of the contract and not the terms of the original
contract in their entirety. After excluding those cities that did not respond, sent the incorrect
document, or sent the employment contract of an interim city manager, I successfully collected
75

332 city manager contracts (representing approximately 78% of all California municipalities
operating under the council-manager form of government).
As I briefly mentioned in the previous paragraph, having collected a portion of all city
manager employment contracts, sample bias was a concern. One of the primary goals of this
study is to better understand the principal-agent relationship between council and manager in
cities of all sizes. Much of the previous research on the council manager form of government has
included data and analysis only on medium or large municipalities, typically excluding cities
with populations under 50,000 people from the analysis (Poister and Streib 1999, Poister and
McGowan 1984, Usher and Cornia 1981, Cope 1992, among others). Although only arbitrary
distinctions, much of the previous work on local government has considered cities of up to
25,000 residents small cities, those between 25,000 and 100,000 as medium sized cities, and
those with more than 100,000 residents as large cities. While many of these studies choose to
focus their analysis on medium or large municipalities, they do so at the expense of the ability to
generalize their results to smaller municipalities. This is particularly problematic as the council-
manager form of government is particularly popular among smaller and medium-sized cities, and
approximately 41 percent of the municipalities in California which operate under the council-
manager form of government have a population of 25,000 residents of fewer. Therefore,
empirical analysis excluding these smaller cities is excluding a sizable portion of all
municipalities operating under the council-manager form of government. The council-manager
form of government is most popular among cities between 10,000 and 100,000 in population,
while the mayor-council form is more popular among very small cities or large cities with
populations greater than 100,000 residents (National League of Cities 2012). I felt it was
76

imperative to include municipalities of all sizes in this analysis in order to present results that can
be generalized to council-manager municipalities throughout the state.
Analysis suggests that being included in the sample is not significantly associated (at the
p < 0.05 level of significance) with municipal population, municipal deficit spending, or
municipal political cohesiveness (see Table 1). In other words, the sample includes cities of all
sizes, economic conditions and political leanings, and is not biased in either direction with
regards to any of these factors. I estimated two probit models of inclusion in the sample
(essentially the likelihood of receiving a compliant response to my public records request). The
first model includes the variable population, which is a measure of municipal population in
thousands. In this model, population approached significance at the p < 0.10 level although it did
not reach significance at the p < 0.10 level. As one of the primary contributions of this analysis
to scholarly knowledge of local government management is including smaller municipalities in
the sample and subsequent analysis, municipalities that are frequently excluded from other
empirical work on local government management, I estimated a second probit model. This model
includes a dichotomous indicator of whether municipal population is below 25,000 people
(1=yes) to test if these small cities are less likely to be included in the sample. This model did
not offer sufficient evidence to reject the null hypothesis that municipalities of 25,000 people or
fewer and cities of 25,000 people or more are equally likely to be included in the sample.

77

Table 1. Sample Selection Analysis - Probit Model (1=inclusion in sample)

Including Population
(in thousands)
Including Binary Measure
Population < 25,000
Population Thousands 0.00
(0.00)

Deficit Spending -0.08
(0.18)
-0.35
(0.18)
Political Cohesiveness -0.25
(0.69)
-0.32
(0.69)
Population Under 25,000
(1=yes)
-0.17
(0.13)
constant 9769.68 23621.10
Log-psuedolikelihood -182.013 -187.67
N 389 389
*** p<.01, ** p<.05, * p<.10. Robust standard errors (in parentheses) are clustered at the county level to
account for any variation within counties due to unobserved variables.

While the first model suggests that the relationship between population and inclusion in
the sample may approach statistical significance at the p < 0.10 level, the second model suggests
that thinking more simplistically about municipal size in a dichotomous manner identifying
either small municipalities (which are typically excluded from empirical analysis) or
medium/large municipalities, both categories of municipalities are represented in the sample.
Additionally, neither deficit spending nor political cohesiveness was significantly related to
inclusion in the sample in either model. This suggests municipal condition and political
ideological cohesiveness do not impact inclusion in the sample. For a full description of the
measurement of each of these variables see the following section on the measurement of
independent variables. The two models estimated to analyze the potential of sample bias suggest
that there is no reason to reject the null hypothesis that the cities included in my analysis are not
significantly different from all California cities.
78

After collecting a total of 332 city manager employment contracts, a research assistant
and I conducted a thorough content analysis of each city manager employment contract, coding
for several key terms of the contract.
17
Each variable drawn from the contracts was numerical in
nature, reducing the risk of coding bias or error. I found several overarching themes among these
employment contracts. Most contracts were 5 to 10 pages in length and contained similar
sections, though frequently varying in order. Most contracts included the term of the agreement
in years, provisions for automatic renewal if applicable, the duties of the city manager, the hours
of work, annual compensation and benefits, termination requirements and removal procedures,
resignation conditions, severance pay, and boilerplate legal provisions on indemnification,
bonding, and notices.

Roles and Responsibilities
After conducting a thorough content analysis of each contract, a few patterns emerged.
First, it is common practice for the section of the contract covering the duties and authority of the
city manager to be relatively vague considering the complex nature of the job itself. Most
contracts simply state that the city manager will serve as the chief executive officer and will be
responsible to the city council for the administration of all municipal affairs, subject to the
provisions of the municipal code. It is typical for a contract to clearly state that the manager will
be expected to administer and enforce the policies enacted by the city council and to impose

17
I developed a list of the key pieces of numerical information to be drawn from the contracts (ex. annual
salary, number of months of severance protection, number of months of political protection). I read and
coded approximately 80 percent of the employment contracts (n=270), and I instructed the research
assistant to read through the contracts and enter these numerical values into the data set for approximately
20 percent of the contracts (n=62). I reviewed ten percent of the contracts coded by the research assistant
to check for inter-rater reliability and found no differences in the research assistants coding results and my
own.
79

rules and regulations as necessary to implement council policy. Most contracts include some
mention of the city manager’s authority with regards to human resources management,
specifically the hiring and firing of municipal employees. For example, one contract stated, “the
city manager shall have the power and shall be required to direct the work of all elective and
appointive city officers and departments that are the concern and responsibility of the city
council, expect those that are directly appointed by or report directly to the city council”
(Murrieta City Manager Employment Contract 2007). This particular employment contract went
further than most in expanding on this authority by stating, “the city manager shall have the
power and shall be required to consolidate or combine offices, positions, departments or units
under the city manager’s jurisdiction” (Murrieta City Manager Employment Contract 2007).
Such a clear statement of the nuances of the authority of the city manager over human resources
decisions and the restructuring of municipal departments is relatively rare in the employment
contracts of city managers in California.

Performance Expectations
Another area worth examining more closely is that of performance review provisions and
incentives linked to performance goals or outcomes. As previously discussed in Chapter Three,
the employment contracts of city managers in California are surprisingly vague with regard to
performance expectations and incentives. The overarching absence of contractual incentives
linked to performance goals or outcomes is surprising given the overall growth in the popularity
of the idea of using pay for performance schemes in the public sector as a means to address the
principal-agent problem and increase efficiency (Binderkrantz and Christensen 2012, Brewer and
Walker 2013, Jensen and Murphy 1990, Weibel, Rose and Osterloh 2010). While much of the
80

empirical attention to pay for performance has been directed at other levels of government or
even other countries, many municipalities have employed performance measurement programs
and pay for performance for their employees (Poister and Streib 1999, Poister and McGowan
1984, Cope 1987, Usher and Cornia 1981, Cope 1992, O’Toole and Stipak 1988). Despite the
wide adoption of performance measurement and pay for performance schemes, there is no
evidence in any of 332 city manager employment contracts analyzed for this study that local
governments are applying performance measurement to their top managers. I did not find a
single contractual statement providing for a defined incentive (monetary compensation or
otherwise) linked to the achievement of a measurement performance outcome. For example, a
typical city manager contract often contains a provision similar to the following two provisions
from the cities of Adelanto and Dana Point.
“Employee shall meet with the City Council in closed session at least annually or as
otherwise determined by the City Council to review and discuss Employee’s
performance, and to consider any adjustments in Employee’s duties and compensation as
deemed appropriate by the City Council” (Adelanto City Manager Contract 2010).

“The City Council shall have the authority to determine the specific duties and functions
that MANAGER shall perform in accordance with the City Council-Manager ordinance
and shall from time to time evaluate MANAGER’S performance, including the means
and manner that he uses to perform his duties. MANAGER agrees to devote all of his
business time, skill, attention, and best efforts to the discharge of the duties and functions
assigned to him by the City Council during his employment” (Dana Point City Manager
Contract 2002).

These two contract provisions and their similarly vague treatment of performance
evaluation are typical of the California city manager employment contracts I obtained and
examined during the course of this project. The vagueness regarding the expectation of the
manager and the flexibility given to the council for determining both the timing and the content
of the performance evaluation is characteristic of the average California city manager contract. If
81

Deno and Mehay (1987) are correct in theorizing that city councils are responsive to the median
voter, whose preferences can change over time, it logically follows that they would have an
incentive to construct vague performance evaluation expectations for the city manager. Doing
otherwise would be to tie their own hands with regard to a preference for certain performance
outcomes relative to others and to declare their priorities for the city manager’s performance in a
publicly available legal contract.
18
As Brian Nakamura, city manager of Chico, said, “ Councils
don’t put metrics in the contract because their targets change frequently and it is hard for them to
hold the manager accountable for clearly defined metrics when they are at the whim of the
electorate. The council doesn’t want to apply metrics because they aren’t sure what they want
and it changes too often” (Nakamura 2014). In addition to a lack of any one over-arching goal,
the degree to which local councils find it advantageous to closely monitor managers varies; some
council may not want to commit to expending the time or money required to monitor managers
carefully enough so as to accurately measure their performance.
While nearly all contracts contained a section directly addressing performance evaluation
and even merit based compensation increases, they were imprecise and left all discretion in the
hands of the council, without any clear definition of measurable performance goals. The
following contractual provisions are typical of the performance review provisions contained in
most city manager employment agreements in California:

18
For example, if a council were to clearly state a goal of “improving public safety” in the manager’s
contract that would remain vague enough to allow some change in preference by the council, but is too
vague with regard to measureable goals to tie incentives to the goal. It is worth noting that even goals this
specific with regard to one policy area are rarely seen in contracts. On the other hand, if the council were
to more clearly state a measurable performance metric such as “reduce violent crime by 10 percent,” the
council is essentially formalizing their commitment to this goal and to the means the manager chooses to
use to achieve this goal. Additionally, stating such clear measurable goals essentially prioritizes those that
are included over those that are not, something that the council has clear incentives to avoid as they have
electoral incentives to be responsive to changes in voter preferences.
82

“The City Council shall conduct a performance evaluation after the first year of
employment and annually thereafter. At the request of Employer or Employee, such
evaluation shall be professionally facilitated by a facilitator approved by the City
Council. Employee will request and schedule the evaluation, as appropriate under City
Council agenda procedures, or as otherwise directed by City Council” (Beverly Hills City
Manager Contract 2009).

“The Mayor and City Council, in closed session, shall review and evaluate the
performance of the Manager at least once year during the term of this agreement. The
review and evaluation shall be completed in accordance with specific criteria developed
jointly by the Mayor and City Council and the Manager. Such review will include a
written statement of the findings provided to the Manager, and the adequate opportunity
for the Manager to discuss this evaluation with the Council. Failure of the Mayor and
City council to review and evaluate the performance of the Manager pursuant to this
section shall not effect the right of the city to terminate his employment. Upon
completion of each such review and evaluation process, the Mayor and City Council shall
determine the amount, if any, of a merit salary adjustment to be provided” (Redondo
Beach City Manager Contract 2004).

These performance evaluation provisions do little to illuminate the specific measures or
outcomes upon which the council will base their performance evaluation of the manager. There
is no suggestion of any particular goals with regard to service provision, human resources
management, financial management, economic development, or any other municipal function.
Rather that seeking to align the incentives of council and manager through the delineation of the
interests of the council and rewards linked to serving those interests on the part of the manager,
these provisions only serve to reiterate the position of the council as principal and of the manager
as agent. The second clause, despite being the more in-depth of the two, simply declares that a
performance review, the content of which is entirely up to the mayor and council, will take place
at least once each year in closed session and that the mayor and council may or may not decide
to provide a merit salary adjustment based on this review. These clauses are absent of the type of
incentive structure proponents of pay for performance schemes would propose including in an
employment contract designed to mitigate the principal-agent problem in local government
83

management. Rather than linking defined incentives to measurable outcomes, these provisions
simply make clear that the city manager is the agent of the council, that he serves at their will,
and that performance evaluations (the content of which is yet to be determined) will occur
annually. The contracts did not define the discretion or authority of the manager, nor the manner
in which the council would judge his behavior and hold him accountable for both his inputs and
outputs. However, despite the lack of measurable performance goals, these evaluation provisions
are typical of city manager employment contracts in California.
Given that California city manager employment contracts do not seek to align manager
and council incentives through the use of monetary rewards linked to measureable performance
outcomes, it was necessary to examine severance as the primary means by which council and
manager seek to align their incentives. Considering that council members are concerned with re-
election and are accountable to the voters, whose preferences can shift, councils want to
maximize their ability to control the actions of the manager, including their ability to terminate
the manager without cause if they so choose. On the other hand, city managers recognize that the
likelihood of meeting council expectations is dependent on the nature of the job, a function of
municipal financial and economic condition and of the political climate within the organization.
The city council’s primary control mechanism over the manager is the power of termination.
Managers recognize and consider the risk of termination in light of the nature of the job and will
negotiate to shift that risk back onto the council via two key contractual provisions.

Risk Protection
Perhaps the most novel contribution of this project is the empirical measurement of city
manager protection from employment risk. Acting as chief executive of the municipality, a
84

typical city manager has numerous job responsibilities, ranging from overseeing administrative
operations, implementing council policy, and preparing the budget for council consideration.
City managers balance these tasks while serving at the pleasure of a city council made up of
multiple elected officials. Most city councils in California consist of an odd number of
representatives, typically five but sometimes more, who are popularly elected to their positions.
As previously discussed in Chapter Three, the nature of the principal-agent relationship between
council and manager is such that the manager is susceptible to varying degrees of career risk
should the council choose to terminate the manager. Managers are acutely aware of this risk, and
as a result of the turnover rate in local government management, they are conscious of
maintaining their professional mobility and reputation.
There are a number of ways in which managers can seek to protect themselves from the
career risk of termination. For the purposes of this study I examine two means of risk protection,
one financial and the other political. I argue that the primary means of risk protection for city
managers is severance protection, which is contractual, determined ex ante, and financial in
nature. Severance protection is negotiated for prior to the manager agreeing to take on any given
position. Severance packages included in city manager employment contracts provide managers
with a financial lump sum payout in the event that the council terminates the manager without
cause. The conditions under which termination would be considered “for cause” vary from
contract to contract, but in general, termination for cause applies in situations of willful
misconduct or criminal behavior on the part of the city manager. For example, some typical
circumstances under which termination is considered “for cause” and the severance provision
would be voided are: improper or unauthorized use of city property, falsification of city reports
or records, being under the influence of drugs or alcohol while conducting city business,
85

conviction of a crime or violation of state law, or neglect of duty. While nearly all city manager
employment contracts in California contain a set of conditions under which the manager may be
terminated with cause, this scenario is actually rare in practice as most city manager turnover is a
result of voluntary resignation or termination without cause (Kammerer at al. 1962, DeHoog and
Whitaker 1990). As such, severance agreements are an important piece of the protection from
career risk granted to city managers. Severance clauses are essentially career and financial
insurance policies for city managers, as they give the manager a certain number of months of
continued salary and benefits during which time they can seek employment with another
organization. Severance clauses, when activated, impose a direct cost on city governments and
serve as a means of shifting the risk that stems from termination from the city manager onto the
city council (Feiock and Stream 1998).
Empirical analysis has presented evidence that a turbulent municipal environment
negatively impacts city manager tenure and that the presence of an employment agreement plays
an intervening role in terms of how municipal conditions impact tenure (Feicok and Stream
1998). While this study presents evidence that employment agreements do impact the career
mobility and career paths of city managers, the content of the agreements was not examined,
only their existence. However, the content of these employment contracts is the key institutional
structure which defines the cost of termination on the part of the council. As Feicok and Stream
(1998) point out, “these [severance] benefits provide flexibility to both the managers and the
council and influence their stake in either the continuation or termination of the existing
employment situation. Councils that incur these costs might be less likely to terminate their
manager because it is seen as costly to force the manager out of a job” (Feicok and Stream 1998,
122). While this assertion is a key assumption of my own argument, in order to understand the
86

institutional relationship between council and manager it is imperative to understand the
variation in these employment agreements rather than treating them all alike. This distinction is
particularly important given that my analysis of city manager employment agreements in
California suggests that the number of months of severance pay and the total dollar amount of
severance pay due city managers in the event of termination without cause varies significantly
19

from manager to manager.
City councils and city managers, as party to a contractual employment agreement,
typically use a provision for severance compensation in the employment contract to shift risk
between the council and the manager and to align the incentives of principal and agent.
Severance compensation serves as risk protection for city managers who are concerned with their
professional reputation and mobility, by shifting a degree of the risk of termination onto the city
council, who must make a lump sum payment to the manager should he be terminated without
cause. However, the amount of severance protection granted to city managers varies greatly
throughout the state of California, up to the legally allowed maximum of 18 months of severance
compensation. In the sample of city manager contracts obtained for this study, 7.56% of city
managers received zero months of severance protection, the remaining 92.44% of city managers
had between 1 and 18 months of severance protection, with a mean of 7.4 months of severance
protection.
The typical contract describes severance as a number of months of salary that will be paid
to the manager in the event of termination without cause during the term of the contract. The

19
After examining and coding 332 employment contracts for current city managers in the state of
California, my analysis suggests that a large majority of city manager contracts (more than 90 percent)
include severance protections from 1 month to 18 months with an average total dollar value of $116,000
but going all the way up to $290,000.
87

following provisions are typical of a California city manager contract, although of course the
number of months of severance varies.
“In the event Stewart’s employment is terminated by the City Council during such time
that Stewart is willing and able to perform the City Manager’s duties under this
Agreement, then in that event the City agrees to pay Stewart a lump sum cash payment
equal to six-months’ (6-months) base salary then in effect as provided in 4.A(1) above”
(Bellflower City Manager Employment Contract 2012).

“In the event that Employee and this Agreement are terminated without cause, City
agrees to provide Employee with severance pay in a lump sum cash payment equal to (6)
months’ base salary and benefits less deductions required by law, and subject to the
following condition. Payment of severance pay shall be conditioned upon Employee
executing a release and waiver of any and all claims against City, its officers, employees
and agents, arising out of his employment with City, and termination thereof, in such
form as may reasonably be required by City. Severance pay shall not be payable unless
and until Employee executes such release and waiver and until expiration of all waiver
and revocation rights as provided by law at the time of termination of Employee’s
employment and this agreement” (Escalon City Manager Employment Contract 2013).

Most severance provisions in city manager employment contracts are relatively straightforward.
They define the number of months of salary that will be paid to the manager in the event of
termination without cause.
The other key source of career risk for city managers is political risk resulting from the
nature of their principal-agent relationship with the city council, a body of elected officials
chosen by the voters. While severance protection offers financial protection from career risk, it
does not place any restrictions on the city council’s ability to terminate the manager. However,
the contractual provision stipulating that a council may not terminate a manager without cause
for a certain number of days before or after an election does tie the hands of the council in the
wake of political change. For example, the following is a typical provision of protection from
termination before or after election.
88

“Employee shall serve at the will and pleasure of the City Council and, notwithstanding
provisions in the municipal code to the contrary, may be terminated at any time with or
without cause upon a majority vote of the City Council; provided, however, that this
Agreement may not be terminated within thirty (30) days prior to a general or special city
election, or ninety (90) days following a general or special city election, at which one or
more Council members are elected, except for cause” (Adelanto City Manager
Employment Contract 2010).

Protection from termination before or after an election is typically written into the employment
contract in this manner - as a number of days before or after a municipal election during which
the city manager cannot be terminated without cause. This type of protection from career risk is
distinct from severance protection in several ways. Political protection does not provide any
financial protection from the manager; in fact, it is of no use to him after termination. Rather,
political protection stipulates a period of time after the potential change of council membership
for the manager to survey and adjust to the new political environment and the new members of
the council. Many municipalities who do negotiate some sort of political protection with the
manager mention that such protection is designed to give the council and manager a defined
number of days or months to develop a working relationship before the council may decide to
terminate the relationship. This type of protection can be particularly valuable to managers
entering municipalities in which political turmoil is a primary source of career risk.
The city manager employment contracts I collected and coded all shared some similar
patterns. Each contract defined the terms of employment between council and manager (all at-
will), the compensation and benefit package, the provisions for performance evaluation, and risk
protection provisions shifting career risk from the manager to the council. While these contracts
were relatively similar in some regards (for example, a lack of clearly defined performance
metrics or pay for performance incentives) they exhibited wide variation in other terms
89

(compensation, benefits, length of political protection, and severance protection). The systematic
analysis of these contracts not only advances our understanding of the institutional structure
which defines the environment in which city managers operate, but, through the coding of such
important terms as severance pay and political protection allows for empirical analysis to test the
key hypotheses derived from my theory of risk shifting between manager and council.
90

CHAPTER SIX: EMPIRICAL ANALYSIS
In the last chapter, I discussed the collection and coding of city manager employment
contracts throughout the state of California, the source of the data used to measure my key
dependent variables: severance protection and political protection. In addition to the city
manager employment contracts, I rely on several data sources to examine the relationship
between municipal condition and city manager risk protection. The various sources of data used
in this study are described in detail below, but all are collected at the individual (city manager) or
municipal level. I constructed a unique dataset comprised of information on all California
municipalities and a sample of California city managers based on publicly available government
data. Municipal level data were collected from a variety of sources including the U.S. Census
Bureau, the U.S. Bureau of Labor Statistics, the California State Controller’s Office, the RAND
Local Government Database, and the California Elections Data Archive (CEDA)
20
. The
municipal data collected from these organizations and agencies is supplemented with additional
data on individual city managers (each representing a single municipality) collected from the
International City/County Management Association (ICMA) People and Places Knowledge
Network.

Measuring Risk Protection
In order to measure city manager protection from employment risk, I focused on two key
contractual provisions as indicators of the degree of risk protection negotiated for between the
manager and the council. These provisions shift risk back onto the council by making it more

20
The CEDA is a joint project of the Center for California Studies, the Institute for Social Research of
California State University, Sacramento, and the California Secretary of State.
91

difficult to terminate the manager, essentially giving the manager varying degrees of latitude to
make decisions that may not be completely aligned with council interests without fear of losing
his or her job. The first way in which the manager can shift risk back onto the council and the
first dependent variable is severance protection, a measure of the financial protection from
termination afforded the city manager in his employment contract if terminated without cause.
The total financial payout the manager would receive in the event of termination without cause is
calculated by multiplying the number of months severance times the manager’s monthly salary.
Severance protection varies from $0 to just over $331,025 with a mean of $117,922 (See Table
2). A higher value indicates that the city manager has a greater degree of risk protection.
Table 2. Summary Statistics

Mean St. Dev. Min. Max.
Severance (2012 dollars) 117,922.20 69541.89 0 331257.60
Unemployment Change 0.04 1.39 -2.5 4.9
Deficit Spending (percent of rev) 4.79 20.07 -90.88 92.93
Political Cohesiveness 0.14 0.10 0 0.45
Local Direct Democracy 0.39 0.97 0 11
Electoral Competitiveness 0.51 0.18 0 0.86
Population (thousands) 58.01 81.29 0.36 931.34
Gender (1=male) 0.83 0.37 0 1
School Rank 54.42 22.16 1.30 98.90
Years Experience 10.32 8.17 1 38
MPA Education (1=yes) 0.46 0.49 0 1
Unemployment Change is the change in the municipal unemployment rate from the year prior to the manager
signing contract to the year of signing contract.
Deficit Spending is based on expenditure and revenue data from 2002 (for all CM who signed contract from 2003 to
2008) and from 2008 (for all CM who signed contract from 2008 to 2013).
Political Cohesiveness Data from 2008 presidential election.
Political Engagement and Electoral Competitiveness based on most recent municipal election prior to CM contract.
Population is for the year CM signed contract.
School Rank is based on 2006 – 2012 data.
92

The second key dependent variable, political protection, is a binary measure of whether
or not the contract stipulates that the city manager cannot be terminated without cause for a
period of time before or after a local election. City managers face career risk due to the principal-
agent relationship they engage in with the city council, a body of elected officials chosen by the
voters. While severance protection offers financial protection from career risk, it does not place
any restrictions on the city council’s ability to terminate the manager. However, the contractual
provision stipulating that a council may not terminate a manager without cause for a certain
number of days before or after an election does tie the hands of the council. For example, the
following is a typical provision of protection from termination before or after election.
“Employee shall serve at the will and pleasure of the City Council and, notwithstanding
provisions in the municipal code to the contrary, may be terminated at any time with or
without cause upon a majority vote of the City Council; provided, however, that this
Agreement may not be terminated within thirty (30) days prior to a general or special city
election, or ninety (90) days following a general or special city election, at which one or
more Council members are elected, except for cause” (Adelanto City Manager
Employment Contract 2010).

Protection from termination before or after an election is typically written into the employment
contract in this manner - as a number of days before or after a municipal election during which
the city manager cannot be terminated without cause. While I originally recorded both the
number of days before and the number of days after an election for which a given manager is
protected from termination, because the nature of the contracts is to fall into 30-day intervals,
there was little variation (in terms of total number of days) among the city manager contracts
included in the sample. The number of days of political protection ranged from 0 to 360 days,
with a mean of 41 days, and a standard deviation of 67 days. Despite the seemingly continuous
nature of this variable, upon closer inspection, it is clear that the total number of days of political
93

protection is not normally distributed, as approximately 68% of city managers receive no
political protection, and those who do receive protection are grouped into just a few categories
(77% of city managers with political protection have either 90 days or 180 days of protection).
Because of these limitations in the data, and the bimodal, non-normal distribution of the total
number of days of political protection, I chose to employ a binary measure of whether or not a
city manager received political protection of any number of days. This variable was coded with a
value of “1” if the city manager received any number of days of protection from termination
before or after a municipal election and a “0” if the manager received no such political
protection. Among the 332 contracts examined, approximately 28.09% of managers receive
some degree of political protection in their employment contract (See Table 1).

Measuring Municipal Condition
The central argument of this analysis requires appropriate measures of municipal
condition, which I break down into three categories: economic, financial, and political. Each of
these three factors is linked to the typical tasks of city managers who prepare municipal budgets
for council review, implement municipal budgets after council passage, play a leading role in
economic development, and manage a principal-agent relationship with numerous popularly
elected officials, who have the authority to fire the manager with or without cause (ICMA 2006).
Economic conditions can play an important role in the allocation of local government
resources. Research has suggested that municipalities with more poverty and unemployment
must spend more on public welfare, housing, and other types of redistribution which necessitates
a decrease in spending on such things as economic development or other city services like parks
and recreation or libraries (Hajnal and Troustine 2010). The presence of poor economic
94

conditions can not only lead to decreased revenues overall, but can also tie both the council and
manager’s hands with regards to how limited resources must be spent. In order to capture the
influence of economic condition at the municipal level on manager contractual protection from
risk, I include a measure of the change in unemployment rate in the municipality from the year
before to the year in which a manager signed his contract. This measure is based on information
from the U.S. Bureau of Labor Statistics and is collected at the municipal level. Unemployment
Change varies from -4.4 percent to 4.9 percent with higher numbers indicating a worse economic
condition. In other words, in a city with a score of 4.9 on this measure, the unemployment rate
increased by 4.9 percentage points during the year before the city manager signed his or her
employment contract.
Municipal financial condition is measured as the percent by which total municipal
expenditures exceed total municipal revenues. This measure is calculated as 1 minus the ratio of
total municipal expenditures to total municipal revenues times 100. For example, in a
municipality with $11,000,000 in expenditures and $10,000,000 in revenues, expenditures
exceed revenues by 10 percent. Therefore, increasing values of this measure indicate an
increasingly unstable financial condition. It is important to note that prior studies have used
various measures of local financial condition and even ICMA defines numerous appropriate
measures including total revenues, total expenditures, unreserved fund balance, appropriated
fund balance, gross expenditures, gross revenues, salary and fringe benefit obligations, cash and
investments, tax levy, tax limit, operating position, debt structure, unfunded liabilities, and
condition of capital plant.
21

21
These metrics are just a sampling of the ICMA suggested measures analysts interested in local
government finance should consider when seeking to evaluate the overall fiscal health of a municipality.
95

There is no one measure of financial standing that offers a complete picture as to the
fiscal health of the municipality. According to Honadle, Cigler, and Costa (2003) “there are so
many disparate factors affecting the fiscal health or condition of local governments that it is a
challenge for local officials to sort them out and make sense of them from their perspectives as
managers” (5). Ladd and Yinger (1989) present a “need-capacity gap” framework which
considers municipal expenditure needs in relation to revenue-raising capacity as a measure of
local fiscal health. Brown (1993) and Maher and Nollenberger (2009) use a ten-point set of
indicators of financial condition which includes total revenues per capita, intergovernmental
revenues as a percentage of total revenues, property tax revenues as a percentage of total
revenues, expenditures per capita, operating surplus or deficit as a percentage of total revenues,
general fund balance as a percentage of general fund revenues, enterprise funds, long term debt,
debt service as a percent of operating revenues, and postemployment benefit assets. While such a
detailed set of indicators provides a nuanced picture of local government fiscal health, many of
these measures are not readily available for cities of all sizes and would require thorough
examination of individual budgets from multiple years to compile.
For the purposes of this study, due to the city manager’s integral role in the preparation of
the budget and implementation of the budget, I focused on the budget deficit as an indicator of
financial condition for two reasons. First, a city manager is likely to be able to easily ascertain
and understand this information as a potential indicator of poor financial condition. Secondly, a
manager entering a position in a city operating under a deficit is likely to have to implement
measures to reduce the deficit, particularly due to California state laws requiring local
governments to balance their budgets. Despite the balanced budget requirement in California, the
ability of local public officials to accurately track expenditures, and forecast revenue makes
96

budgeting an imprecise science at best. A municipality is in the best position when it can
accurately forecast so as to avoid deficit spending and ideally to end each fiscal year with some
degree of a surplus so as to protect from the potential shortfalls that could occur in any given
year. The measures necessary to balance the budget may not be politically popular and could
present career risk to the manager, who must balance adherence to their professional training and
norms with the reality of accountability to an elected city council facing political pressures.
Scholarship on local government finance has previously used bond ratings or municipal
interest rates on bonds as a measure of financial health as a result of their ability to capture
numerous indicators in one easily available measure (Honadle and Lloyd-Jones 1998). As
Gianakis and McCue (1999) explain, “A [bond] rating downgrade can severely impact the
financial position of a local government and could potentially place the entity into receivership”
(155). Bond ratings are a particularly useful proxy for municipal fiscal health because in one
measurement they capture a variety of factors that impact fiscal health. For example, “[Bond-
Rating] agencies look for evidence of professionalism and financial responsibility in account
structures, reporting procedures, budgetary and financial planning processes, and financial
control mechanisms” (Gianakis and McCue 1999, 155). While these measures do offer a
reasonable proxy for financial health at the local level, they were not practical to use in this
analysis as many smaller cities do not take out any bonds and thus bond-rating data is not
available for these cities.
22
A key aim of this project was to better understand the relationship
between municipal condition and city manager risk protection for city managers in municipalities

22
I did collect bond-rating data for each municipality in California for the 8 years prior to the city
manager accepting employment, but nearly 45% of cities in California had not issued any municipal
bonds in that time period. Further analysis revealed that municipal population was significantly associated
with whether or not a city had issued any bonds in the eight years prior to the city manager signing his or
her contract.
97

of all sizes through results which can be generalized, using bond-rating data would have
excluded smaller towns and cities from the analysis, fundamentally altering the nature of the
project.

Political Constraints
In addition to economic and financial concerns, the central argument of this project
requires a thorough and appropriate set of measures of political conditions that could present
career risk to city managers. I develop and use several measures of political conditions within a
municipality. The nature of municipal elections is an important consideration. Some political
scientists have suggested that the electorate’s choices on Election Day are a reflection of their
evaluation of the past performance of politicians (Key 1966), but little of the work on
determining the predictors of election outcomes has focused on local elections.
23
While scholars
have not offered a clear picture of what factors drive local election outcomes, they have
suggested, “competitive elections create a relationship of formal accountability between policy
makers and citizens” (Ashworth 2012, 183). An incumbent running for reelection in a
competitive political environment faces an “incentive to choose a policy that is ex ante popular
with voters, even though [the incumbent] knows it is the wrong decision, given all of the
information” (Ashworth 2012, 189). The implication of this work is that incumbents running for
office are more likely to pander to voters and behave in a way that may not produce policies that
serve the long-term interests of the voters (Ashworth 2012). The degree to which council
elections are competitive can directly impact the likelihood of change on the council and the

23
According to Berry and Howell (2007) who conducted an analysis of all 212 articles on elections
published between 1980 and 2000 in five top political science journals, less than 1 percent examined local
elections.
98

likelihood of politicians voting for policies designed to improve the long-term condition of the
municipality. For the city manager, a professionally trained administrator interested in his own
professional mobility and thus the improvement of overall municipal condition, an increase in
the competitiveness of elections can mean an increase in the divergence of his own preferences
from the council and an increase in the likelihood of a change in the makeup of the council. As
such, electoral competitiveness is a key source of career risk for managers.
In this analysis, electoral competitiveness is a measure of the competitiveness of local
elections for city council in the most recent election prior to the city manager accepting the
position. Electoral Competitiveness is a continuous variable based on the ratio of the number of
available council seats in the prior election to the total number of candidates for those seats. For
example a city with 3 available seats and 4 candidates would be coded a 0.75 (out of 1) on this
scale, meaning for any given candidate, holding all other variables equal, there is a 0.75
probability of winning a seat on the council. In contrast, in a city with 3 available seats and 9
candidates, the ratio would be 0.33, meaning that for any given candidate there is only a 0.33
probability of winning a seat on the council. The final step in creating this variable was to
subtract the probability of winning (ratio of available seats to candidates) from one such that
increasing values of this measure are associated with increasing political stability and decreasing
electoral competitiveness within a municipality. In other words, as the number of candidates
running per open seat on the council increases the value of electoral competitiveness increases as
well.
The second measure of political condition in a municipality is political cohesiveness,
which as per my prior discussion is not to be confused with political polarization. While Fiorina
and Abrams (2008) define polarization as “the simultaneous presence of opposing or conflicting
99

principles, tendencies, or points of view… an intuitive notion of polarization is as a bimodal
distribution of observations” (566), my concept of political cohesiveness does not consider the
bimodal distribution of preferences, but rather the homogeneity of preferences among voters.
Political cohesiveness is a measure of the like-mindedness of voters, not the ideological distance
from one key group to another. Political cohesiveness is based on the percentage of the
municipal population that voted for President Obama in the 2008 presidential election. The
political cohesiveness measure is calculated by taking the absolute value of the difference
between the proportion of the municipal population that voted for Obama in 2008 and 0.50.
Increasing values on this variable indicate increasing political cohesiveness or ideological
homogeneity among municipal residents. For example, in the city of Berkeley, 92% of residents
voted for Obama in 2008, meaning Berkeley’s score on this measure (the absolute value of the
difference between 0.92 and 0.50) would be 0.42 (out of 0.50) indicating a high level of political
cohesiveness among municipal voters. On the opposite end of the spectrum, in the city of
Auburn, 50% of residents voted for Obama in 2008, meaning Auburn’s score on this measure
(the absolute value of the difference between 0.50 and 0.50) would be 0.0 (out of 0.50) indicating
a low level of political cohesiveness and a low degree of ideological homogeneity among
municipal voters.
Political cohesion is an important part of the consideration of political risk because the
more likeminded the electorate, the more pressure they are able to impose on politicians to
support the policies they favor; as the percentage of voters who share the same ideology grows,
the size and relative influence (in terms of the number votes) of the group that counters their
ideology lessens, allowing the majority group to impose increasing electoral pressure on the
politician. For example, in a city in which voters are relatively evenly split in their political
100

ideologies, one groups’ preferences balance the others’ and politicians should be responsive to
the median voter by supporting more moderate policies (Dahl 1961). On the other hand, in a
municipality in which voters are ideologically homogeneous, regardless of if they are
conservative or liberal, the median voter is no longer moderate, but rather located further at one
end of the ideological spectrum and the policies supported by politicians are expected to be more
extreme as well. This dynamic creates risk for a manager who was professionally trained to
competently manage the city so as to improve the overall municipal condition and provide long-
term stability for the community. Such a manager is often not interested in pleasing ideologically
extreme voters, but rather, has a professional interest in the passing and implementation of
policies designed to improve the overall municipal condition. As the electorate becomes more
politically cohesive, they are, through the council, able to place more constraints on the behavior
of the manager, increasing his career risk.
24

The third measure of political condition at the municipal level is local direct democracy,
a measure designed to capture the engagement of municipal citizens in the political process at the
local level. Local direct democracy is a measure of the total number of ballot initiatives or recalls
included on the municipal ballot in the most recent election (prior to the city manager entering
into the contract with the municipality) and ranges from 0 to 11. For example, a city with 0
citizen initiated ballot measures would have a low level of local direct democracy, while the

24

It is worth noting that the importance of the municipal level political cohesiveness among voters with
regard to municipal election outcomes may vary depending on the election format. In municipalities
employing district level voting for council seats the political cohesiveness of the entire electorate may be
less important than in a city employing at large elections for council seats. This variable was not included
for two reasons. The municipal level election data used in this study did not specify whether municipal
elections were at-large or district. Additionally, in the wake of Voting Rights Act and the California
Voting Rights Act, many California cities have changed their system of voting for council seats from one
of at-large elections to one of district level elections in order to decrease the incidence of minority group
disenfranchisement. These changes have happened at various points in time and some of these changes
resulted from council policy while others followed legal challenges brought forward in the courts.

101

municipality with 11 ballot measures present on the municipal ballot in the prior election would
be considered to have a high degree of local direct democracy. This is an important consideration
with regard to risk shifting between council and manager, as more initiatives at the local level
means that there is a greater strength of direct democracy at the local level and citizens are
aware, engaged, and likely to make their preferences known to council members. While there are
certainly other ways that citizens can be involved in local politics, the action of placing a citizen
written initiative on the local ballot is one of the strongest and most direct ways that citizens can
influence the local policy making process as it essentially takes policy making power away from
local elected politicians. This ability to craft local policy and circumvent the authority of the
council gives citizens a degree of control over local political behavior, and this degree of control
increases as citizens become more engaged and exhibit a willingness to go through the process
required to place initiatives on the ballot.
Control Variables
In order to parse out the specific effects of municipal condition on city manager risk
protection, I control for other factors that previous literature suggests could influence city
manager employment contracts. First, there is some empirical evidence in the sociology and
psychology literature of gender differences in negotiation outcomes (Stuhlmacher and Walters
1999, Kimmel et al 1980, O’Shea and Bush 2002). The explanations for these gender differences
in negotiation include variation in professional aspiration, trust and interpersonal behavior
(Watson 1994, Meeker and Weitzel-O’Neill 1977). Fully understanding the factors which impact
the gender gap in negotiation outcomes is beyond the scope of this project. However, since
previous empirical research has offered evidence of gender gaps in negotiation outcomes, it is
necessary to control for the role of gender in a model of contractual risk protection negotiated for
102

between city managers and city council members. I include a dichotomous variable, gender,
which is given a value of “1” if the city manager is male and a value of “0” if a city manager is
female. Approximately 84 percent of city managers included in the sample used in this study are
male (See Table 1), which reflects a historical trend of local government management as a
profession dominated by males (Svara 2003).
When thinking about the city manager position and the terms of the employment contract
as a portfolio, or a collection of assets, benefits, and responsibilities held by the city manager,
each piece of the contract can add or subtract from the sum amount of compensation, in the form
of both monetary and non-monetary compensation. One key municipal attribute that contributes
non-monetary value to the compensation of the city manager is the quality of the schools within
the municipality. Empirical research has shown that public school quality is a key determinant of
housing prices (Kane, Staiger and Samms 2003). School quality can also be an indicator of an
overall desirable place to live as “good schools usually come bundled with other neighborhood
qualities such as proximity to employment, shopping and recreational conveniences, and
neighborhood peers” (Kane, Reigg and Staiger 2006, 1).
In addition to their role as indicators of generally desirable places to live, school quality
also directly affects managers with children. Most managers are under pressure (if not explicitly
required in their contracts) to live within the limits of their municipality, meaning their children
will either attend the local public schools in that municipality or the manager will pay the cost of
sending the children to private school. The presence of high quality schools in the municipality
in which the manager will work adds value to the compensation package and can be thought of
as an asset by the manager. As such, the manager would want to hedge against the divestment of
this asset, or the loss of access to high quality schools that would accompany termination. As the
103

school rankings in a city where the manager is accepting employment increase, so does the
probability that upon termination the manager will not be able to secure employment in another
municipality with a better school ranking. In other words, the higher the quality of the schools in
the city, the higher the risk of divestment of that piece of his overall compensation package upon
termination. Consider a manager accepting a position in a city ranked in the 90
th
percentile for
school quality. Upon termination, only ten percent of municipalities in the state would be able to
offer higher quality schools, and if the manager seeks to maintain or exceed the current school
quality, his employment opportunities would be limited to those municipalities. School quality is
effectively a piece of the overall compensation package, because it is a private benefit, and is an
indicator of overall neighborhood desirability (Kane, Reigg, and Staiger 2006). As such, the
higher the quality of the schools in a municipality, the more the manager seeks to strengthen his
protection from termination and the potential divestment of the value of that piece of his
compensation package.
To control for the impact of school quality on manager protection from termination, I
include the variable school rank, a measure of each municipality’s school ranking as a percentile
compared to all other schools in California. This rating is based on data from the California
Department of Education and is published by School Digger.
25
The variable school rank varies
from 0 to 100 and indicates an individual city’s percentile rank in school quality compared to all

25
School Digger is a website (www.schooldigger.com) which compiles test scores, rankings, and
student/teacher ratios for schools at either the district or municipal level. The website provides a measure
of school quality that is particularly useful for this study in that it measures quality at the municipal level
and all rankings are relative to other schools in California. Rather than using raw test scores, the measure
provided by School Digger compares the quality of one school to the others, for a relative measure of
school quality.
104

other municipalities in California.
26
The rankings are based on data compiled from all schools
located within the municipality that are designated as “regular” by the U.S. Department of
Education with available test scores for both math and English. The average math score across
all grades in all schools within the municipality is added to the average English score across all
grades in all schools within the municipality for a combined school quality score. The
municipality with the highest score (out of all municipalities in California) is given a rank of 1,
and so on and so forth. The school rank measure used in this analysis is the percentile rank of a
given municipality’s school quality ranking, as published by School Digger, among all California
municipalities. Ranging from 0 to 100, this measure represents the percentage of California
municipalities that fall below the given municipality in school rank, meaning that higher scores
indicate higher school quality. This measure offers an appropriate way to gauge school quality at
the municipal level, which has frequently presented a challenge to researchers due to the fact that
many school districts are not clearly organized around the geographic boundaries of municipal
governments. For example, during the 2011 to 2012 school year, there were 10,296 public
schools in California organized into 1,043 different school districts (California Department of
Education 2013). The advantage of the school rank measure used in this analysis is that it
measures school quality at the municipal level, regardless of the structure of the larger school
district within which the municipal schools operate.
City manager skill and professional qualifications could also impact negotiation
outcomes, either via improved negotiation skill or stronger negotiating position. Most simply,
executives who have spent more time in the role of city manager in the past have more

26

A regression analysis offers no evidence that school rank is significantly related to economic
conditions, financial conditions, or population at the municipal level.

105

experience negotiating with city councils and working under city councils. This prior experience
could give more seasoned city managers an advantage over less experienced city managers in
terms of negotiation skill since they have already gone through previous negotiations with other
city councils. In addition, executives with more experience have a stronger bargaining position
as they have evidence of their value as a manager (Rau and Xu 2013, Lys and Sletten 2006,
Rusticus 2006).
At the heart of the principal-agent problem in local government management is the idea
that there is an information asymmetry between the manager and the council, the latter not
having full information on the quality and quantity of the effort the manager will expend once in
the position. More experienced managers have a more thorough record of prior work on which
the council can base its assessment. While this linkage has not been directly studied in the
previous literature on local government management, research on private sector executive
management has shown that among employees starting new positions, individuals with more
prior work experience are more likely to successfully negotiate an increase in salary (O’Shea and
Bush 2002). Additionally, CEOs with higher bargaining power, based on their previous
experience, are more likely to receive severance agreements (Rusticus 2006). Empirical evidence
from an empirical setting offers support for the theory that more experienced negotiators claim a
larger share of resources from the party they are bargaining with than less experienced
negotiators (Thompson 1990). The previous empirical work suggests that prior work experience
can impact general contract negotiations and severance negotiations, specifically. Scholars have
theorized that the key relationship between experience or skill and severance is that with the
hiring of a relatively poorly qualified candidate, the more likely it is for a higher quality manager
to become available to the organization in the future. The implications being that the less
106

qualified an incoming manager is, the more inclined a board will be to terminate that manager in
the future so as to employ a higher quality executive (Almazan and Suarez 2003). This
relationship is such that the less experienced manager will seek a larger severance package to
protect from the increased risk of termination associated with their lack of prior experience.
Additionally, there is support for the notion that more experienced managers are more likely to
have solidified their professional reputation, meaning they are less likely to suffer harm to their
professional reputation in the event of termination. With each additional year of experience, the
potential negative impact of future termination on professional reputation decreases, suggesting
that the need for severance protection decreases as well.
In order to control for the impact of skill and qualifications on risk protection, I include
two variables in the model: experience and education.
27
Both the experience and education data
included in this analysis are gathered from the International City County Management
Association (ICMA) People and Places Knowledge Network, an online directory of members of
the ICMA. This was a logical source to gather such information as ICMA has become the most
prominent professional organization for city management professionals and almost all city
managers have chosen to be members. Many city manager employment contracts go so far as to
stipulate that the manager is expected to abide by the ICMA Code of Ethics and that the
municipality will pay for the manager’s professional ICMA dues and travel to the ICMA

27
In order to gather information on the experience and education of city managers, I initially sought to
gather city manager resumes. The public records requests sent to each municipality in California included
a request for a copy of the current city manager’s resume. While a small number of cities supplied this
document, most cities were not willing to provide this document as the city manager’s resume is not
considered public record. The California State Supreme Court has ruled that professional resumes contain
information of a personal and private nature, and as such, public agencies are not required by law to
provide the resumes of public employees under public records laws. While the full resumes of each city
manager would have been the preferred source of information on the experience and educational
background of each manager, they were not a viable option for data collection. I turned to information
self reported by city managers through the ICMA.
107

conference. However, the background information on each manager in the People and Places
Knowledge Network on the ICMA website is all collected voluntarily. A manager inputs this
information into his or her own profile and can choose what information to include and what
information not to include. As such, I was only able to collect experience and education
information on about 59 percent of the city managers in my sample. Having education and
experience information on only 160 managers was problematic, as this was far too low a
proportion of the total sample to use any advanced statistical techniques to estimate the missing
values. Therefore, the variables related to background experience and education described below
are included in a separate model of city manager severance protection to test for their impact and
to confirm the validity of the initial model, but these variables are not included in the primary
model as it would have resulted in the loss of more than 40% of the cases.
The first measure, experience, is a measure of the total number of years an incoming city
manager has served as a full city manager in previous positions. This measure was collected for
each city manager who reported their employment history on the ICMA website and is a simple
continuous measure of the total number of years the manager reported having served as a full
city manager prior to taking their current position. This measure excludes years as assistant city
manager but does include years as interim city manager in which the manager was performing
the same essential duties as a full city manager. This measure ranges from 1 year to 38 years
with a mean of 10.3 years (see Table 1).
The second measure, MPA Education, is a dichotomous measure of whether or not the
city manager has earned a Master of Public Administration (MPA) degree. This measure is
coded a “1” if the city manager does have an MPA degree and a “0” if otherwise. This measure
was also collected based on the self-reported education history provided by some members of
108

ICMA on the People and Places Knowledge Network. While there are several degrees that a city
manager could have, including other Masters degrees, the MPA is the standard and most
specialized graduate degree directly related to local government management. Accredited MPA
programs are governed by the National Association for Schools of Public Affairs and
Administration (NASPAA) which requires curriculum to “ensure excellence in education and
training for public service and to promote the ideal of public service” (NASPAA 2009).
According to NASPAA standards, MPA programs should focus primarily on preparing students
to be leaders and managers in the professions of public affairs and public administration with a
focus on public service values such as accountability, transparency, professionalism, efficiency,
objectivity, equity, fairness, and ethics. While other graduate programs can provide managers
with knowledge that is valuable in the profession of local government management, the MPA is
the most common graduate degree that focuses specifically on public sector management. In
order to test for the impact of holding an MPA degree on managerial risk protection, the variable
MPA Education is included in the second model of severance protection as well. Again, this
variable is not included in the primary model as educational information was only available for
149 city managers, leaving far too many missing cases to attempt to impute the missing values.

109

Methods
28

To test the impact of municipal level factors on city manager severance protection, I
employ a tobit regression model. The tobit regression model “assumes that the dependent
variable has a number of its values clustered at a limiting value, usually zero” (McDonald and
Moffitt 1980, 318). In this context, city managers can negotiate for a severance amount of
anywhere from zero dollars to an unlimited dollar amount (theoretically). While the state of
California has limited the total months of severance that a city manager can receive at 18
months, because severance is a function of both the number of months of severance and the
monthly compensation of the manager, the upper bound on severance pay is limited only so
much as compensation is restricted. At writing, there were no limits on the compensation of city
managers in the state of California, although given recent local government management
scandals in the state, this policy could be forthcoming.
While severance protection has no upper bound, it is limited by a lower bound of zero,
meaning that severance pay cannot be a negative number, even if a city manager’s evaluation of
risk is such that he would be willing to incur a loss in order to accept the position. In other
words, the severance measure is a function of both the career risk present in each municipal
position and the risk aversion of the respective managers. However, severance cannot be
negative, meaning that the latent trait of true risk aversion cannot be fully captured at the lower

28
Before proceeding with the empirical analysis, I removed all observations determined to be outliers
based on statistical analysis of the distribution of the data (n=18). The data exhibited a negatively skewed
distribution, with observations prior to 2002 determined to be outliers (See Appendix A Figure 3). These
observations were removed from the sample to prevent inflated error terms and substantial distortions of
parameter estimates. Substantively, these observations were problematic because the average tenure of
city managers in a position is 7 years, and it is likely that managers in place since 2001 or earlier (13
years or more in one position) are systematically different than those with shorter tenures, who are more
representative of city managers as a whole.

110

values of the severance measure because the measure has a lower bound of zero. As such,
managers with a low degree of severance protection are clustered at the lower bound of zero and
the distribution of the data reflects this clustering at zero (see Figure 2).
Figure 2. Distribution of Data Based on Severance Protection

The tobit regression model uses all observations, both those above the limit of zero and
those at the limit of zero to estimate a regression line. There exists a continuous latent variable
of risk protection that cannot be fully measured because of the lower bound of zero on severance
pay. As such, the tobit model estimates this latent variable of monetary protection from risk, Y
i
*
,
using the indicator variable of severance pay, Y
i
, where:
Y
i
*
= X
i
β + u
i
However, I observe the following realization of Y
*
(severance protection):
Y
i
= Y
i
*
if Y
i
*
> 0
111

Y
i
= 0 if Y
i
*
≤ 0
This tobit model allows for the estimation of severance protection, which is left-censored at zero,
based on some independent variable X
i
, assuming a normally distributed error term u
i
.
To test the impact of municipal level factors on political protection, I estimate a probit
model of the likelihood of political protection, which predicts the probability that a manager
receives any form of political protection based on various municipal and individual level factors.
The probit model is appropriate in this scenario because I am estimating a latent variable, degree
of political protection from risk, based on a dichotomous dependent variable (political protection
or no political protection).
My approach postulates a continuous underlying variable of political protection from
risk, but given the constraints of my data, it cannot be measured, and instead I use a dichotomous
indicator, Y (which measures whether or not a manager has political protection), of that latent
variable, Y*. The probit model relies on the assumption that the error term, u
i
, is distributed
according to some symmetrical distribution. As such, the underlying model for political
protection is:
Y
i
*
= X
i
β + u
i
However, I observe the following realization of Y*:
Y
i
= 0 if Y
i
*
< 0
Y
i
= 1 if Y
i
*
≥ 0
This probit model allows me to estimate the latent variable Y
*
despite the limitations of the data.
Post-2008 Sample
Throughout the empirical analysis, I present several models which consider managers
who accepted positions in 2008 or later separately from those who accepted their position before
112

2008. There are several reasons to split the sample in this manner. First, in the aftermath of the
great recession and the unprecedented fiscal stress facing many municipalities, there is reason to
suspect that the nature of the profession of city management and the way managers considered
risk changed post-2008. In the wake of the great recession, especially with property values
falling, many local governments saw drastic reductions in revenue, but could not have predicted
these reductions in time to adjust their expenditures accordingly. As the story of San Bernardino
shows, falling revenues especially hurt local governments, many of which saw their pension and
healthcare obligations consume a larger portion of their revenues. As one scholar recently
concluded, “many of the promises [local governments] made to public employees are simply not
sustainable, and many jurisdictions are struggling to make payments into these systems, leaving
less each year to spend on core governmental services” (Reilly 2012).
California state law requires that local governments balance their budgets each year, and
despite some attempts to borrow from one fund to mitigate shortages in another, the fiscal shock
that occurred in 2008 was severe enough that it forced many local governments to reduce overall
expenditures. Local governments faced choices regarding the degree to which they would scale
back on service provision, reduce their municipal workforce, and reduce the salaries of the
workers who remained. In the prosperous years leading up to 2008, many city managers were
accustomed to being able to increase both the quality and quantity of services their cities
provided and the resulting quality of life in their communities. However, in the wake of the
recession, many city managers took on the task of addressing massive budget shortfalls. As one
city manager told me, “It is really easy to make decisions as a manager when you have money,
but it is very difficult to continue to create value for your community when you don’t have
money and you have to think outside the box and make hard decisions” (Brian Nakamura 2014).
113

In other words, there is anecdotal evidence to support the argument that the profession of
city management simply got harder in the wake of the 2008 recession as the landscape of local
government finance changed. There is empirical evidence of senior managers exiting the
profession in the wake of the great recession (Gabris, Davis, and Nelson 2010) although it is not
clear whether the increase in the retirement of senior managers was due to the aging of the baby
boom generation as Svara (2010) suggests or because “senior managers left after the recession
because they realized that it was going to be harder than before and they wouldn’t be able to
provide services in the way that they had in the past” (Nakamura 2014).
The second reason to consider managers who accepted positions after the recession
separately is that, among those senior managers who did remain in the profession, many
approached taking positions in struggling cities as way to “give back to the profession” or to help
a community in need. Thinking about the varying degrees of innate public service motivation
present in public managers (Crewson 1997, Perry 1997, Perry 2008), it is possible that in the
wake of the recession, senior managers who had already developed a solid professional
reputation, were willing to suspend their career concerns in order to accept positions in
struggling municipalities that under different circumstances they may have evaluated differently.
To this point I can only speculate, but after conducting a number of qualitative interviews of both
senior and junior managers, I identified a pattern among senior managers, many of whom
discussed the motivations behind their decisions to accept positions in financially or
economically unstable municipalities. These senior managers described the decision to take these
difficult jobs not as career moves that would better their reputation and bring them higher
compensation or better quality of life, but rather as personal sacrifices made to help struggling
114

communities achieve financial solvency. Pat Martel, city manager of Daly City, discussing her
decision to accept her current position, said,
“You can do a lot when you have a lot of money. When you want to do good things but
you don’t have a lot of money it is more challenging and requires more creativity. But
struggling communities need people who are committed to providing quality services and
quality of life even though there is not a lot of money” (Martel 2014).

Martel speaks to the supposition that in the wake of the great recession, senior managers who
value quality service provision and community quality of life, who possess a certain degree of
public service motivation, were willing to take on challenging positions in financially struggling
municipalities as a way of performing a public service. As Brian Nakamura, who was hired as
the city manager of Chico in 2012, said,
“They [Chico] hired me because they had a financial problem. The mayor told me that
the reason the city recruited me was because I had a track record of identifying and
solving financial problems. I took a pay cut to come to this city, knowing that they had a
$32 million budget deficit, but my whole goal in this profession is to make sure that the
community’s needs are met and that the community is a better place to live when I leave.
I knew coming into a city with financial instability and attempting to address the
underlying financial problems meant that the employees were never going to like me and
the community would go through difficult times. I told the council from the beginning of
my employment that within two years they would have to ask me to leave because I
would be that unpopular in the community.”

Stories like Nakamura’s and Martel’s raise a legitimate concern that the way managers evaluated
the nature of city manager positions changed after the shock of the 2008 recession. As such, I
present four tobit models of severance protection.
Results
Table 3 presents the results of three tobit models of severance protection, all estimated
with the exclusion of the personal education and experience variables due to data availability.
Including these two variables results in the loss of more than half the cases, and as such separate
115

tobit models are estimated with the inclusion of those variables (discussion below). The first
tobit model specification includes all cases in the sample, while the second specification includes
all cases and a dummy variable for contracts signed in 2008 or later, while the third specification
excludes those managers who signed their employment contract in 2007 or earlier.
Table 3. Tobit Model of Severance Protection (in dollars)

Full Sample Full Sample Excluding Pre-2008
Unemployment Change 9446.74***
(2639.27)
9770.57***
(2652.44)
9780.21***
(2495.32)
Deficit Spending 370.77**
(160.34)
336.13*
(195.63)
333.76
(231.73)
Political Cohesiveness 1475.67***
(507.09)
1447.92***
(432.24)
1435.44***
(44840.18)
Local Direct Democracy 4389.88
(3372.27)
5139.95
(3919.68)
6379.39**
(2813.91)
Electoral
Competitiveness
77930.23***
(21251.47)
78198.78***
(25733.87)
52723.25**
(25779.79)
Population (in thousands) 125.71**
(61.99)
121.41*
(73.59)
188.78**
(86.08)
Gender (1=male) 4049.89
(12058.44)
2934.88
(11652.43)
-1964.51
(13477.31)
School Rank 657.61***
(141.90)
635.91***
(180.02)
546.72***
(121.51)
Post 2008 Dummy
(1=yes)
-16234
(10859.41)

constant 6216.48 22052.39 23621.10
σ 66837.46
(4365.31)
66574.88 66091.94
(4639.16)
Log-psuedolikelihood -3172.32 -3171.21 -2591.00
N 272 272 224
AIC 6363.65 6364.42 5201.63
*** p<.01, ** p<.05, * p<.10. Robust standard errors (in parentheses) are clustered at the county level to account
for any variation within counties due to unobserved variables. Models have 21 and 19 left-censored observations,
respectively.

In terms of the relationship between the political condition of a municipality, the tobit
model of severance protection offers evidence that both political cohesiveness and electoral
competitiveness are related to city manager protection from risk, although there was no evidence
116

of a relationship between local direct democracy and city manager risk protection. In terms of
political cohesiveness, the first specification of the tobit model indicates that political
cohesiveness, (measuring how ideologically homogenous a community is) is associated with an
increase in city manager risk protection, offering support for Hypothesis 3. In all tobit models
robust standard errors are clustered at the county level to account for any variation at the county
level which is unobserved in the data. For example, the types and degree of services offered by
the county could vary which could potentially impact the service obligations of the municipality.
Since these variations at the county level are unobservable, robust standard errors are clustered at
the county level, which leads to more conservative estimates and a decreased likelihood of
concluding that a relationship is significant when in fact it is not.
Table 4 presents the results of the third, fourth and fifth tobit models for severance
protection which include the personal education and experience variables based on the data that
was available (as discussed in Chapter Four). My discussion will generally focus on the first and
third tobit models to avoid any generalizations based on the smaller subset of data which was
gathered through the self-reporting of a smaller number of city managers. However, this second
model offers robust support for the supposition that the coefficient estimates in the first model
hold true even when controlling for education and experience.

117

Table 4. Tobit Model of Severance Protection (in dollars)
Full Sample
w/ Personal
Characteristics
Full Sample
w/ Personal
Characteristics
Excluding Pre-2008
w/Personal
Characteristics
Unemployment
Change
10999.38***
(3641.98)
11458.34***
(3821.63
11763.01***
(3225.76)
Deficit Spending 355.26
(296.07)
262.76
(337.16)
547.22
(408.68)
Political
Cohesiveness
1597.01**
(688.14)
1531.70***
(626.58)
1572.87***
(754.62)
Local Direct
Democracy
9900.65***
(3427.16)
11513.54***
(3433.67)
14905.63**
(3583.57)
Electoral
Competitiveness
118771***
(39278.6)
123771.8***
(34963.57)
98200.37**
(39851.46)
Population (in
thousands)
-10.27
(52.45)
-24.41
(56.15)
4.70
(127.86)
Gender (1=male) -15774
(16810)
-16038.51
(16695.61)
-16552.46
(17060.2)
School Rank 455.32***
(167.93)
410.83
(285.17)
157.831
(159.76)
Years Experience -694.18
(977.38)
-918.18
(878.55)
-1558.441*
(822.29)
MPA Education 6671.16
(13701.65)
3662.08
(12214.71
-1630.35
(12060.76)
Post 2008 Dummy
(1=yes)
-29840.81*
(16584.62)

constant 25296.56 55156.87 56762.93
σ 69885.32
(6117.95)
69136.97
(4556.91)
68796.14
(5851.40)
Log-
psuedolikelihood
-1676.24 -1674.52 -1358.14
N 145 145 119
AIC 3377.67 3375.048 2740.48
*** p<.01, ** p<.05, * p<.10. Robust standard errors (in parentheses) are clustered at the county level to
account for any variation within counties due to unobserved variables. Models have 13 and 12 left-
censored observations, respectively.

The Impact of Economic Condition on Severance Protection
The first tobit model for severance protection includes all city managers in the sample
who accepted their current position as manager between 2002 and 2013 and excludes both
118

personal experience and education. This specification maximizes the number of cases included in
the model and includes robust standard errors adjusted for variation between counties otherwise
not captured in the model. Table 3 presents the results of the first specification of the tobit model
for severance protection which suggests that numerous factors are significantly associated with
city manager protection from risk.
The key measure of municipal economic condition is unemployment change, the change
in unemployment rate from the one year before the city manager signed his contract to the year
the city manager signed his contract, with increasing values indicating an increase in the
unemployment rate in the municipality. In support of the expectation that increasing economic
instability is positively associated with severance protection, the model suggests that there is a
statistically significant, positive relationship between unemployment change and severance
protection, in adjusted 2012 dollars.
Substantively, a one-percentage point increase in unemployment from the year before the
city manager signed his contract to the year in which he signed his contract is associated with a
$9,446 [$4,265, $14,627] increase in severance protection. Another way to think about this result
in a city where the unemployment rate increased from 8 percent to 10 percent, I would expect an
incoming city manager to successfully negotiate for an increase in severance protection of nearly
$19,000, hedging against the increased probability that he, as a manager, will fail to meet council
expectations and will face termination.
When the sample is split and city managers who signed their employment contracts prior
to 2008 are excluded (the second tobit model), the positive relationship between economic
instability, or unemployment change, and severance protection holds. The coefficient is slightly
larger, suggesting that among these managers who took their positions in the wake of the
119

recession, a one unit increase in unemployment from the year before they signed their contract is
associated with a $9,770 [$4,547, $14,993] increase in severance protection. These results
support the hypothesis that increasing economic instability is associated with increasing city
manager risk protection.
In the model, which includes city managers from all years but excludes those who did not
provide information regarding their personal experience and education, the relationship between
unemployment change and city manager protection from risk remains positive and statistically
significant. A one-percentage point increase in unemployment is associated with a $10,999
[$3,594, $18404] increase in severance protection. The magnitude of the impact of economic
instability is slightly larger than when personal characteristics are not considered. This provides
internal validity to the results of the original models based on the larger sample.
In the tobit model, which includes city managers from 2008 or later and excludes those
who did not provide information regarding their personal experience and education, the
relationship between unemployment change and city manager protection from risk remains
positive and statistically significant. This model predicts that a one-unit increase in
unemployment is associated with a $11,763 [$4,170, $19,355] increase in severance protection.
This model includes only 119 cases, which is problematic as the risk that the results are purely
due to chance is greater. However, again, as this fourth model confirms the relationship between
unemployment and risk protection seen in the first three models, it does offer another check of
internal validity. Examining the results of all tobit model specifications for severance protection
offers full support for Hypothesis 1, which stated that increasing economic instability is
associated with increasing city manager severance protection.

120

The Impact of Financial Condition on Severance Protection
Turning to the impact of financial instability on severance protection, the four model
specifications for severance protection offer mixed support for Hypothesis 2, that municipal
fiscal stress is positively associated with city manager protection from risk. The primary measure
of fiscal stress within the municipality is deficit spending, the percent by which expenditures
exceeded revenues in the most recent year for which data is available (either 2002-2003 or 2007-
2008). Increasing positive values of deficit spending indicate higher levels of deficit spending,
which can strain local government finances over time, while decreasing negative values indicate
a surplus in the municipal budget. I hypothesized that as deficit spending increased, city manager
protection from risk would also increase, and the model supports this hypothesis. There is a
statistically significant positive relationship (at the p < 0.05 significance level) between the
degree of deficit spending in a municipality and the severance protection of the city manager in
the first model, which includes the full sample. According to this model specification, a one-unit
increase in deficit spending is associated with an increase in severance of approximately $371
dollars.
Substantively, if comparing two cities, one in which expenditures were 2 percent less
than revenues (indicating a relatively small surplus) and one in which expenditures were 2
percent more than revenues (indicating a relatively small deficit), I expect the city manager in the
municipality with deficit spending to have $1,481 more in severance protection. The effect of
deficit spending may not seem substantial given that the average amount of severance protection
among city managers in the sample is nearly $118,000 but when one considers that the mean
level of deficit spending among cities is 4.8 percent and more than one third of local
governments in California have deficit spending such that expenditures exceed revenues by 10
121

percent of more, the impact of financial instability on city manager risk protection could be
substantial in cities with high levels of deficit spending. For instance, in a city with deficit
spending of 10 percent, I would expect city manager severance to be $3,707 dollars more than in
a city with a balanced budget.
However, when considering the model excluding those city managers who signed their
employment contracts prior to 2008, this statistically significant relationship between deficit
spending and city manager severance protection does not hold, although it does approach
statistical significance. While this was unexpected and does not offer support for Hypothesis 2, I
would conjecture that this result could be linked to the outliers, or city managers willing to
accept positions in cities with high levels of deficit spending without relatively high protection
from termination. This likely stems from some intrinsic public service motivation to go to these
communities that are struggling financially and attempt to bring them back to financial solvency
without further exacerbating their budgetary problems by requiring a large severance package.
Unfortunately, this innate public service motivation on the part of managers that I hypothesize
became more pronounced in the wake of the recession and cannot be measured with the data
available at this time.
Another factor to consider with regard to the relationship between deficit spending and
city manager severance protection is the change in the behavior of local governments before and
after 2008. Further analysis reveals that prior to 2008, there was a smaller range of deficit
spending behavior, with the highest incidence of deficit spending value occurring in a
municipality in which expenditures exceeded revenues by approximately 55 percent. However,
when considering only those cases in which the city manager signed his contract in 2008 or later,
there is a wider range of deficit spending behavior, with the most severe deficit spending
122

reaching a level with expenditures exceeding revenues by 92.9 percent (See Figure 7 for a
graphical representation of the central tendency and dispersion of the data by pre and post-2008
categories). Another possible explanation for why the impact of economic difficulty holds but
deficit spending does not is that perhaps city managers are less weary of fiscal stress as their
ability to directly influence the financial position of the city is greater than their ability to impact
overall economic condition in the municipality. While city managers do play some role in local
economic development, their ability to restructure the budget within a municipality is far greater
than their ability to propose and implement policies, which would have a marked impact on
unemployment. In other words, perhaps city managers are more confident in their ability to
address financial instability (increasing the likelihood that they will meet council expectations)
than they are in their ability to address economic instability.
Moving to the fourth, fifth, and sixth specifications of the model of severance protection
(see Table 4), when controlling for city manager personal experience and education, there is no
statistical support for the hypothesis that deficit spending is positively associated with severance
protection. Overall, these results offer mixed support of Hypothesis 2 that deficit spending is
positively associated with severance protection and suggest that perhaps other measures of fiscal
stress should be considered as well as individual city manager personality traits and
characteristics that might influence their assessment of fiscal stress within a municipality.

The Impact of Political Conditions on Severance Protection
While economic and financial conditions were hypothesized to impact city manager
protection from risk because of their direct relation to the duties and responsibilities of the city
manager, political conditions are an important factor to consider because of their role in
123

accountability and the relationship between city council and city manager. One of my goals with
this project was to consider the impact of economic and financial factors, which are traditionally
examined in studies of local government management, but to also incorporate an examination of
the role of political factors on public management within the same model. City managers operate
in a unique principal-agent relationship in which they are hired by an elected body accountable
to the people, yet they are tasked with using their professional skills and training in management
(which typically spans financial management, human resources management and economic
development among others) to professionally manage the operations of the municipality, much
like a CEO of a public company is expected to manage all aspects of the firm. Given the frequent
turnover in the profession, city managers must constantly balance the need to serve professional
norms and protect their professional reputation with the need to meet the expectations of the
council for which they serve. No study of the concerns of public managers is complete without
the consideration of the political connection between agent and principal, no matter how weak or
tenuous that connection may be.
While the public sector literature offered little guidance with regard to protection from
termination risk, the private sector literature suggested the important role that the difficulty of the
job (based on organizational factors) played in risk protection for managers. However, much of
this work in labor economics has focused on the agency relationship between CEO and board in
corporate firms, firms which operate under the profit motive and have a duty to return value to
shareholders. Municipal government does not seek profit, but rather serves the people and their
many, disparate, ever changing preferences. City councils, and managers as their agents, have
numerous goals, which often include efficiency, effectiveness, equity and fairness, but these
values are often difficult to define. As Brian Nakamura, current city manager of Chico said,
124

“ In this profession, we say the goal is to improve quality of life. Councils come up with
broad goals each year, such as improving public safety or economic development, but
when they get down to core objectives, it gets difficult because they aren’t sure what they
really want, and what they want changes too often. The council members want to be able
to go to a group of constituents and say, yes we should focus on that, and then two weeks
later, in front of another group of constituents, change their mind” (Nakamura 2014).

Such is the nature of city council politics and its impact on the principal-agent relationship
between council and manager. City council members are accountable to the people and
susceptible to public preferences and pressures, and as public preferences change, so too can the
council’s goals and their expectations of the manager.
Because of the unique relationship between council (principal) and manager (agent), I
include several measures of the political environment within a municipality to test their impact
on city manager protection from risk. Each of the three variables included to measure political
condition are included because they represent the likelihood of either the makeup of the council
changing or the preferences of the citizens changing and pressuring council goals and
expectations to change as well. A city manager going into a new position in which the political
climate is unstable and it seems likely that the preferences of the voters or the makeup of the
council will change is at greater career risk as the likelihood of his meeting council expectations
diminishes.
In the first specification of the tobit model estimating severance protection from risk,
which includes the full sample and excludes the personal characteristics of previous experience
and education, there is some support for the hypotheses that political factors are related to
severance protection from risk. Political cohesiveness is significantly and positively related to
severance protection. As the political cohesiveness, or ideologically homogeneity, of the
municipal voters increases, I would expect city manager severance protection to increase. This
125

supports the argument that ideologically extreme electorates are less likely to be supportive of
the type of financially moderate policies that a professionally trained manager might seek to
implement. This positive relationship holds in the second, third, and fourth specification of the
tobit model.
Analysis of the second key political variable, local direct democracy, a measure of the
involvement of the citizenry in the local political process, offers mixed support for Hypothesis 4.
The local direct democracy variable is not significantly related to severance protection when the
full sample is included, but when considering only city managers who accepted their positions in
2008 or later, local direct democracy is positively related to severance protection. When
considering the fourth, fifth, and sixth tobit models, those including the personal characteristics
of experience and education, the results suggest that there is a statistically significant, positive
relationship between local direct democracy and severance protection. These results offer
support for Hypothesis 4, which predicted an increase in local direct democracy would be
associated with an increase in city manager protection from risk. Substantively, based on this
estimation, a one unit increase in local direct democracy is associated with a $6,379 [$757,
$12,007] increase in severance protection. For example, I would expect a city manager taking on
a position in a city in which the citizens placed two ballot measures on the ballot in the most
recent municipal election to negotiate for approximately $12,700 more severance than a city
manager entering a city in which the citizens had placed zero ballots measures on the local ballot
in the last municipal election, holding all other variables at their means.
The final measure of political climate at the municipal level is electoral competitiveness,
a measure of the degree of competitiveness of the local elections for city council with increasing
values indicating increasing competitiveness of elections. In all six tobit models of severance
126

protection, electoral competitiveness is positively associated with city manager severance
protection. In the first (and preferred) specification of the tobit model of severance protection,
the one including the full sample, a one unit increase in electoral competitiveness is associated
with an increase in severance of $78,199 [$27,528, $128,869]. Given that the electoral
competitiveness variable ranges from 0 to 1, this is essentially the predicted difference in
severance between a municipality with completely noncompetitive elections (for example, 2
candidates running for 2 seats) and a municipality with perfectly competitive elections (in theory,
a municipality with an infinite number of candidates for only one seat). To offer a more practical
substantive effect, I consider the predicted difference in severance between two cities, both with
three available council seats but one municipality with 9 candidates running for council and one
municipality with four candidates running for council. Using these two cities for illustrative
purposes, the results of the tobit estimation lead me to predict that a city manager entering the
municipality with relatively competitive elections (nine candidates running for three seats) would
receive approximately $33,950 more in severance protection than a city manager taking a
position in the municipality with relatively uncompetitive elections (four candidates running for
three seats). While each of the four models of severance protection results in a different
coefficient for the impact of electoral competitiveness on severance, the results clearly support
Hypothesis 5 that as the competitiveness of municipal council elections increases, severance
protection from risk increases.
Finally, in the specification of the tobit model for severance protection, three additional
variables were included to control for other factors that may impact severance protection from
risk. In the first two model specifications, which exclude personal background and education
variables, population is positively related to severance protection. The larger the municipality in
127

terms of population, the more severance protection the manager receives. Based on the results of
the first specification of the tobit model, I would predict that an increase in population of 10,000
people would be associated with an increase in severance of $1,257. Moving to a similarly
situated city with 40,000 more residents is associated with an increase in severance of $6,285.
One possible explanation for the relationship between population and severance is that severance
is a function of salary and past empirical evidence has shown that salary is linked to population
size. A second possible explanation for this relationship is that the larger the municipal
population, the greater the scale of service provision a city manager must manage, and the more
constituents he must serve, which could contribute to an increase in the difficulty of the job.
However, in the third and fourth specifications of the tobit model, controlling for city manager
education and experience, population is not associated with severance protection. The second
control, gender, was not significantly associated with severance protection in any of the four
tobit models.
Finally, I considered school rank as a control variable, as quality schools can be thought
of as an additional valuable piece of the city manager’s total compensation and an indicator of
the overall desirability of a community as a place to live. With increasing school quality, the city
manager would have increasing incentives to hedge against the divestment of that piece of his
compensation package or the loss of that overall benefit of living in a desirable community. In
each of the four specifications of the tobit model for severance protection, school rank was
positively associated with severance protection. In the first model of severance protection, which
includes the full sample, a one unit increase in school rank is associated with a $636 [$281,
$990] increase in severance protection. Substantively, this suggests that holding all other
variables at their means, a city manager serving in a municipality with school quality in the 90
th

128

percentile would have approximately $13,152 more severance protection than a city manager
serving in a municipality with school quality in the 70
th
percentile in the state. This suggests that
city managers do view school quality as a valuable asset of the compensation package when
evaluating the nature of the municipality and the career risk and divestment of their total
compensation package that would result upon termination.
The inclusion of city manager previous experience and education in three of the model
specifications for severance protection offered limited evidence of personal characteristics
impacting severance protection. City manager years of experience was negatively associated
with severance protection, but only in the sixth specification of the tobit model, the specification
excluding all city managers who signed their contract prior to 2008. This specification predicts a
one-year increase in experience as a full city manager is associated with a $1,558 [-3,278, $161]
decrease in severance protection. Substantively, this leads to the prediction that in the wake of
the 2008 recession, a manager with 20 years of prior experience would have $23,370 less
protection from risk in the form of severance than a city manager with 5 years of previous
experience as a full city manager. While this result might seem counterintuitive to those familiar
with the empirical evidence in the private sector literature, which suggests that experience is
positively associated with improved negotiation outcomes, it does offer support to the hypothesis
of public service motivation and self-sacrifice based on anecdotal evidence from current
California city managers.
Multiple senior city managers stated they were willing to accept positions in financially
and economically unstable municipalities in the wake of the recession as a way of serving the
public interest and improving the quality of life in communities that were financially insolvent.
These senior managers were willing to take these positions without large severance packages for
129

two reasons. First, they were advanced enough in their career that they felt their professional
reputation was relatively well cemented and that assisting in a struggling community would not
damage their reputation. Secondly, they felt that the financial challenges of these communities
and the high likelihood that the decisions they would have to make to improve the condition in
these cities would make them unpopular in the community. Essentially, these senior managers
knew that if they did their job well in the community, their dwindling popularity would likely
necessitate that their tenure be short lived, and as such it would be exploitative to demand a large
severance package. While I only have anecdotal evidence of this phenomenon, and the idea of
local government managers essentially “sacrificing themselves” in the service of the larger good
has not been empirically tested or formally modeled, this anecdotal evidence does offer a
plausible explanation for the relationship between experience and severance when considering
only managers taking on new positions in the wake of the 2008 recession.

A Substantive Example of Predicted Risk-Shifting
In order to present a substantive, graphical presentation of the relationship between
electoral competitiveness and severance protection, based on the degree of deficit spending
present in a municipality, I constructed a categorical variable to group municipalities based on
the competitiveness of local elections, from the least competitive to the most competitive. This
categorical variable, which is used in a graphical depiction of the relationship between electoral
competitiveness, includes seven categories based on the following classifications:

130

Table 5. Measurement of Electoral Competitiveness
Label Minimum Ratio
(Probability of Winning)
Maximum Ratio
(Probability of Winning)
1 (Least Competitive) 0.85 1.000
2 0.75 0.849
3 0.60 0.749
4 0.45 0.599
5 0.30 0.449
6 0.15 0.299
7 (Most Competitive) 0.00 0.149

Imagining four hypothetical city managers each taking on a new position in a municipality and
negotiating and signing an employment contract allows for a clear understanding of the
substantive impact of these factors on severance protection. Imagine Manager One is taking a
position in a city with deficit spending of -15.28 (one standard deviation below the mean rate of
deficit spending) meaning that the city spent 15.28 percent less than they accepted in revenues.
Now imagine Manager Two is entering a municipality with deficit spending of 4.79 (the mean
rate of deficit spending), Manager Three is entering a municipality with deficit spending of 20.07
(one standard deviation above the mean rate) and Manager Four is entering a municipality with
deficit spending of 44.93 (two standard deviations above the mean rate). Holding all other
variables at their means, Figure 3 presents a graphical representation of the impact of electoral
competitiveness and deficit spending on severance protection.

131

Figure 3. Predicted Severance for Four Hypothetical Managers

Manager One: Deficit Spending = -15.28 (one standard deviation below mean)
Manager One: Deficit Spending = 4.79 (mean)
Manager One: Deficit Spending = 24.86 (one standard deviation above mean)
Manager One: Deficit Spending = 44.93 (two standard deviation above mean)

Severance protection increases as electoral competitiveness increases, but there are stark
differences between the predicted severance of a manager in a city with surplus spending and a
manager in a city with extreme deficit spending. Manager One, whose hypothetical city is
operating under a surplus, is expected to receive severance protection of approximately $75,000
when electoral competitiveness is at its lowest, and severance protection of approximately
$145,000 when electoral competitiveness is at its highest. This represents a nearly 100 percent
132

increase in expected severance based solely on municipal elections moving from the least
competitive category to the highest, all other variables are held at their means. Now consider
Manager Four, operating in a city with the highest level of deficit spending. Under conditions of
very low electoral competitiveness, this manager is predicted to received approximately $95,000
in severance - $25,000 more than Manager One who is also in a city with uncompetitive
elections but is operating under a surplus. When the electoral competitiveness of Manager Four’s
municipality increases to the highest level, the model predicts that the manager will receive
approximately $165,000 in severance. This hypothetical scenario and graphical representation
illustrate the substantial impact of both deficit spending and electoral competitiveness on the
predicted level of severance protection granted a manager.

The Impact of Municipal Conditions on Political Protection
The results of the estimation of several tobit models of severance protection offers
support for the theory that managers and councils seek to align their incentives prior to
employment via the employment contract. They also support the hypotheses that municipal
conditions perceived to increase the risk of termination are related to the degree of protection
from termination present in manager contracts. However, severance protection is not the only
way a city manager can seek protection from the career risk presented by instability in a
municipality. While severance offers financial protection and insurance against the repercussions
of termination, contractual provisions restricting termination before or after a municipal election
offer protection from termination by granting the manager a period of time after a potential
change in the makeup of the council to have a strong degree of job security. As described in
Chapter Four, political protection typically takes the form of a number of days (often in 30 day
133

increments) before or after a municipal election during which a city manager cannot be
terminated without cause. These provisions allow for a period of time for new council members
to integrate with the incumbent council members, for new council members and the manager to
develop a working relationship, and for new council members to familiarize themselves with the
work of the manager. In order to test the impact of various economic, financial, and political
conditions on city manager risk protection via political protection, I estimate four probit models
of the probability of a manager receiving political protection in his employment contract. As in
the previous section on the estimation of severance protection, I estimate one model with the full
sample, one model with only those city managers who signed their employment contract in 2008
or later, one model with city managers from all years including variables for experience and
education, and one model with only city managers in the 2008 or later class including variables
for experience and education. Due to the loss of cases and smaller sample size associated with
the last three models, the first model including the full sample is the preferred specification.
In order to test the impact of municipal factors on city manager political protection, I
estimate a probit model of the probability of political protection being included in a city manager
contract (See Table 6). The results of this model suggest that neither unemployment change nor
deficit spending are significantly associated with the probability of political protection. This
result holds in the second model, which excludes city managers who accepted their position prior
to 2008.

134

Table 6. Probit Model of Political Protection from Risk

Full Sample Excluding Pre-2008
Unemployment Change -0.0554
(0.065)
-0.076
(0.069)
Deficit Spending -0.004
(0.005)
-0.002
(0.005)
Political Cohesiveness 1.722**
(0.72)
1.570**
(0.74)
Local Direct Democracy 0.036
(0.063)
0.013
(0.069)
Electoral Competitiveness 0.144
(0.530)
0.612
(0.591)
Population (in thousands) -0.004**
(0.0015)
-0.003**
(0.0015)
Gender (1=male) 0.033
(0.239)
0.062
(0.296)
School Rank 0.007**
(0.003)
0.008**
(0.004)
constant -0.906 56762.93
Log-psuedolikelihood -156.72 -1358.14
N 270 222
AIC 342.29 284.21
*** p<.01, ** p<.05, * p<.10. Robust standard errors clustered on the county in
parentheses.

With regard to the political climate in the municipality, neither local direct democracy
nor electoral competitiveness was significantly associated with the probability of political
protection in any of the four models. Political cohesiveness was the only political variable of the
three to show any association to the probability of political protection. As political cohesiveness,
or ideological homogeneity, increases, the model predicts that the probability of a manager’s
contract including protection from political termination increases. A .10 increase in political
cohesiveness is associated with an increase in the probability of political protection of 0.17,
suggesting that the more ideologically homogenous a community is, at the ends of the
ideological spectrum, the more risk of termination around an election the manager perceives.
135

These results suggest that city managers are rather sophisticated with regard to evaluating
potential career risk and shifting that risk onto council. That financial and economic factors do
not impact political protection suggests that managers recognize the source of the career risk they
face (political) and shift that risk back on to the council through the form of protection (political)
that most closely maps to the risk.
The third and fourth probit models (See Table 7) which include the variables for city
manager professional experience and education, also offer no support for the hypotheses that
increasing municipal economic instability or increasing financial instability are related to the
probability of political protection in the manager’s employment contract. While this result was
unexpected, it is consistent across all models.

136

Table 7. Probit Model of Political Protection from Risk w/ Personal Characteristics
Full Sample
w/ Personal Characteristics
Excluding Pre-2008
w/ Personal Characteristics
Unemployment Change -0.142
(0.099)
-0.163
(0.102)
Deficit Spending -0.003
(0.006)
-0.002
(0.008)
Political Cohesiveness 2.012*
(1.295)
1.633
(1.40)
Local Direct Democracy -0.005
(0.111)
-0.0149
(0.117)
Electoral Competitiveness 0.178
(0.967)
0.144
(1.072)
Population (in thousands) -0.004*
(0.0024)
-0.002
(0.002)
Gender (1=male) -0.221
(0.312)
-0.073
(0.371)
School Rank 0.005
(0.004)
0.0065
(0.0049)
Years Experience 0.0243*
(0.0136)
0.011
(0.012)
MPA Education -0.1654
(0.186)
-0.149
(0.246)
constant -0.807 -0.903
Log-psuedolikelihood -83.923 -69.818
N 142 116
AIC 192.78 164.34
*** p<.01, ** p<.05, * p<.10. Robust standard errors (in parentheses) are clustered at the
county level to account for any variation within counties due to unobserved variables.

A substantive example illustrates the impact of political cohesiveness, measured as the
absolute value of 0.50 minus the proportion of the municipal population that voted for Obama in
the 2008 presidential election. Based on the measurement of the variable, increasing values do
not indicate a conservative or a liberal leaning, rather they indicate the relative homogeneity of
the municipal electorate in their preferences. For example, Berkeley, in which 92 percent of the
citizens voted for Obama, receives a score of 0.42 on this scale, while Maricopa, in which 19
percent of the citizens voted for Obama, receives a score of 0.31 on this scale. This does not
137

mean that the two municipalities are both approaching similar degrees of conservativeness or
liberalness, but rather, they are both approaching similar degrees of the homogeneity of political
preference. The first model (in Table 6) predicts that an increase in the political cohesiveness
score from zero to a score of 0.35 (representing a relatively high degree of homogeneity)
increases the probability of political protection in the contract from 0.253 to 0.465. In other
words, the manager in the more politically cohesive city (0.35) has approximately a one in two
chance of receiving political protection, while a manager in the less politically cohesive city
(zero) has approximately a one in four chance of receiving political protection. These results
offer clear support for the hypothesis that increasing political cohesiveness or ideological
homogeneity is associated with an increase in the probability of political protection. The more
politically cohesive (or ideologically homogenous) the electorate becomes, the median voter
essentially moves further into the extremes of the ideological spectrum, meaning that the council,
who is responsive to the median voter is also likely to support more ideologically extreme
policies. This increases the manager’s risk of failing to please the council as their policy
pressures move further into the tails of the ideological spectrum and they likely become less
supportive of moderate policy proposals on the part of the manager.
Several control variables are significantly related to the probability of political protection.
Population has a negative association to political protection in three of the four specifications of
the probit model of political protection. This result suggests that as population increases the
likelihood of receiving protection from political termination decreases. Additionally, there is
some evidence that school rank is positively related to the probability of political protection
based on the results of the first and second specifications of the probit model. When excluding
controls for personal experience and education, the results predict that a one unit increase in
138

school quality is associated with an increase in the probability of political protection of
approximately 0.007. Considering the more conservative estimate of the impact of school rank
on political protection from the first specification of the probit model, the substantive effect is
small. Based on the model predictions, a city manager in a municipality ranked in the 90
th

percentile of school quality would have an increased probability of .001 over a city manager in a
municipality with school’s ranked in the 50
th
percentile of school quality.
With regard to the controls for experience and education, there is some evidence that the
years of experience a city manager has may be associated with the probability of political
protection. In the third specification of the probit model, which includes all years and includes
the personal background variables, years of experience is positively associated with the
probability of political protection (at the p < 0.10 level of significance). Substantively, the model
predicts that a city manager with 25 years of experience would have an increased probability of
political protection of 0.167 over a city manager with 5 years of experience, holding all other
variables at their means. While this result suggests that city manager with more experience may
be more successful in their negotiations for political protection, the result does not hold in the
second model, meaning that, based on the results of both models, there is not strong evidence of
a relationship between experience and political protection.
Overall, the probit models of political protection present mixed evidence, at best, of the
relationship between municipal level factors and city manager protection from risk. While
political cohesiveness was related to the probability of political protection in two of the four
specifications of the model, the result did not hold when city manager experience and education
were considered, meaning that no real conclusions about this relationship can be drawn.
139

One possible explanation of the failure of these models to offer support to the hypotheses
presented in Chapter Four is that political protection is more the result of the municipal status
quo than of an evaluation of risk and conscious negotiation between manager and council. Only
about one-third of managers in the sample had any degree of political protection from
termination in their employment contracts. There is also the potential of measurement error, as
some municipalities do include provisions forbidding the termination of the city manager without
cause for a certain period of time after municipal elections within their municipal code. These
provisions were enacted with the intent of allowing a standard amount of time for new council
members to familiarize themselves with municipal affairs before making any determinations
about changes in leadership. In the cases where these provisions do exist in the municipal code,
they are not negotiated for by the council and manager, as they apply to each manager who
works for that municipality, so they cannot be thought of as the result of any sort of evaluation of
the difficulty of the job, risk calculus on the part of the manager, or negotiations between council
and manager. However, their very existence in the municipal code removes any need for them to
be considered for inclusion in the employment contract. Unfortunately, quantitative information
about political protection clauses in municipal codes in California is unavailable, and conducting
a content analysis of all municipal codes as well as all employment contracts would have been
prohibitively time consuming.
Another possible explanation for the lack of evidence connecting municipal condition
and political protection is the relative impact of political protection as insurance in relation to
severance protection. Several city managers suggested that severance functioned as a much
better insurance policy against the risk of termination. As they mentioned, once they council
decides to terminate, they do not typically change that decision, meaning political protection
140

only delays termination, but offers no benefit to the manager once termination does occur. As
such, political protection offers minimal protection from termination risk compared to severance
protection. In contrast to political protection, severance protection gives the manager a sizeable
lump sum payment, which can be used to support the payment of debts and liabilities and to
maintain a certain lifestyle while seeking employment with another organization.
Lastly, many managers mentioned that one of the most important factors when
considering a move to a new position was finding the right fit. The larger a severance package a
manager receives upon termination, the longer that manager has to fully explore his other options
of employment, increasing his likelihood of finding a position that offers a good fit. Pat Martel,
city manager of Daly City, said in reference to the decision to accept a new position,
“I have always wanted to be in a community and organization where I would reflect the
community and have the community see someone that reflected their backgrounds and
their values and concerns working in city hall, so I have always selected communities
that were racially and culturally diverse, because those communities were the best fit for
me as a woman and person of color” (Martel 2014).

If managers are concerned with finding the right fit in accepting a new position, this could offer
an explanation as to why managers would want to maximize the amount of time they have to
seek employment after a termination. The larger a severance package is, the longer a manager
will have to seek employment. Political protection offers no similar benefit to managers. While
each of these possible explanations are based primarily on qualitative interviews, they do suggest
that the relationship between political protection and municipal condition may not operate in the
way I originally hypothesized, and as such both the nature of political protection theories of the
factors that impact political protection, and the measurement of political protection should be
reexamined in the future.
141

The results of the empirical analysis presented in this chapter offer substantial support for
the theory of risk shifting between council and manager I outlined in earlier chapters. As the
career risk associated with a given manager position increases, the manager shifts that risk back
onto the council through either severance protection or political protection. Particularly
interesting are the results suggesting that not only do managers shift career risk back onto the
council, but they are acutely aware of the sources of career risk and map their protection from
that risk to its original source. An increase in risk stemming from financial or economic
conditions increases the degree of protection in the form of severance, which provides protection
for the manager in the event of termination, while risk stemming from political conditions
increases the degree of protection before or after elections, providing a more direct insurance
against the specific type of risk present in the municipality. These results suggest that the
conditions of the municipality, in regard to economic, financial, and political condition, do
impact the degree of career risk facing an incoming city manager, and as such, the manager seeks
to shift that risk back onto the city council through insurance against and protection from
termination.

142

CHAPTER SEVEN: CONCLUSION
Local public management is an important area of scholarly inquiry, yet little work thus
far has examined the institutional arrangements determined in city manager employment
contracts that structure the principal-agent relationship between manager and council. A more
thorough understanding of the incentive structures and risk protections within city manager
employment contracts is critical to developing a clear theory of how city managers behave within
the municipal organization. While this project advances scholarly knowledge with regard to the
ex ante institutional arrangements between principal and management in local government, it is
only the first step toward developing a viable theory of how the institutional structure under
which city managers operate influence ex post outcomes.
Municipal governments seek to bring in top talent to manage the provision of services,
implementation of the budget, and other administrative tasks within the city. However, potential
managers evaluate cities as potential employers, in terms of both the organizational and political
environment they would be entering and the challenges, as top management, they would be
taking on in terms of managing the daily operations of the city. When a city manager is
considering a new position, they negotiate some level of compensation based on their individual
characteristics (education and experience) and the size of the city, but beyond those two factors,
previous work in local government management failed to account for the manner in which the
career risk associated with a given municipality is priced into the total compensation package
negotiated between manager and council. The institution of the city manager is organized by
contract, but any focus on negotiation for discretion in the contract is misguided; city manager
employment contracts do not specify varying degrees of discretion or performance metrics or
goals linked to incentives. The questions that this project set out to answer were (1) in the
143

absence of these provisions for discretion and performance incentives, how do city manager
employment contracts shift risk between council and manager ex ante? And (2) how do
municipal factors, above and beyond personal experience and education, impact the degree of
risk shifted from manager to council?
Each city offers a different kind of job in terms of the level of difficulty the job presents
and the likelihood that the city manager will succeed in meeting council expectations. The results
of the analysis of this project provide clear empirical evidence that the inherent career risk of a
given position, which varies from city to city, is manifested in the protection from risk, or
severance protection, that the city manager and city council negotiate before the city manager
starts his first day of work in city hall. The empirical results of my analysis have overarching
implications for scholars and practitioners of local government as I have essentially estimated the
risk premium associated with hiring a city manager in a city facing economic, financial, or
political distress. The results provide evidence in support of my predictions that the more career
risk a manager believes he is taking by accepting a position, with regard to the difficulty of the
job, the likelihood of termination, and the potential for harm to his professional reputation, the
more protection from termination he will negotiate for in his employment contract.
Using data representative of municipal governments throughout the state of California, I
have developed a unique dataset of both municipal economic, financial, and political conditions.
Most importantly for the advancement of our understanding of how the institution of the city
manager is organized, I developed a dataset that for the first time quantifies the terms of city
manager employment contracts. While this dataset was instrumental in testing my predictions
regarding municipal condition and protection from risk, it will prove to be a fruitful dataset into
144

the future for many more empirical studies related to the principal-agent relationship between
council and manager.
With the development of this dataset and the subsequent empirical analysis, I have
advanced the current scholarship on the principal-agent problem at the local level, which prior to
this project was unsystematic and lacked the ability to generalize results to cities of all sizes and
conditions. Furthermore, this project moves the study of local government beyond the
dichotomous distinction between the council-manager form of government and the mayor-
council form. As this analysis suggests, focusing on that simple distinction in government type
muddies the waters of local government scholarship and misses the point. Now that a large
majority of municipalities in the southeast and western United States operate under the council-
manager form of government, and as trends suggest that these numbers will only grow, local
government scholarship must focus on the city manager as the unit of analysis in order to
understand differences between managers themselves.
Moving forward, there are several directions in which future research will evolve. As
unelected public officials, city managers operate in an environment that is largely shaped by the
council who hired them. Within the municipal organization, policy choices and the decisions of
city managers reflect a response to the rules, laws, and conventions built into the environment
(Powell 2007). While this project has helped to define the rules, laws, and conventions built into
the environment, moving forward, scholars should examine the impact of this institutional
environment on city manager behavior and performance outcomes at the local level.
Generally speaking, a fruitful research agenda on understanding phenomena of
importance to local government management must account for the relationship between public
servants and democratically elected officials. While this project focuses on how city managers as
145

public servants, evaluate the risk of taking a position and how they seek to protect themselves
from that risk, there are limitations to the data which I could attempt to address in the future.
First and foremost, a reconsideration of the role of political protection and its measurement is
warranted. Secondly, a survey of city managers could offer insight into several factors that this
analysis was unable to touch on, such as city manager public service motivation, previous career
history, career concerns, and their risk aversion. The results suggest that there are systematic
differences in the risk-shifting behavior of managers before and after 2008, which warrants
further examination of the impact of the recent recession on the ways in which city managers
approach risk. Particularly interesting is the idea that in the wake of the recession senior
managers set aside their concerns of career risk in order to perform a public service for
struggling communities. Although the current data available did not allow me to test any
hypotheses related to this topic directly, it is worth exploring in the future by building on this
current dataset. One logical way to expand the data to test this relationship directly is with a
survey of city managers designed to tap their intrinsic public service motivation.
Another direction in which this research agenda will move is toward a comparative focus.
While this project offers empirical evidence of the impact of municipal factors on city manager
risk protection, it is based solely on data from California cities. While California is frequently a
leading state with regard to progressive institutional change, a comparison to another state,
perhaps a state with different privacy laws (allowing public access to résumés) or with different
budgetary requirements would offer useful insight into the impact of state level institutions on
council-manager relationships. For example, Florida’s sunshine laws allow for greater public
access to documents related to the recruitment of job candidates at the local level. Any candidate
for a city manager position is required to submit his resume with the understanding that his entire
146

application for the position becomes public record. While this offers access to more data, the
more interesting aspect of this dynamic is the ability to empirically study the impact of these
open access laws on the ways in which managers approach the consideration of these jobs and
the career risk they present. For, while municipal conditions in a potential city of employment
create future career risk, the awareness by a manager’s current council of the manager’s intention
to or interest in leaving their city creates an entirely different layer of career risk for managers to
consider.

147

APPENDIX A:

Figure 4. Dispersion of Data Based on Year

Figure 5. Box Plots of Deficit Spending by Pre or Post-2008 (0 = before 2008)
148

BIBLIOGRAPHY

Abowd, John M., and Orley C. Ashenfelter. 1981. "Anticipated unemployment, temporary
layoffs, and compensating wage differentials." In Studies in labor markets, pp. 141-186.
University of Chicago Press, 1981.

Aggarwal, Rajesh, and Andrew A. Samwick. 1998. The other side of the tradeoff: The impact of
risk on executive compensation. No. w6634. National Bureau of Economic Research.

Almazan, Andres, and Javier Suarez. 2003. "Entrenchment and severance pay in optimal
governance structures." The Journal of Finance 58, no. 2: 519-548.

Ammons, David N., & C. Newell. 1989. City Executives. Albany, NY: State University of New
York Press.

Ashworth, Scott. "Electoral accountability: recent theoretical and empirical work." Annual
Review of Political Science 15 (2012): 183-201.

Banovetz, J. M. 1995. Council-manager government’s response to economic development. Ideal
and practice in council‐manager government, 213-226.

Barber, Daniel M. 1988. "Newly promoted city managers." Public Administration Review: 694-
699.

Bendor, Jonathan, Serge Taylor, and Roland Van Gaalen. 1987. "Politicians, bureaucrats, and
asymmetric information." American Journal of Political Science: 796-828.

Benton, J. Edwin. 2002. “County Service Delivery: Does Government Structure Matter?” Public
Administration Review 62 (4): 471–479.

Berry, Christopher R., and William G. Howell. "Accountability and local elections: rethinking
retrospective voting." Journal of Politics 69, no. 3 (2007): 844-858.

Bertelli, Anthony. 2012. The Political Economy of Public Sector Governance. New York, NY:
Cambridge University Press.

Bertelli, Anthony M., and Laurence E. Lynn. 2006. Madison's managers: Public administration
and the Constitution. JHU Press.

Binderkrantz, A. S., & Christensen, J. G. (2012). Agency performance and executive pay in
government: An empirical test. Journal of public administration research and
theory, 22(1), 31-54.

149

Booms, Bernard H. 1966. “City Government Form and Public Expenditure Level.” National Tax
Journal 19(2): 187-199.

Bowlin, W. F. (1997). A proposal for designing employment contracts for government
managers. Socio-Economic Planning Sciences, 31(3), 205-216.

Bowman, Ann and Richard Kearney. 2008. State and Local Government: The Essentials.
Houghton Mifflin Harcourt: Boston, MA.

Bowman, Ann and Richard Kearney. 2011. “Local Government Structures and Leadership” in
State and Local Government: The Essentials. Wadsworth: Boston, MA.
Brewer, G. A., & Walker, R. M. (2013). Personnel constraints in public organizations: the impact
of reward and punishment on organizational performance. Public Administration
Review, 73(1), 121-131.

Brown, Heather. 2014. Good Question: Why Do We Pay CEOs So Much To Leave. CBS
Minnesota. May 5.

Brown, Kenneth W. 1993. "The 10-point test of financial condition: Toward an easy-to-use
assessment tool for smaller cities." Government Finance Review 9: 21-21.

California Department of Education. 2013. Fingertip Facts on Education in California.
http://www.cde.ca.gov/ds/sd/cb/ceffingertipfacts.asp

Carpenter, Daniel. 2006. “Reputation and the Regulator,” presented at the University of Bath
Center for Regulated Industries Conference on Frontiers of Regulation – Assessing
Scholarly Debates and Policy Challenges. Bath, United Kingdom.

Carpenter, Daniel P., and George A. Krause. 2012. "Reputation and public
administration." Public Administration Review 72, no. 1: 26-32.

Clingermayer, James C., and Richard C. Feiock. 1997. "Leadership turnover, transaction costs,
and external city service delivery." PUBLIC ADMINISTRATION REVIEW-
WASHINGTON DC- 57: 231-239.

Cope, Glen Hahn. 1987. "Local Government Budgeting and Productivity: Friends or
Foes?." Public Productivity Review: 45-57.

Cope, Glen Hahn. 1992. "Walking the fiscal tightrope: Local government budgeting and fiscal
stress." International Journal of Public Administration 15, no. 5: 1097-1120.

Crewson, P. E. (1997). Public-service motivation: Building empirical evidence of incidence and
effect. Journal of public administration research and theory,7(4), 499-518.

150

Dahl, Robert Alan. 1961. Who governs?: Democracy and power in an American city. Yale
University Press.

Deno, Kevin T., and Mehay, Stephen L. 1987. “Municipal Management Structure and Fiscal
Performance: Do City Managers Make a Difference?” Southern Economic Journal 53(5):
627-641.

Deno, Kevin T., and Stephen L. Mehay. 1987. "Municipal management structure and fiscal
performance: do city managers make a difference?." Southern Economic Journal 1987:
627-642.

DeHoog, R., & Whitaker, G. P. 1990. Political Conflict of Professional Advancement. Journal of
Urban Affairs, 12(4), 361-377.

DeSantis, Victor S., and Charldean Newell. 1996. "Local government managers’ career
paths." The Municipal Year Book 1996: 3-10.

DeSantis, Victor S., and Tari Renner. 1994. "The impact of political structures on public policies
in American counties." Public Administration Review: 291-295.

DeSantis, V. and Renner, T. 2002. “City Government Structures: An Attempt at Clarification.”
In H.G. Frederickson and J. Nalbandian (Eds.), The Future of Local Government
Administration: The Hansell Symposium. Washington DC, International City/County
Management Association.

Donahue, Amy Kneedler, Sally Coleman Selden, and Patricia W. Ingraham. 2000. "Measuring
government management capacity: A comparative analysis of city human resources
management systems." Journal of Public Administration Research and Theory 10, no. 2:
381-412.

Donovan, Todd, and Max Neiman. 1992. "Community social status, suburban growth, and local
government restrictions on residential development." Urban Affairs Review 28, no. 2:
323-336.

Edwards, Andrew. 2009. “San Bernardino’s New City Manager, Police Chief to Start Work
Monday.” San Bernardino Sun. May 31.

Edwards, Andrew. 2009. “Four Candidates Quality for Fourth Ward Election.” San Bernardino
Sun. January 9.

Edwards, Andrew. 2009. “San Bernardino’s Election Less Than a Week Away.” San Bernardino
Sun. March 10.

Epstein, David, and Sharyn O’Halloran. 1994. "Administrative procedures, information, and
agency discretion." American Journal of Political Science 38, no. 3: 697-722.
151

Feiock, R. C. (2004). Politics, institutions and local land-use regulation. Urban Studies, 41(2),
363-375.

Feiock, R. C., & Tavares, A. F. 2002. County government institutions and local land use
regulation. Cambridge, MA: Lincoln Institute of Land Policy.

Feiock Richard C., and Clingermayer, James C. 1997. "Leadership turnover, transaction costs,
and external city service delivery." PUBLIC ADMINISTRATION REVIEW-
WASHINGTON DC- 57: 231-239.

Feiock, Richard C., and Christopher Stream. "Explaining the tenure of local government
managers." Journal of Public Administration Research and Theory8, no. 1 (1998): 117-
130.

Feiock, Richard, and Christopher Stream. "Local government structure, council change, and city
manager turnover." In The Future of Local Government Administration: The Hansell
Symposium,(Washington, DC: ICMA), pp. 118-123. 2002.

Feiock, Richard C., Moon‐Gi Jeong, and Jaehoon Kim. 2003. "Credible Commitment and
Council‐Manager Government: Implications for Policy Instrument Choices."Public
Administration Review 63, no. 5: 616-625.

Feiock, Richard C., and Jae-Hoon Kim. "Form of government, administrative organization, and
local economic development policy." Journal of Public Administration Research and
Theory 11, no. 1 2001: 29-50.

Feldman, Martha S., and Anne M. Khademian. 2002. "To manage is to govern."Public
Administration Review 62, no. 5: 541-554.

Fiorina, Morris P., and Samuel J. Abrams. 2008. "Political polarization in the American
public." Annu. Rev. Polit. Sci. 11: 563-588.

Francis, Nathan, Richard C. Feiock, and Anthony Kassekert. "Should I Stay or Should I Go? The
Clash between Push and Pull Factors Effecting Administrative Turnover." In 10th
National Public Management Research Conference, Columbus, OH. 2009.

Frederickson, H. George and Gary Alan Johnson. 2001. “The Adapted American City: A Study
of Institutional Dynamics.” Urban Affairs Review 36(6): 872-884.

Frederickson, H. George, Gary Alan Johnson, and Curtis H. Wood. 2003. The Adapted City:
Institutional Dynamics and Structural Change. M.E. Sharpe.

Frederickson, H. George, and Kevin B. Smith. 2003. The Public Administration Theory Primer.
Bouder, CO: Westview Press.
152

Gabris, G. T., Davis, T. J., & Nelson, K. L. (2010). Demand Versus Supply: Can MPA Programs
Satisfy the Need for Professional Management in Local Government?. Journal of Public
Affairs Education, 379-399.

Gianakis, Gerasimos and Clifford McCue. 1999. Local Government Budgeting: A Managerial
Approach. Westport, CT: Praeger.

Goetz, Edward G. 1994. "Expanding possibilities in local development policy: An examination
of US cities." Political Research Quarterly 47, no. 1: 85-109.

Goldstein, Gerald S., and Ronald G. Ehrenberg. 1976. "Executive compensation in
municipalities." Southern Economic Journal: 937-947.

Gould, Rodney. 2014. Personal Communication, Santa Monica, CA. April 14.

Hambrick, D.C., and S. Finkelstein. 1987. “Managerial Discretion: A Bridge Between Polar
Views of Organizations.” Research in Organizational Behavior 9: 369-406.
Hagen, Ryan. 2012. San Bernardino City Manager, Other Officials Resign; Toxic Atmosphere
Blamed. San Bernardino Sun. March 22.

Hagen, Ryan. 2013. San Bernardino Bankruptcy Timeline. San Bernardino Sun. July 27.

Hagen, Ryan. 2013. Blame Game on how San Bernardino Came to File for Bankruptcy. San
Bernardino Sun. July 27.

Hajnal, Zoltan L. and Troustine, Jessica. 2010. Who or What Governs?: The Effect of
Economics, Politics, Institutions, and Needs on Local Spending. American Politics
Research, 38, 6: 1130-1163

Hayes, Kathy, and Semoon Chang. 1990. “The Relative Efficiency of City Manager and Mayor-
Council Forms of Government.” Southern Economic Journal 57 (1): 167–177.

Hirschman, Albert O. 1970. Exit, voice, and loyalty: Responses to decline in firms,
organizations, and states. Vol. 25. Harvard university press.

Honadle, Beth Walter. 1981. "A capacity-building framework: A search for concept and
purpose." Public Administration Review: 575-580.

Honadle, Beth Walter, Beverly Cigler, and James M. Costa. 2003. Fiscal Health for Local
Governments. 1st ed. Academic Press.

Honadle, Beth W. and Lloyd-Jones, M. 1998. “Analyzing Rural Local Governments’ Financial
Condition: An Exploratory Application of Three Tools.” Public Budgeting & Finance,
18(2): 69-86.
153

Huber, John D., and Charles R. Shipan. 2002. Deliberate discretion?: The institutional
foundations of bureaucratic autonomy. Cambridge University Press.

International City/County Management Association. 1996. Municipal Form of Government
Survey.

International City/County Management Association. 2006. Municipal Form of Government
Survey.

International City/County Management Association. 2007. Council Manager Form of
Government.

Jensen, Michael C., and William H. Meckling. 1976. "Theory of the firm: Managerial behavior,
agency costs and ownership structure." Journal of financial economics 3, no. 4: 305-360.

Jensen, M. C., & Murphy, K. J. (1990). Performance pay and top-management
incentives. Journal of political economy, 225-264.

Kemp, Roger. 1999. Local Government Election Practices: A Handbook for Public Officials and
Citizens. Jefferson, NC. : McFarland & Company.

Kimmel, M. J., Pruitt, D. G., Magenau, J. M., Konar-Goldband, E., & Carnevale, P. J. (1980).
Effects of trust, aspiration, and gender on negotiation tactics. Journal of Personality and
Social Psychology, 38(1), 9.
Kammerer, Gladys M. City Managers in Politics. University of Florida Press, 1962.

Kane, Thomas J., Stephanie K. Riegg, and Douglas O. Staiger. "School quality, neighborhoods,
and housing prices." American Law and Economics Review 8, no. 2 (2006): 183-212.

Kane, Thomas J., Douglas O. Staiger, Gavin Samms, Edward W. Hill, and David L. Weimer.
"School Accountability Ratings and Housing Values [with Comments]." Brookings-
Wharton Papers on Urban Affairs (2003): 83-137.

Kaatz, James Bryan. 1996. City manager tenure stability: A predictive model based on perceived
council conflict, leadership effectiveness, and conflict resolution ability. Northern Illinois
University.

Key Jr, V. O. 1966. "The Responsible Electorate." Harvard University, Cambridge, MA.

Ladd, H.F., and Yinger, J. 1989. America’s Ailing Cities: Fiscal Health and the Design of Urban
Policy. Baltimore, MD: Johns Hopkins University Press.

154

Levin, Jonathan, and Steven Tadelis. 2010. “Contracting for Government Services: Theory and
Evidence from U.S. Cities.” Journal of Industrial Economics 58(3): 507-541.

Lipsky, Michael. 1980. "Street level bureaucrats." Nova York: Russel Sage.

Loveridge, Ronald O. 1971. City managers in legislative politics. Indianapolis: Bobbs-Merrill.

Lovett, Ian. 2013. A Poorer San Bernardino, and a More Dangerous One, Too. The New York
Times. January 14.

Loviscek, A. and F. Crowley. 1990. “What is in a Municipal Bond Rating?” The Financial
Review 25(1): 25-53.

Lubell, Mark, Richard C. Feiock, De La Cruz, and Edgar E. Ramirez. 2009. "Local institutions
and the politics of urban growth." American Journal of Political Science 53, no. 3: 649-
665.
Lupia, Arthur, and Mathew McCubbins. 1994. "Who controls? Information and the structure of
legislative decision making." Legislative Studies Quarterly 19, no. 3.

Luckerson, Victor. 2014. This Guy Got $58 Million for 15 Months of Work. Time. April 17.

Lys, T., and E. Sletten. 2006. "Motives for and risk-incentive implications of CEO
severance." Unpublished working paper, Northwestern University, Evanston, IL.

Maher, Craig S., and Karl Nollenberger. "Revisiting Kenneth Brown’s “10-Point
Test.”." Government Finance Review 25, no. 5 (2009): 61-66.

Martel, Patricia. 2014. Personal Communication. April 21.

Martin, David L. 1990. Running City Hall: Municipal Administration in America. Tuscaloosa:
University of Alabama Press.

Martin, J. Schiesl. 1977. "The Politics of Efficiency: Municipal Administration and Reform in
America, 1880-1920." Berkeley: University of California Press 112 (1977).

McCabe, B. C., Feiock, R. C., Clingermayer, J. C., & Stream, C. 2008. Turnover among city
managers: The role of political and economic change.Public Administration
Review, 68(2), 380-386.

McCubbins, Mathew D., and Thomas Schwartz. 1984. "Congressional oversight overlooked:
Police patrols versus fire alarms." American Journal of Political Science: 165-179.

McDonald, J. F., & Moffitt, R. A. (1980). The uses of Tobit analysis. Review of economics and
statistics, 62(2), 318-321.
155

Meek, J.W., Schildt, K. and Witt, M. 2002. “Local Government Administration in a Metropolitan
Context.” In H.G. Frederickson and J. Nalbandian (Eds.), The Future of Local
Government Administration (145-153). Washington, DC: International City/County
Management Association.

Meeker, B. F., & Weitzel-O'Neill, P. A. (1977). Sex roles and interpersonal behavior in task-
oriented groups. American Sociological Review, 91-105.

Morgan, David R., and Kenneth Kickham. 1999. “Changing the Form of County Government:
Effects on Revenue and Expenditure Policy.” Public Administration Review 59 (4): 315–
324.

Morgan, David R., and Sheilah S. Watson. 1992. “Policy Leadership in Council-Manager Cities:
Comparing Mayor and Manager.” Public Administration Review 52 (5): 438–446.

Nakamura, Brian. 2014. Personal Communication, Santa Monica, CA. April 18.

Nalbandian, J. 1989. The contemporary role of city managers. The American Review of Public
Administration, 19(4), 261-278.

Nalbandian, J. 1999. Facilitating community, enabling democracy: New roles for local
government managers. Public Administration Review, 187-197.

Nalbandian, John. 2004. "Politics and Administration in Council‐Manager Government:
Differences between Newly Elected and Senior Council Members." Public
Administration Review 64, no. 2 2004: 200-208.

National Association of Schools of Public Affairs and Administration. 2009. Commission on
Peer Review and Accreditation. Presented at the NASPAA Annual Business Meeting in
Arlington, VA. October 2009.

National League of Cities. 2012. Forms of Municipal Government. Washington, DC: National
League of Cities.

Niskanen, William A. 1975. "Bureaucrats and politicians." Journal of law and economics: 617-
643.

Northup, Nancy. 1987. "Local nonpartisan elections, political parties and the first
amendment." Columbia Law Review (1987): 1677-1701.

O’Leary, Rosemary. 2013. The ethics of dissent: Managing guerrilla government. CQ Press.

O'Shea, P. G., & Bush, D. F. (2002). Negotiation for starting salary: Antecedents and outcomes
among recent college graduates. Journal of Business and Psychology, 16(3), 365-382.
156

O'Toole, Daniel E., and Brian Stipak. 1988. "Budgeting and productivity revisited: The local
government picture." Public Productivity Review: 1-12.

Osborne, David, and Ted Gaebler. 1992. "Reinventing government: How the entrepreneurial
spirit is transforming government." Reading Mass. Adison Wesley Public Comp.

Perry, J. L. (1997). Antecedents of public service motivation. Journal of public administration
research and theory, 7(2), 181-197.

Perry, J. L., & Hondeghem, A. (Eds.). (2008). Motivation in Public Management: The Call of
Public Service: The Call of Public Service. Oxford University Press.

Peterson, Paul E. 1981. City limits. University of Chicago Press.

Pinkerton, Steve. 2014. Personal Communication, Santa Monica, CA. April 16.

Poister, Theodore H., and Robert P. McGowan. 1984. "The use of management tools in
municipal government: A national survey." Public Administration Review(1984): 215-
223.

Poister, Theodore H., and Gregory Streib. "Performance measurement in municipal government:
Assessing the state of the practice." Public Administration Review: 325-335.

Powell, W. 2007. “The New Institutionalism.” In The International Encyclopedia of
Organization Studies. Thousand Oaks, Ca: Sage Publishers.

Protasel, G. J. 1988. Abandonments of the council-manager plan: A new institutionalist
perspective. Public Administration Review, 807-812.

Rau, Raghavendra, and Jin Xu. 2012. "Paying for risk or shareholder rip-off? An analysis of ex-
ante severance pay contracts." An Analysis of Ex-Ante Severance Pay Contracts (March
16, 2011). AFA.

Reilly, Thom. 2012. “Rethinking the Role of the Profession on Public Sector Compensation.”
Public Administration Review 73(1): 8-9.

Renner, Tari. 1988. “Elected Executives: Authority and Responsibility.” In Baseline Data Report
20, no. 3. Washington, DC. International City/County Management Association.

Renner, Tari. 2001. "The local government management profession at century’s end." The
municipal yearbook 2001: 35-46.

157

Renner, Tari, and Victor S. DeSantis. 1998. “Municipal Forms of Government: Issues and
Trends.” In Municipal Year Book 1998. Washington, DC. International City/County
Management Association.

Roberts, Norm. 2013. Personal Communication. June 13.

Rusticus, Tjomme O. 2006. "Executive severance agreements." PhD diss., University of
Pennsylvania.

San Bernardino Sun. 2009. Election Results Released by Registrar of Voters. San Bernardino
Sun. November 4.

Santerre, Rexford E. 1990. "A test of executive behavior in the local public sector." The Review
of Economics and Statistics: 546-550.

Santerre, Rexford E., and Laurie J. Bates.1996. "Performance and Pay in the Public Sector: the
Case of the Local Tax Assessor." Public Finance Review 24, no. 4: 481-493.

Schneider, Mark, and Kee Ok Park. 1989. "Metropolitan counties as service delivery agents: The
still forgotten governments." Public Administration Review: 345-352.

Simondes, Marge. 2009. Cooler Heads Welcome. San Bernardino Sun. May 16.

Smith, Adam. 1776. "The Wealth of Nations. Modern Library." New York 1937.

Stein, Robert M. 1990. "Urban alternatives." Public and Private Markets in the Provision of
Local.

Svara, James H. 1998. "The politics-administration dichotomy model as aberration."Public
Administration Review: 51-58.

Svara, J.H. 1999. U.S. City Managers and Administrators in a Global Perspective. Municipal
Year Book 1999. Washington, DC: International City/County Management Association.

Svara. 2003. Two Decades of Continuity and Change in America’s City Councils. Washington,
DC: National League of Cities.

Svara, J. H. (2010). The Next Generation Challenge: Incorporating the Local Government
Managers of the Future. Journal of Public Affairs Education, 361-377.

Stuhlmacher, A. F., & Walters, A. E. (1999). Gender differences in negotiation outcome: A
meta-analysis. Personnel Psychology, 52(3), 653-677.

158

Stillman, R. J. 1974. The rise of the city manager: A public professional in local government.
Albuquerque: University of New Mexico Press.

Stone, Harold Alfred, Don Krasher Price, and Kathryn Haeseler Stone 1940.. City manager
government in the United States: a review after twenty-five years. Public Administration
Service.

Tekniepe, Robert J., and Christopher Stream. 2012. "You’re Fired! County Manager Turnover in
Large American Counties." The American Review of Public Administration 42, no. 6:
715-729.

Thompson, Leigh. (1990). An examination of naive and experienced negotiators.Journal of
Personality and Social Psychology, 59(1), 82.

Tiebout, Charles M. 1956. "A pure theory of local expenditures." The journal of political
economy: 416-424.

Usher, Charles L., and Gary C. Cornia. 1981. "Goal setting and performance assessment in
municipal budgeting." Public Administration Review: 229-235.

Vuong, Zen. 2013. San Bernardino City Council Stands by New City Manager Allen J. Parker.
San Bernardino Sun. February 21.

Watson, C. (1994). Gender versus power as a predictor of negotiation behavior and outcomes.
Negotiation Journal, 10(2), 117-127.

Watson, Douglas J., and Wendy L. Hassett. 2003. "Long–Serving City Managers: Why Do They
Stay?." Public Administration Review 63, no. 1: 71-78.

Watson, Douglas J., and Wendy L. Hassett. "Career Paths of City Managers in America's Largest
Council‐Manager Cities." Public Administration Review 64, no. 2 (2004): 192-199.

Weibel, A., Rost, K., & Osterloh, M. (2010). Pay for performance in the public sector—Benefits
and (hidden) costs. Journal of Public Administration Research and Theory, 20(2), 387-
412.

Wilson, Woodrow. 1887. "The study of public administration."

Willon, Phil, and Abby Sewell. 2012. Plenty of Blame on Long Raod to San Bernardino
Bankruptcy. Los Angeles Times. July 12.

Wright, Deil S. (1969). The City Manager As A Development Administrator. In R. T. Daland
(Ed.), Comparative Urban Research (pp. 203-248). Beverly Hills, CA: Sage.

159

Zhang, Yahong, and Richard C. Feiock. 2007. "Career incentives and city managers’ policy
leadership." In Public Management Research Conference, Tucson, AZ, October, pp. 26-
27.

Abstract (if available)

Abstract While numerous public administration studies examine principal‐agent problems, only a narrow set of work investigates this relationship in the local context, and most of this work focuses on ex post municipal level outcomes. In this study, I advance the literature by linking municipal conditions to the ex ante contractual specification of principal‐agent relationships between council and manager. I develop a novel theory of risk shifting between principal and agent at the local level and predict that political constraints and financial conditions drive the risk shifting behavior of managers and councils. ❧ I create a unique municipal‐level data source containing economic, financial and political conditions in cities and towns throughout the state of California as well as information on the risk protection provisions granted to the city managers serving in these cities via their employment contracts. Using this data, I empirically test the predictions of the risk‐shifting theory. The results of the empirical analysis offer substantial support for the theory of risk‐shifting between council and manager and suggests that city managers recognize potential career risk when entering into a new employment arrangement and they seek to protect themselves and their professional reputation through risk‐shifting. As the career risk associated with a given manager position increases, the manager shifts that risk back onto the council through either severance protection or political protection. These results suggest that municipal economic, financial, and political conditions do impact the degree of career risk facing an incoming city manager, and as such, the manager seeks to shift that risk back onto the city council through insurance against and protection from termination. This study advances our understanding of principal‐agent problems within local government and presents implications for future research.

Linked assets

University of Southern California Dissertations and Theses

Conceptually similar

PDF

The effects of interlocal collaboration on local economic performance: investigation of Korean cases

PDF

The functions of the middleman: how intermediary nonprofit organizations support the sector and society

PDF

Essays on fiscal outcomes of cities in California

PDF

Why go green? Cities' adoption of local renewable energy policies and urban sustainability certifications

PDF

Governing regional collaboratives: institutional design, management and leadership

PDF

Civic associations, local governance and conflict prevention in Indonesia

PDF

Governing public goods: how representation and political power in local and regional institutions shape inequalities

PDF

How can metrics matter: performance management reforms in the City of Los Angeles

PDF

Governance networks: internal governance structure, boundary setting and effectiveness

PDF

Three essays on the causes and consequences of China’s governance reforms

PDF

Local votes and outside money: campaign contribution geographic origins and their impact on Los Angeles City Council election outcomes

PDF

Emergency-war machine: national crisis, democratic governance, and the historical construction of the American state

Asset Metadata

Creator Connolly, Jennifer Marie (author)

Core Title The risks and rewards of city management: how city managers evaluate the nature of the job and compensation

School School of Policy, Planning and Development

Degree Doctor of Philosophy

Degree Program Public Management

Publication Date 08/05/2014

Defense Date 08/05/2014

Publisher University of Southern California (original), University of Southern California. Libraries (digital)

Tag city management,city manager,employment contract,Local government,OAI-PMH Harvest,principal agent,Public Administration,Public Management,risk negotiation,risk shifting

Format application/pdf (imt)

Language English

Contributor Electronically uploaded by the author (provenance)

Advisor Bertelli, Anthony M. (committee chair), Grose, Christian (committee member), Tang, Shui Yan (committee member)

Creator Email jenniferconnolly2@gmail.com

Permanent Link (DOI) https://doi.org/10.25549/usctheses-c3-451798

Unique identifier UC11286990

Identifier etd-ConnollyJe-2759.pdf (filename),usctheses-c3-451798 (legacy record id)

Legacy Identifier etd-ConnollyJe-2759-0.pdf

Dmrecord 451798

Document Type Dissertation

Format application/pdf (imt)

Rights Connolly, Jennifer Marie

Type texts

Source 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 a...

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