University of Southern California Dissertations and Theses

Essays on the politics of the American welfare state

(USC Thesis Other)

Essays on the politics of the American welfare state

PDF

Download a page range

Download transcript

Copy asset link

Request this asset

Transcript (if available)

Content ESSAYS ON THE POLITICS OF THE AMERICAN WELFARE STATE
by
Dimitris Pipinis
A Dissertation Presented to the
FACULTY OF THE USC GRADUATE SCHOOL
UNIVERSITY OF SOUTHERN CALIFORNIA
In Partial Fulfillment of the
Requirements for the Degree
DOCTOR OF PHILOSOPHY
(ECONOMICS)
August 2012
Copyright 2012 Dimitris Pipinis
Acknowledgments
During my years at USC, a number of individuals played an important role in the
development of my research and the completion of this dissertation. Specically, I would
like to thank Jerey Nugent, who as an advisor was so supportive since the beginning of
my academic career at USC. I greatly appreciate his advice and understanding especially
since my research interests gradually diverged from his and economics in general. I would
also like to thank Janelle Wong for her encouragement at a very crucial stage of my studies
at USC. I consider myself very lucky to have taken Janelle's course in American Politics in
Fall 2008, as this experience played a decisive role in my decision to pursue my interests
in political science. I would nally like to thank Richard Easterlin both for serving in my
dissertation committee as well as for his interest in my work during the last phase of this
dissertation.
My experience at USC was made richer by the presence of a select group of friends and
colleagues. Special thanks go to Mohamed Saleh who has been a great friend and as a true
scholar himself has always been interested in providing feedback on my research. I would
also like to thank Martin Weidner and Giorgos Eraimidis who, apart from being close
friends, were very eager to share their expertise on some rather esoteric econometric issues.
My development as an individual and a scholar has been decisively shaped by my family.
I am grateful to my parents who always, and by example, taught me the value of questioning
established knowledge and who unyieldingly supported me in my decision to go down less
well trodden paths. For this reason, I am forever indebted to them.
ii
Table of Contents
Acknowledgments ii
List of Tables iv
Abstract v
Chapter 1: Bureaucracy's \Color" and the American Welfare State: The Adminis-
tration of Sanctions in TANF 1
1.1 Context:Temporary Assistance for Needy Families . . . . . . . . . . . . . . . 3
1.2 Data . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4
1.3 Empirical Strategy . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10
1.4 Sanctions and the Racial Characteristics of Welfare Bureaucracy . . . . . . . 11
1.5 Interpreting the Findings . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14
1.6 Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20
Chapter 2: Late Champions of Healthcare Reform: Two Tier Contracts, \Fair Share"
Bills and a \Coalition to Advance Healthcare Reform" 27
2.1 Theoretical Background and the Argument . . . . . . . . . . . . . . . . . . . 34
2.2 On The Sidelines: Healthcare Reform in 2003 . . . . . . . . . . . . . . . . . . 37
2.3 The Largest Retail Workers' Strike in the American History . . . . . . . . . . 39
2.4 The "Big Three's" Stance with Respect to Proposition 72 . . . . . . . . . . . 40
2.5 Into the Ring: Healthcare Reform in 2007 . . . . . . . . . . . . . . . . . . . . 42
2.6 Alternative Interpretations . . . . . . . . . . . . . . . . . . . . . . . . . . . . 48
2.7 Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 53
Bibliography 59
iii
List of Tables
1.1 Reasons for case closure by race of the household head . . . . . . . . . . . . . 21
1.2 Summary Statistics for individual/household level variables . . . . . . . . . . 22
1.3 Summary Statistics for county level variables . . . . . . . . . . . . . . . . . . 23
1.4 Sanctions and the characteristics of the Welfare Bureaucracy . . . . . . . . . 24
2.1 Chronology of a Puzzle . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 56
2.2 Key events 2003/04 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 57
2.3 Key events, 2007/08 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 58
iv
Abstract
The dissertation deals with key aspects pertaining to the politics of the Welfare State.
In the rst part of the dissertation, I examine whether, and how, policy implementation is
aected by the personal characteristics of the street level welfare bureaucrats. I then turn my
attention to issues surrounding the role of business in the development of progressive social
legislation. Specically, in the second part of the dissertation, I analyse the circumstances
under which a segment of the big business community came to adopt a positive stance
towards healthcare reform. The aforementioned questions are investigated in the context
of the United States. In what follows, I provide an overview of the questions and the key
ndings of the dissertation.
Chapter 1 examines how the personal characteristics of the street level bureaucrats aect
the implementation of welfare policy. While both theory and anecdotal evidence point to the
central role of low level bureaucracy in the administration of welfare programs in the U.S., no
previous empirical study has examined whether, and how, the personal characteristics of the
street level bureaucrats indeed aect how benets are allocated and sanctions are imposed
on welfare clients. This work attempts to ll this gap in the literature. By matching
individual level data on welfare clients with county level data on welfare workers, I nd that
the racial composition of the welfare bureacracy matters for clients. Specically, I show
that as the presence of whites among welfare workers grows stronger, both black and white
clients face an increased likelihood of being cut o from welfare. I argue that these results,
which are robust to controlling for the racial composition of the county, re
ect the more
conservative attitudes held by whites towards welfare and welfare recipients in general. This
v
work provides thus, for the rst time, systematic evidence that the racial characteristics of
the welfare bureaucracy aect how welfare policy is implemented at the street level.
Chapter 2 investigates the reasons behind big business mobilization in favor of health-
care reform in the years preceding the passage of the landmark Patient Protection and
Aordable Care Act of 2010. I specically examine why large unionized rms, such as the
\Big Three" food retailers, breaking ranks with the major business organizations, spear-
headed some of the most important state level reform eorts in the wake of Clinton's plan
debacle. I argue that their embrace for comprehensive health reform and employer man-
dates in particular re
ects their failure to vastly reduce their labor costs through radically
restructuring, along two-tier lines, their contracts with the unions. In light of their failure
to privately solve the health cost crisis, these rms turned to government for a solution.
Initially through the support for \Fair Share" legislation in a number of states and, later,
through the formation of ad hoc business coalitions, such as the \Coalition to Advance
Healthcare Reform" in California, the \Big Three" signaled their openness to employer
mandates, initially, and to more comprehensive solutions, eventually, along lines fundamen-
tally similar to the federal legislation passed in 2010. This work contributes both to the
understanding of the politics of healthcare reform in the United States as well as to the
larger literature on the role of business in the development of progressive social legislation
in capitalist societies.
vi
Chapter 1
Bureaucracy's \Color" and the American Welfare
State: The Administration of Sanctions in TANF
In the realm of the U.S. social policy, there are few programs that have been so decisively
linked to race as welfare.
1
As one author puts it \welfare is frequently the object of racially
focused antagonism and resentment, directed particularly at the apparent conditions of life
in our decaying inner cities-idleness, immorality, family decay, and crime."(Lieberman 1998).
An important line of literature has documented how racial considerations have shaped the
scope and nature of the welfare state in the United States (Lieberman 1998; Quadagno
1994). Racial considerations, however, may have not only in
uenced policy design but
policy implementation itself. In the literature, one nds ample, albeit anecdotal, evidence
that welfare caseworkers have historically used their discretion to discriminate against black
recipients (Bell 1965; Lieberman 1998; Piven and Cloward 1971).
2
In other words, race
has been a factor aecting how, once designed, the rules of the program are applied by the
\street level" bureaucrats, to use Lipsky's terminology.
1
Welfare here refers to programs such as the Temporary Assistance for Needy
Families (TANF). While TANF constitutes only a small slice of the welfare state in
the United States, it has been the object of most attention in the American public's
mind (Lieberman 1998).
2
For example, in their classic study Piven and Cloward (1971) argue that \if an
impoverished woman and her children who apply for welfare in Mississippi happen to
be black, they will have to contend with welfare ocials who will try to embarrass,
scold, or frighten them into withdrawing their application" (italics mine).
1
Building on Lipsky's theoretical insight on how low level bureaucracy shapes public
policy, this paper aims at shedding light on the interplay between the race of \street level
bureaucrats" and that of clients in the context of the prototypical welfare program in the
United States, namely, the Temporary Assistance for Needy Families (hereafter, TANF).
TANF oers a particularly suitable area for the study of how bureaucrats' race aects policy
implementation and, eventually, client outcomes for two reasons. First, welfare caseworkers
have long been recognized as a prime example of street level bureaucrats.
3
Second, as it
was argued above, programs such as TANF are considered racially charged and, thus, one
would expect that race would be one of the most salient characteristics in that context.
Despite both the salience of race in the context of welfare and the wide discretion welfare
caseworkers enjoy in applying the rules of the program, no previous work has examined if,
and to what extent, the race of welfare caseworkers aects the way benets are allocated
and penalties are imposed across clients of dierent backgrounds. This, however, is an
unfortunate omission given that both theoretical work (Lipsky 1980) and anecdotal evidence
(Handler and Hasenfeld 2007) strongly suggest that welfare workers' personal characteristics
continue to play a crucial role in how welfare policies are implemented.
4
This article aims
at lling this gap in the literature.
In order to tackle this question, I rely on two sources of data. First, I employ individ-
ual/household level data of TANF recipients which are available through the Department
of Health and Human Services (hereafter, DHHS). The DHHS data allow me to distinguish
between clients who were involuntarily cut o from the program (i.e. due to sanctions) and
those who were not (e.g. no longer satisfying the eligibility criteria). The administration of
sanctions is an appropriate area to study how the characteristics of low level bureaucrats
3
Lipsky (1980) identies welfare workers as a typical example of \street level
bureaucrats".
4
For example, Handler and Hasenfeld (2007), writing in the context of the TANF
program, argue that \the worker's own personal beliefs and experiences in
uence how
they implement welfare policies."
2
aect policy implementation as it is well documented that caseworkers have signicant dis-
cretion in applying the rules of the program, especially with regards to granting \good cause"
exceptions from work requirements. As Fording et al. (2011) put it \it is the case manager
who must decide whether to initiate sanction procedures in response to an infraction and
whether the circumstances of the infraction justify a \good cause" exception to the rules."
(italics mine). Second, I use data from a special tabulation of the 2000 U.S. census titled
the Equal Employment Opportunity tabulation (hereafter, EEO). The EEO data allow me
to construct a measure of the racial composition of welfare caseworkers at the county level.
By matching the individual level DHHS data to the county level EEO data, I am able to
examine whether, and how, the racial composition of the low level bureaucracy aects each
client's likelihood to receive a sanction.
1.1 Context:Temporary Assistance for Needy Families
TANF, like its predecessor, the Aid to Families with Dependent Children (hereafter,
AFDC), is designed to work as a safety net for poor families with dependent children.
However, and consistent with the increased focus on the promotion of work norms, TANF
includes not only stricter eligibility requirements and time limits but it is also characterized
by the more extensive use of sanctions. As some authors have argued, \sanctions have come
to play a more prominent role in welfare implementation; compared to the AFDC program,
the TANF program compels clients to face stricter work requirements, narrower exemption
criteria, a greater number of behaviors subject to sanction, and stronger penalties for non-
compliance" (Hasenfeld and Powell 2004). While sanctions can be applied for a variety of
reasons, the vast majority is applied for non compliance with the work requirements.
5
Par-
alleling the increasing use of sanctions is the greater devolution of power to the states and,
in some cases, to the local governments (i.e. counties). In its turn, the devolution of power
5
According to a report prepared by the General Accounting Oce in 1998, the
majority of sanctions were imposed for failure to comply with work related require-
ments. Other reasons include failure to cooperate with child support enforcement
and non compliance with immunization or children's school attendance responsibili-
ties (GAO 2000).
3
to the states has led to wide interstate variation in the formal rules of the program. Many
states, for example, have opted for stricter policy rules compared to the federal guidelines
as far as sanction policies, work requirements and time limits are concerned.
6
Moreover,
the emphasis put on the provision of
exible services as well as on the use of sanctions
meant that \decision making has been pushed downward to local implementers, including
case managers, who serve as the primary street level bureaucrats in the new world of welfare
service delivery" (Fording et al. 2007). Although there does not seem to exist a consensus
in the literature on whether the discretionary powers of the caseworkers actually increased
or not in the aftermath of the 1996 reform, there is little doubt that caseworkers continue to
exercise signicant discretion in how they implement program rules (Sandfort 2000). The
imposition of sanctions is a prime area where this discretion can be exercised.
In its turn, the substantial discretion that welfare workers enjoy in implementing the
rules of the program implies that any prejudices and biases permeating a given society can
nd their expression in the decisions of the low level bureaucracy (Lipsky 1980). As a con-
sequence, rules may be interpreted arbitrarily depending on the characteristics of both the
caseworker and the client. To examine how the interplay of caseworker and client character-
istics aects policy implementation, one needs information on both sets of characteristics.
The next section is dedicated in describing in detail the data used to that end.
1.2 Data
My empirical strategy is based on matching the individual/household DHHS data with
the county level data from the EEO tabulation and the 2000 U.S. census.
7
For reasons
6
Previous research has shown that racial considerations may have in
uenced policy
choices at the state level. For example, Soss et al. (2001) show that states with a
higher percentage of African-Americans were more likely to adopt \get tough" policies
(i.e. shorter time limits, stricter work requirements etc).
7
I matched the county level data to the DHHS data based on the 2000 Federal
Information Processing Standards (FIPS) codes. In two states (Maryland and Mon-
tana) FIPS codes were not available in the DHHS data. In these cases I assigned
FIPS codes to each household based on the zip code provided in the data. I excluded
households residing in zip codes which could not be uniquely matched to a county.
4
which will be discussed later I split my sample based on the racial/ethnic characteristics
(i.e. Hispanic or not, white, black, etc) of the clients. This approach works well for white
and black clients but not for Hispanics due to the relatively few Hispanic households in my
sample.
8
As a result, the focus of my analysis is on non Hispanic black and white house-
holds.
9
I, therefore, create two samples. The white (black) sample consists of households
where the household head was identied as non Hispanic white (black). In what follows, I
discuss in some detail my primary sources of data.
Information pertaining to TANF clients is drawn from administrative data provided
through the DHHS. The data which are available for all years between 1998 and 2007
include information on both active as well as closed cases. In this paper, I make use of the
closed cases le for 2000 which includes data from 48 states and the District of Columbia.
10
The le has information on households that left the program during a twelve month period
starting in October 1999 and ending in September 2000. Following Fording et al. (2011),
I restrict my analysis to adult TANF recipients identied as household heads in the data.
This means excluding \child only" cases. In these households, only the minor members are
eligible for assistance and, therefore, the household head is not subject to work requirements
and she rarely faces sanctions (Keiser et al. 2004). The DHHS data include a wide range of
individual as well as household level variables. More importantly, they contain information
that allows me to distinguish between households that left TANF due to sanctions and those
that left TANF for other reasons. Table 1.1 provides a breakdown based on the reason of
8
For example, there are 9273 household heads identied as non Hispanic white but
only 1018 white Hispanic household heads. The contrast is even sharper for black
households with 10197 household heads identied as non Hispanic black but only 127
identied as Hispanic black.
9
These two groups combined accounted for almost three quarters of the house-
holds who left welfare during the period under examination (Department of Health
and Human Services, 2001). Specically, 37.5% of the closed cases involved white
households, 36% black households and 21% Hispanic (of any race). The remaining
(< 5%) includes Asian, Native American, Hawaiian, etc.
10
Oregon did not report data for that year, while data from Louisiana are reported
in error (Department of Health and Human Services website).
5
exit from the program for the black and white samples separately. All tables presented in
this section are based on the matched data used in the empirical analysis.
The DHHS data identify four broad categories of reasons for case closure. The rst
category includes cases closed because units were no longer eligible for benets (e.g. excess
earnings). In the second category, all cases closed due to sanctions are included. The
third category includes case closures related to \state (tribal) rules". The vast majority of
households belonging to this category have their cases closed due to failure to submit the
verication materials required for their applications to be processed. The remaining cases
are closed either because \the family voluntarily closes the case" or for other reasons. These
cases comprise the fourth and last category. A few observations are in order here. First, a
non negligible number of cases are closed due to sanctions (8% in the \black" sample and
6% in the \white" sample). Second, the vast majority of sanctions are imposed for failure
to satisfy work requirements (71% in the \black" sample and 62% in the \white" sample).
This later nding highlights the central role of the work requirements as a disciplinary tool
in the post 1996 period.
In the empirical analysis, the outcome variable is binary (whether a household was
sanctioned or not) and, therefore, I use logit/probit models which are standard in the
analysis of such data. In choosing the independent variables, I rely on previous literature on
the determinants of sanctions. I follow most closely Fording et al. (2011) as it is one of the
few papers that are employing the DHHS data.
11
In addition to the controls used in Fording
et al. (2011), I include a variable measuring the time the household has been on TANF (i.e.
months the household has been receiving benets) as previous research has shown that it
aects the likelihood of the imposition of a sanction (Wu et al. 2006). Summary statistics
of the variables constructed from the DHHS data are presented in table 1.2.
As it is evident from table 1.2, single parenthood is substantially more common among
households headed by blacks than among those headed by whites (82% vs 54%). Blacks,
also, appear to stay in the rolls longer (15 vs 13 months) and to have more children on
11
Kim (2008) and Monnat (2010) are also using data from the same source.
6
average. Finally, black households, compared to white households, are more likely to reside
in public housing. Few other noticeable patterns emerge. I now turn to the discussion of
the data on welfare bureaucrats.
Information on the characteristics of the local welfare bureaucracy is gleaned from a spe-
cial tabulation of the 2000 U.S. census. The EEO tabulation provides information on the
821 detailed occupations of the Standard Occupational Classication (hereafter, SOC) bro-
ken down by race, sex, education and age. In my analysis, I use worksite data on \Eligibility
Interviewers, Government programs" (SOC code 43.4061.00). According to the description,
eligibility interviewers are required to \Determine eligibility of persons applying to receive
assistance from government programs and agency resources, such as welfare, unemployment
benets, social security, and public housing." Among the tasks these workers are required
to perform is to \determine eligibility status and initiate procedures to grant, modify, deny,
or terminate assistance, or refer applicants to other agencies for assistance" (italics mine).
An examination of the sample of job titles provided makes it clear that these workers are
the street level bureaucrats charged with how welfare programs are administered at the
frontlines.
12
The EEO data allow me to construct measures of the racial, sex, age and
educational distribution of eligibility interviewers at the county level. Apart from the racial
composition, which is of prime interest here, there are good theoretical reasons for including
the additional variables. For example, Meyers and Riccucci (2004) use caseworker's race,
gender, education and years of experience in the job (in my case proxied by age) as predic-
tors of caseworker attitudes which, as they claim, may eventually in
uence policy outcomes
12
The following job titles are reported: \Workforce Services Representative (WSR),
Eligibility Specialist, Career Consultant, Work Force Advisor, Business Employment
Specialist, Client Services Representative, Eligibility Technician, Employment Adjudi-
cator, Employment Specialist, Family Independence Case Manager". Relying both on
job search banks (e.g. http://www.simplyhired.com) as well as on secondary sources
(e.g. Lurie 2006; Riccucci 2005), I was able to verify that nine out of ten occupational
titles refer to workers employed as frontline personnel in the administration of TANF
and the related workforce services. An exception is \employment adjudicators", who
are responsible for unemployment benets.
7
(such as the imposition of sanctions).
13
Overall, the EEO data allow me to control for a vari-
ety of caseworker characteristics that may in
uence policy implementation. However, they
are not without limitations. First, due to condentiality reasons, detailed tabulations based
on race and education/age are available only for counties with a population greater than
100000. This limits the number of available counties and, therefore, reduces the variation
I am able to use in order to identify the eect that the characteristics of the local welfare
bureaucracy have on sanctioning rates.
14
Second, and perhaps more serious, is the fact
that among eligibility interviewers are counted frontline workers employed in government
programs other than welfare (i.e. unemployment insurance, social security etc). There
is, of course, no theoretical reason to expect that the characteristics of these street level
bureaucrats would be relevant for how TANF clients fare. Obviously, to the extent that the
characteristics of eligibility interviewers do not systematically dier among agencies, there is
little reason for concern. However, this assumption does not necessarily hold. For example,
eligibility interviewers working in federally administered programs (i.e. social security) may
be systematically dierent from those working in state or locally administered programs (i.e.
welfare). In technical terms, the inclusion of these workers means that the characteristics
of the welfare workers, which are of theoretical interest here, would be measured with some
noise (i.e. measurement error). While, usually, the presence of measurement error leads to a
downward bias for all regression coecients (i.e. classical measurement error), this need not
be the case here. Specically, since the range of the variable is limited (e.g. the percentage
white among eligibility interviewers is bounded between 0 and 1), the measurement error
will be non classical (Angrist and Krueger 1999). The presence of non classical measurement
error implies that predicting the direction of the bias is not straightforward. In either case,
13
They claim that \To the extent that social origins predict the values or attitudes
of caseworkers in welfare agencies, social origins may explain variation in policy out-
comes (for example, the granting of cash benets or the imposition of sanctions and
obligations)."
14
I did not include parishes/counties from Louisiana and Oregon since the later did
not report data for that year, while data from Louisiana are reported in error. Also,
in Wyoming no county has population greater than 100000.
8
however, the impact of the measurement error will crucially depend on how \noisy" the
data are. Fortunately, only a small fraction (around one fth) of eligibility interviewers are
employed in the federal government, while the vast majority of interviewers are employed
either in state or local governments.
15
To the extent, therefore, that the prime source of
measurement error stems from dierences between eligibility interviewers working in federal
vs state/local programs, the impact of measurement error may indeed be limited. Conse-
quently, I argue that the characteristics of \eligibility interviewers, government programs"
would adequately re
ect the characteristics of the welfare workers administering TANF.
To summarize, while the existence of measurement error recommends additional caution in
interpreting the empirical ndings, I expect that the aforementioned data limitations impose
constraints of rather second order importance.
Apart from the characteristics of eligibility interviewers, I include a number of contextual
variables which have often been used in the literature on sanctions (i.e. % unemployment, %
under poverty, per capita income, population, % Republican). I, also, account for the racial
composition of the county at large. While both theory and empirical evidence support the
notion that racial context in
uences sanction rates (e.g. Keiser et al. 2004; Soss et al. 2009),
my primary motivation for accounting for the racial context is methodological. Specically,
since my interest lies in how the racial composition of the local welfare bureaucracy, and
not of the county in general, aects sanctioning, I have to warrant against the possibility
that any eect I nd is a re
ection of the later. To that end, in my analysis, I control for
the racial composition of the county at large. Summary statistics for all the county level
variables are presented in Table 1.3.
15
In 2006 there were 104240 workers classied as eligibility interviewers. According
to the North American Industry Classication System (hereafter, NAICS), 23970
(22.9%) of the eligibility interviewers were employed in the federal government
(NAICS code 999100) and 76220 (73.1%) were employed in the state or local gov-
ernments (NAICS codes 999200 and 999300). The remaining 4% were employed in
other sectors (i.e Grantmaking and Giving Services, Social Advocacy Organizations,
etc). These calculations are based on data that are available through the Department
of Labor website.
9
Note that the number of counties diers between the white and black sample. The reason
is that there are counties where only black or only white clients are observed.
16
Looking
at Table 1.3, however, reveals that, apart from the fact that black clients tend to reside in
more populated counties on average, there are few noticeable dierences between the two
samples as far as county characteristics are concerned. Crucially, the main ndings of the
paper are robust to including all counties or restricting attention only to counties where
both white and black clients are observed. Table 1.3 also shows that, typically, eligibility
interviewers are female, non Hispanic white and have some college education. Other groups
have, also, substantial representation with 16-18% being black on average and around 8%
being Hispanic. Finally, there is substantial variation in the characteristics of eligibility
interviewers across counties.
1.3 Empirical Strategy
I estimate the following equation using standard logit/probit models.
Probability(Sanction) =F (X
ics
;Z
cs
;D
s
) +u
ics
(1.1)
F, the probability of sanction for client i, residing in county c and state s, is a function of
individual level characteristics X and county level characteristics Z. I, also, include a set of
state dummies D. As mentioned earlier, in the aftermath of the 1996 reform, wide variation
in policy rules developed across states. Thus, including state xed eects allows me to
examine the eect of bureaucratic characteristics in an environment where the formal rules,
by which welfare workers should abide, are kept constant across localities (i.e. counties).
17
16
The counties where only black or white clients are observed are 7 and 86 respec-
tively.
17
Some states have devolved greater authority for the administration of the program
to counties. While this means that each county may have more leeway in designing
how to implement the program, it is still the case that program rules (i.e. severity of
sanctions) are set at the state level. As Gainsborough (2003) informs us \in almost
all instances the state still retains control over who is eligible for services and benets,
how much in benets they will receive, and what types of services must be provided".
10
I estimate (1.1) separately for the black and the white sample. This strategy allows me to
directly test how, keeping clients' race/ethnicity constant, the characteristics of the local
welfare bureaucracy aect the imposition of sanctions. It, also, addresses some concerns
about endogeneity. For example, early work on sanctions acknowledged that higher sanction
rates for blacks may be driven by unobserved factors such as labor market discrimination
(Kalil et al. 2002). Blacks may, also, face higher sanction rates due to residential segregation
which limits their employment opportunities and may, thus, increase diculty in obtaining
and keeping jobs (i.e. \spatial mismatch" hypothesis).
18
By estimating (1) separately by
race, I avoid these complications without incurring any cost. Recall that my aim here
is not to examine how the race of the clients aects sanctioning rates, a topic of a well
developed literature, but rather to assess how the racial characteristics of the bureaucracy
aect policy implementation. Finally, the nested nature of the data may raise concerns
about the structure of the error term (i.e. autocorrelation). While the inclusion of state
xed eects corrects for autocorrelation at the state level, it cannot address the possibility
that the errors are correlated at lower geographical levels (i.e. within a county). To that
end, the standard errors are adjusted for clustering at the county level. The results from
the estimation of (1.1) above are presented in the next section.
1.4 Sanctions and the Racial Characteristics of Welfare Bureaucracy
Table 1.4 presents the results from estimating (1.1) using a logit model.
19
I estimate
but do not report state xed eects for all specications.
In the rst two columns, I report the results when I include only the individual and
contextual variables. Although not of prime theoretical interest here, the results on the
individual determinants of sanctions are noteworthy, because they suggest that the eect
of these variables may be contingent on the race of the client. I nd, for example, that
18
See Ihlanfeldt and Sjoquist (1998) for an earlier review of the empirical evidence
on the spatial mismatch hypothesis. For more recent evidence, see Stoll (2006) and
Boustan and Margo (2009).
19
The results are robust to using a probit model.
11
single parenthood increases the risk of sanctions for black clients but not for whites. Also,
while the number of children increases the risk of a sanction for both white and black
clients, the eect is statistically signicant only for black clients. On the contrary, the
length of welfare receipt appears to increase the risk of sanctions for white TANF clients
but not for blacks. At the same time and consistent with previous ndings (Hasenefeld et
al. 2004; Keiser et al. 2004; Fording et al. 2011), I nd that younger clients and those
with older children are more likely to be sanctioned. This holds true for both white and
black clients. Turning to contextual variables few interesting patterns emerge. As it has
been analyzed earlier, it is crucial for the present analysis to control for the composition of
the county in general. I, therefore, include measures of the racial composition of the county
in general (the reference group is Hispanics). In columns 3 and 4, I add the percentage
of non Hispanic whites among eligibility interviewers as a regressor. I, also, control for
the age, sex and educational distribution of eligibility interviewers. As mentioned in the
previous section, all these variables can potentially aect policy implementation as previous
work on public administration has argued. The results indicate that neither education nor
sex distribution aect sanction rates for welfare clients. However, I nd some evidence
that the age distribution of eligibility interviewers may aect sanction rates. Specically,
I nd that as the percentage of eligibility interviewers over 60 increases, the likelihood of
a sanction decreases for both white and black clients (the reference group is those aged
between 16 and 39). These results are consistent with the ndings reported in Soss et
al. (2009). In their paper, Soss et al. (2009), using hypothetical vignettes, nd that
more experienced caseworkers were less likely to recommend sanctions. Although I only
indirectly observe experience (i.e. through age) here, the fact that sanction rates are lower
when the percentage of older caseworkers increases provides some support to the idea that
more experienced welfare workers may use sanctions less frequently.
20
20
Soss et al. (2009) argue that more experienced caseworkers may be less likely to
impose sanctions for a variety of reasons. For example, \case managers with more
experience are more likely to have been trained in an earlier era of welfare provision
that placed less emphasis on sanctioning" or \alternatively if case managers learn
over time that sanctions have negative consequences for clients or for performance
12
The results presented in columns 3 and 4 provide the rst indication that indeed the
racial composition of the local welfare bureaucracy may aect how sanctions are imposed.
Specically, the coecient on non Hispanic whites is positive and signicant in both the
white and the black sample suggesting that, as the percentage of non Hispanic whites
among eligibility interviewers increases, all clients face a higher likelihood of receiving a
sanction. Alternatively, as non Hispanic whites replace eligibility interviewers of other
races/ethnicities, the sanction rates increase for both black and white clients. In the next
two columns I repeated the analysis by replacing the percentage of non Hispanic whites with
that of blacks. The results indicate that although the percentage of non Hispanic blacks has
a negative impact on the likelihood of a sanction the eect is not statistically signicant.
21
Given that the logit coecients are not directly interpretable, I calculated the odds ratios.
The odds ratio for the percentage white (non Hispanic) ranges from 1.79 (black sample)
to 2.04 (white sample). This means that as one moves from a purely non white welfare
bureaucracy to one that is all white, the odds of receiving a sanction almost double.
While these results are intriguing, the analysis remains incomplete. Specically, by
lumping all other racial groups into the omitted category, one may fail to uncover a more
nuanced picture on how the racial composition of the welfare bureaucracy aects sanction
rates. In order to examine this possibility, I experimented by omitting dierent racial groups
at a time. I nd that while the coecient on non Hispanic whites remains always positive,
it reaches statistical signicance only when the omitted category is Hispanic whites. These
results are reported in columns 7 and 8. To summarize, the results presented here reveal
three important patterns. First, they show that increased presence of non Hispanic whites
among the local welfare bureaucracy is associated with a more frequent use of sanctions.
Second, this eect becomes more pronounced as non Hispanic whites replace Hispanic whites
ratings, then more experienced case managers might become more reluctant to impose
sanctions".
21
The analysis was repeated for the remaining racial groups (\Hispanic white", \His-
panic other" and \other race"). Although the coecients were in all cases negative
only the coecient on white Hispanics reaches statistical signicance.
13
among eligibility interviewers. Finally, a stronger presence of blacks among eligibility inter-
viewers does not have a statistically signicant eect on sanction rates. While a discussion
of the empirical ndings has to be postponed for the next section two observations can be
made.
First, the main ndings of this section are robust to whether I include only the counties
where I observe both white and black clients or all counties available as in table 1.4.
22
Second, it is important to note that since I controlled for the racial composition of the
county at large, the coecient on non Hispanic whites should capture mechanisms related
to the presence of this particular group inside the welfare bureaucracy and not at the county
at large. What these mechanisms may be is the focus of the next section.
1.5 Interpreting the Findings
According to the results of the empirical analysis, as the presence of non Hispanic whites
inside the welfare bureaucracy grows stronger, sanction rates increase for all clients. In
addition, this eect appears to be primarily driven by dierences between Hispanic and non
Hispanic white eligibility interviewers. What are, however, the mechanisms driving these
results?
Work on public opinion and, specically, on dierences in attitudes towards welfare
programs across racial groups motivates a potential explanation. Previous work has shown,
22
The coecient on non Hispanic whites remains positive and signicant when I
limit my sample to counties where both black and white clients are observed. The
coecients for \Black (non hispanic)", \Other Race" and \Hispanic (other race)" are
negative and insignicant when entered separately, while the coecient on \Hispanic
(white)" is negative and signicant. These patterns exactly match the results reported
in table 1.4 as well as in note 21. At the same time, when I restrict my sample to
counties where both black and white clients are observed, I nd that apart from the
coecient on white (non Hispanic), which remains signicant (and positive) in all
cases, the coecients on black (non Hispanic) and Hispanic (other race) also reach
statistical signicance (their signs do not change) in the \white" sample (column 8)
but not in the \black" sample (column 7). A few other changes occur. For example, in
column 7 the coecients on the age groups are no longer signicant, while in column
8 the coecient on percentage Republican becomes signicant. Given, however, that
these results do not hold when I expand the sample to include all available counties,
I opted for the most robust specication (i.e. including all available counties).
14
for example, that whites are less likely to hold positive attitudes towards welfare spending
compared to blacks (Kinder and Winter 2001). This line of research has more recently
expanded to include other ethnic/racial groups. For example, Nicholson et al. (2005)
show that both Hispanics and Blacks, compared to whites, are more likely to show support
for social programs which are only implicitly related to race (i.e. national health care).
Their ndings are consistent with the results presented in Hunt (2007) who, using data
from the General Social Surveys, nds that both Hispanics and blacks are signicantly less
likely than whites to endorse purely motivational explanations for the black socioeconomic
disadvantage.
23
This later nding is important because, in his pioneering work, Gilens
(1999) nds that white support towards welfare is primarily motivated by perceptions about
black work ethic. To the extent, therefore, that Hispanics' attitudes towards welfare are
formed in ways similar to those of blacks, it is to be expected that Hispanics, on average,
would also be more supportive of programs targeting the poor.
Research on public opinion thus reveals systematic dierences between groups as far as
attitudes towards the poor are concerned. It, however, remains unclear how these dierences
are mapped among welfare workers. Theory suggests that this may be so. For example,
Handler and Hasenfeld (2007) argue that the fact that most welfare workers are not pro-
fessionally trained allows more room for their personal values to play an important role in
how they view their authority as policy administrators. Research on welfare bureaucracy
provides some support for this thesis. For example, Riccucci (2005) nds that white welfare
workers are more likely, compared to workers of color, to endorse goals more consistent with
a disciplinary view of TANF. Specically, she nds that white caseworkers, compared to
workers of color, assign higher priority to goals such as deterring welfare use, preventing
fraudulent claims and pushing recipients to accept employment etc. These ndings suggest
23
Respondents were asked to choose between Yes or No on the following statement
\On the average, blacks have worse jobs, income, and housing than white people. Do
you think these dierences are because most blacks just don't have the motivation or
will power to pull themselves out of poverty?"Hunt (2007) nds that after controlling
for a host of predictors (e.g. education, age, class etc) blacks and Hispanics are less
likely, compared to whites, to answer the aforementioned question armatively.
15
that opinion cleavages in the general population may nd their re
ection inside the welfare
bureaucracy. In that case, the more cynical attitudes whites hold towards the \undeserving"
poor are re
ected in the greater emphasis they put on goals which are more consistent with
viewing welfare recipients as \undeserving" poor (e.g. preventing fraudulent claims). As a
result of the generally unfavorable attitudes towards welfare recipients among whites as well
as because welfare departments have been rarely professionalized, it becomes all the more
likely that personal values in
uence workers' conduct toward clients (Handler and Hasen-
feld 2007). In this case, clients residing in counties with more non Hispanic whites among
welfare workers, by virtue of having a higher likelihood to have their cases assigned to non
Hispanic whites, may run an increased risk to have their cases closed due to sanctions. This
is what I nd in this paper.
The rationale developed here can explain why there is a positive association between
the percentage of whites among eligibility interviewers and the risk of sanctions (Table 1.4,
columns 3 and 4). It cannot fully account, however, for why this association is primar-
ily driven by dierences between non Hispanic and Hispanic white eligibility interviewers
(Table 1.4, columns 7 and 8). Recall that when I chose non Hispanic blacks as the reference
group, I found a positive but statistically insignicant association between the percentage
of non Hispanic white eligibility interviewers and sanction rates.
24
This result is particu-
larly interesting since, based on the literature on public opinion, one may have expected
that, if anything, the dierences in policy implementation would have been the starkest
when whites replace blacks among eligibility interviewers. What could explain this nding?
Recent ethnographic work on welfare workers motivates a potential explanation. Specically,
in her work Watkins-Hayes (2009) argues that the precarious economic and institutional
standing of black and Hispanic welfare workers (i.e. themselves members of traditionally
disadvantaged groups) may render them particularly fearful of appearing partial to their
own racial group. For the purposes of the analysis here, this may, on the one hand, make
24
The same holds true when I chose \Hispanic (other race)" and \other race" as
the omitted category.
16
black caseworkers less willing to dierentiate themselves (i.e. by applying sanctions more
leniently) from their white colleagues, especially when treating black clients. On the other
hand, Hispanics may be less constrained in dierentiating themselves from whites when
treating the non Hispanic clients who comprise my sample. As a result, for non Hispanic
clients the greater dierences may emerge as Hispanics replace non Hispanic whites inside
the welfare bureaucracy.
25
The argument I advanced here rests on the assumption that clients are randomly assigned
to caseworkers. At the most basic level this means, for example, that as the presence of non
Hispanic whites among eligibility interviewers increases both black and white clients become
more likely to interact with a white caseworker. It is exactly the higher likelihood of inter-
acting with whites which I argue leads to higher sanction rates for all groups. Furthermore,
my interpretation requires that clients are not sorted across counties in ways that could be
aecting the main ndings of the paper. For example, it is not the case that \harder to
serve" clients (i.e. who are more likely to receive a sanction anyway) reside in counties with
a higher percentage of non Hispanic whites among eligibility interviewers. As both of these
assumptions are crucial for my argument, it is important to discuss their plausibility.
Clearly, the argument will be greatly weakened if assignment of clients to caseworkers
depends on either's characteristics, most prominently race. There is some evidence that,
in the past, such considerations played an important role. For example, Neubeck and
Cazenave (2001) argue that in the pre civil rights South it was \almost unimaginable that
black caseworkers would be assigned to investigate and determine eligibility of white mothers
for ADC". Fortunately, overtly discriminatory practices of that kind appear to be a thing
of the past. According to public administration scholars, today, the assignment of clients
to caseworkers does not depend on either client's or caseworker's characteristics (Meyers
25
This argument cannot obviously account for the weak association between non
Hispanic whites and sanction rates I observed when I chose \Hispanic (other race)"
or \Other race" as the reference group. However, especially the \other race" category
is too heterogeneous (comprises of Asians, Native Americans, Hawaiians etc) to allow
for sharp predictions on how their presence should aect the application of sanctions.
17
2011; Riccucci 2011).
26
The assumption, therefore, of the random assignment of clients
to caseworkers within a county appears plausible. An increase in the percentage of non
Hispanic white eligibility interviewers means that all clients will be more likely to have
their cases assigned to a caseworker who is non Hispanic white.
Even if clients are assigned to caseworkers randomly, it is still possible that clients
residing in counties with more whites among welfare workers are \harder to serve" and,
thus, more likely to receive a sanction anyway. In other words, clients are not randomly
sorted across counties. In that case, the relationship I observe between the likelihood of a
sanction and the presence of whites among eligibility interviewers is not causal but, rather, a
re
ection of unobserved characteristics correlated with both. Obviously, the cross-sectional
nature of the data does not allow me to fully address econometrically this possibility.
27
To
address this issue, I relied on existing literature on sanctions in order to identify the relevant
control variables. The use of an extensive list of controls both at the individual/household
as well as at the county level and the inclusion of state xed eects should alleviate some
concerns about omitted variables. In addition, by splitting my sample by race, I avoid any
complications that may arise due to unobserved factors correlated with race. Even then,
it may still be argued that variables correlated with the racial composition of the welfare
bureaucracy aect sanction rates. The most obvious candidate here is the racial composition
of the county. The racial composition of the county is not only highly correlated with that
of the local welfare bureaucracy but may, also, in
uence sanction rates through mechanisms
operating outside the welfare bureaucracy.
28
For example, Soss et al. (2009) nd that
26
Instead, according to Norma Riccucci (2011), it is concerns about dividing the,
already very high, caseload equally among caseworkers that play a central role.
27
In theory, if panel data were available, one could take rst dierences and eliminate
any unobserved factors correlated with the regressors. In practice, the fact that
tenure among welfare workers is protected by union agreements makes it less likely
that substantial variation in the racial composition of eligibility interviewers can be
observed over a short period of time.
28
The simple correlations between the percentage of whites/blacks/Hispanics in the
county and the percentage of these racial groups among eligibility interviewers are
0.63/0.61 and 0.71 respectively.
18
sanction rates decrease as the percentage of blacks in the county increases. As they argue,
this may be the result of increased contact with non white population, of outside pressure on
local welfare policy makers or, nally, of the increased presence of blacks within the welfare
bureaucracy. By including measures of racial composition of the county, I am however able
to disentangle between these separate mechanisms. I nd that the presence of whites among
eligibility interviewers has an eect that is separate from the one that the presence of that
racial group in the population has.
Finally, additional insight on the plausibility of the sorting story is provided by looking
at the literature on welfare migration. For example, recent work (Bailey 2005; Kennan
and Walker 2010) concludes that interstate dierences in welfare benets provide at best
secondary, if any, migration incentives for welfare-eligible women.
29
If clients, however,
respond only weakly to interstate dierences in welfare benets, arguably the most visible
aspect of the welfare system, it is extremely dicult to imagine that they will make residential
decisions based on the racial composition of the local welfare bureaucracy.
30
To summarize, I argue that, conditional on the controls included, clients do not system-
atically dier across counties in ways that could aect the main ndings of this paper. The
increase in the likelihood of a sanction associated with an increase in the percentage of non
Hispanic whites among eligibility interviewers is better understood rather as the re
ection
of the negative stereotypes hold by non Hispanic whites towards welfare recipients than as
a result of unobserved variables.
29
The inclusion of state xed eects obviously accounts for any state level variables
that may aect such decisions.
30
For the results to be driven by such a mechanism, it has to be the case that harder
to serve clients decide to migrate to counties with a higher percentage of whites among
eligibility interviewers. It is not immediately obvious why they would decide to do so.
If, on the other hand, clients observe that sanction rates are higher in counties with
more whites among eligibility interviewers, then a reverse migration trend should be
observed. In that case, the results I report here constitute a lower bound of the true
eect.
19
1.6 Conclusions
In this paper, I sought to examine how the racial characteristics of the street level bureau-
crats aect service delivery and, specically, the imposition of sanctions. While sanctions
are but an aspect of the post 1996 welfare, they occupy a central role and are considered
the major disciplinary tool in the world of the new \paternalistic" welfare system. I found
that the racial characteristics of the street level welfare bureaucracy do matter for clients.
More \white" bureaucracies tend to be more punitive toward clients, irrespective of the
client's race. These results I argued re
ect whites' largely unfavorable opinions about wel-
fare recipients. I found little evidence that an increased presence of blacks among eligibility
interviewers lowers sanction rates for black clients. This latter nding provides little sup-
port in favor of active representation by the part of black welfare workers, at least in the
realm of sanctions. It is rather more consistent with the assertion by Watkins-Hayes (2009)
that the simultaneous positions of privilege and disadvantage minority welfare bureaucrats
occupy, prescribe their conduct towards minority clients within the larger institutional goals
of promoting employment and curbing welfare \dependency". More broadly, the ndings
here suggest that perceptions over the \deserving" or \undeserving" status of wide segments
of the population may shape, apart from the politics of welfare (Gilens 1999; Winter 2006),
the implementation of welfare policy itself.
20
Table 1.1: Reasons for case closure by race of the household head
Reason for Case Closure Black(sample) White(sample)
Excess Earnings,time limit etc 2938(28.8%) 2762(28.4%)
Sanctions 853(8.3%) 571(5.8%)
State/Tribal Rules 3629 (35.5%) 3546 (36.4%)
Other 2777 (27.2%) 2394 (24.6%)
#Observations 10197 9723
21
Table 1.2: Summary Statistics for individual/household level variables
Variable Black (sample) White (sample)
Age 29.044 (8.03) 29.41 (8.08)
Education (years) 10.46 (3.25) 10.55 (3.21)
Single Parent .82 .54
Male .03 .07
Citizen .97 .97
# of children 1.90 (1.23) 1.64 (1.03)
Age (youngest child) 5.00 (4.35) 5.06 (4.41)
Public Housing .13 .06
OASDI .003 .006
SSI .004 .005
Months on TANF 15.50 (15.45) 13.09 (14.39)
Earned Income (last month) 376.19 (549.07) 344.09 (527.38)
#Observations 10197 9723
Note:Standard deviations are in parentheses. Only children who are eligible for assistance
are counted towards the construction of the \# of Children" variable. \SSI" and \OASDI"
are dummy variables indicating whether the household head was a recipient of Supplemental
Security Income or one of Federal Disability Insurance Benets.
22
Table 1.3: Summary Statistics for county level variables
Variable Black (sample) White (sample)
White (non Hispanic)- eligibility interviewers .69 (0.28) .71 (0.29)
Black (non Hispanic)- eligibility interviewers .18(.23) .16 (.23)
Hispanic (white)- eligibility interviewers .04(.10) .04 (.11)
Hispanic (other race)- eligibility interviewers .04(.09) .04(.11)
Other race- eligibility interviewers .03(.08) .03 (.09)
Male -eligibility interviewers .18 (.20) .18 (.21)
Female - eligibility interviewers .81(0.20) .81(0.21)
Between 16-39 -eligibility interviewers .33(.24) .33(.26)
Between 40-49 -eligibility interviewers .35(.23) .35(.25)
Between 50-59 -eligibility interviewers .26(.22) .26(.24)
Over 60 - eligibility interviewers .05(.10) .04(.10)
High School Dropout - eligibility interviewers .01(.05) .01(.05)
High School - eligibility interviewers .15(.18) .16(.20)
Some college - eligibility interviewers .39(.26) .39(.28)
Bachelor - eligibility interviewers .36(.27) .35(.28)
Graduate Degree -eligibility interviewers .06(.12) .06(.13)
White (non Hispanic) - county .73(.17) .74(.17)
Black (non Hispanic) - county .12(.12) .11(.12)
Hispanic - county .08(.10) .09(.11)
Other Race -county .05(.06) .05(.06)
Income per capita - county 22121.9(4603.66) 21833.79(4754.686)
Population - county 471156.3(695859.1) 425841(654577.6)
% Poverty - county 10.17(3.85) 10.22(4.14)
% Unemployment - county 5.48(1.88) 5.51(2.0)
% Republican- county 48.3(12.02) 49.6(12.3)
#Counties 384 463
Note:Standard deviations are reported in parentheses. Variable \% Republican" measures the percentage voting Republican at the county
in the 2000 Presidential Elections.
23
Table 1.4: Sanctions and the characteristics of the Welfare Bureaucracy
Variable Black(1) White(2) Black(3) White(4) Black(5) White(6) Black(7) White(8)
Individual-Household level
Age -.022** -.02** -.02** -.02** -.021** -.02** -.02** -.02**
(.009) (.01) (.009) (.01) (.009) (.01) (.009) (.01)
Education (years) -.003 .02 -.004 .02 -.003 .02 -.004 .02
(.02) (.01) (.02) (.01) (.02) (.01) (.02) (.01)
Single Parent .34*** .07 .35*** .08 .34*** .08 .34*** .09
(.12) (.12) (.12) (.12) (.12) (.12) (.12) (0.12)
Male .11 -.003 .08 .02 .09 .01 .07 .01
(.21) (.2) (.21) (.19) (.21) (.19) (.21) (.2)
Citizen .18 -.34 .18 -.39 .18 -.37 .19 -.39
(.41) (.31) (.41) (.31) (.41) (.31) (.41) (.31)
# of children .11*** .05 .1*** .05 .1*** .05 .1*** .05
(.03) (.05) (.03) (.05) (.03) (.05) (.03) (.05)
Age (youngest child) .03*** .03* .03*** .03* .03*** .03* .03*** .03*
(.01) (.01) (.01) (.01) (.01) (.01) (.01) (.01)
Public Housing .1 -.13 .12 -.17 .12 -.19 .12 -.17
(.1) (.27) (.09) (.27) (.09) (.27) (.09) (.27)
OASDI -1.25 -1.3 -1.25 -1.31 -1.26 -1.32 -1.24 -1.28
(1.16) (1) (1.17) (1.02) (1.17) (1.03) (1.17) (1.01)
SSI -.49 - -.56 - -.54 - -.54 -
(1.29) (1.3) (1.3) (1.3)
Months on TANF .001 .01*** .002 .01*** .002 .01*** .002 .01***
(.004) (.004) (.004) (.004) (.004) (.004) (.004) (.004)
Earned Income -.002*** -.001*** -.002*** -.001*** -.002*** -.001*** -.002*** -.001***
(.0002) (.0002) (.0002) (.0002) (.0002) (.0002) (.0002) (.0002)
Year .17 .08 .17 .07 .17 .08 .16 .07
(.11) (.12) (.12) (.12) (.11) (.12) (.12) (.12)
24
Table 1.4: Continued
Variable Black(1) White(2) Black(3) White(4) Black(5) White(6) Black(7) White(8)
County level
White (non Hispanic) -.61 -.92 -1.36 -1.9 -1.15 -1.22 -1.46 -2.27
(1.75) (1.69) (1.57) (1.7) (1.61) (1.69) (1.65) (1.7)
Black (non Hispanic) .05 -3.31 -.02 -3.94** .15 -3.59* -.3 -4.58**
(1.74) (2) (1.46) (1.99) (1.53) (2.01) (1.5) (2.04)
Other Race 5.43 -4.91 6.19 -3.47 6.47 -3.81 5.59 -3.09
(4.36) (3.77) (4.11) (3.75) (4.15) (3.73) (4.25) (4.004)
Income per capita (log) .34 .022 .55 -.44 .55 -.41 .51 -.26
(.69) (.85) (.66) (.87) (.67) (.87) (.66) (.84)
Population (log) -.12 .07 -.14 .12 -.15 .11 -.14 .13
(.11) (.12) (.11) (.13) (.11) (.13) (.12) (.13)
% Poverty .07 -.02 .09* .03 .09* .03 .1* .05
(.05) (.06) (.05) (.08) (.05) (.08) (.05) (.07)
% Unemployment -.13 .18** -.11 .03 -.11 .03 -.13 .02
(.12) (.07) (.13) (.11) (.13) (.11) (.13) (.11)
% Republican .-002 .01 .009 .01 .01 .01 .007 .01
(.01) (.01) (.01) (.01) (.01) (.01) (.01) (.01)
White (non Hispanic) .58** .71** 1.64** 1.89*
eligibility interviewers (.24) (.32) (.73) (.97)
Black (non Hispanic) -.45 -.32 1.22 1.49
eligibility interviewers (.3) (.37) (.88) (1.09)
Hispanic (other race) 1.11 1.1
eligibility interviewers (1.06) (1.01)
Other race .31 .2
eligibility interviewers (1.2) (1.46)
Male -.13 -.5 -.13 -.37 .03 -.53
eligibility interviewers (.41) (.4) (.42) (.4) (.45) (.4)
25
Table 1.4: Continued
Variable Black(1) White(2) Black(3) White(4) Black(5) White(6) Black(7) White(8)
Between 40-49 -.06 -.22 -.03 -.12 -.009 -.19
eligibility interviewers (.29) (.24) (.31) (.24) (.32) (.25)
Between 50-59 .08 -.21 .12 -.13 .08 -.17
eligibility interviewers (.35) (.37) (.36) (.38) (.35) (.37)
Over 60 -.73* -1.08** -.70* -.79 -.7* -1.06*
eligibility interviewers (.4) (.55) (.41) (.56) (.41) (.57)
High School Dropout -1.89 -1.12 -2.58** -1.93* -1.08 -.48
eligibility interviewers (1.16) (1.08) (1.07) (.99) (1.2) (1.26)
High School -.28 .55 -.33 .41 -.28 .65
eligibility interviewers (.7) (.62) (.7) (.63) (.68) (.65)
Some college -.86 -.25 -.94 -.22 -.82 -.31
eligibility interviewers (.71) (.63) (.71) (.65) (.72) (.63)
Bachelor -.73 -.22 -.81 -.3 -.69 -.19
eligibility interviewers (.64) (.54) (.63) (.56) (.65) (.55)
State Fixed Eects Yes Yes Yes Yes Yes Yes Yes Yes
# Observations 7327 4569 7327 4569 7327 4569 7327 4569
# Counties 281 305 281 305 281 305 281 305
Note:The results were generated using STATA 11/SE. Cell entries are logit coecients with clustered standard errors reported in parentheses.
Variable \Year" is a dummy variable indicating whether the case was closed in 1999 or 2000. Variable \% Republican" measures the percentage
voting Republican at the county in the 2000 Presidential Elections. In the \white" sample the coecient on the \SSI" variable is not identiable
as it predicts the outcome perfectly. All specications include state xed eects. The inclusion of state xed eects reduces the number of
observations/counties as in a number of states no cases were closed due to sanctions during this period. In these cases observations were
automatically dropped by STATA.***p 0:01;p 0:05;p 0:1, two tailed.
26
Chapter 2
Late Champions of Healthcare Reform: Two Tier
Contracts, \Fair Share" Bills and a \Coalition to
Advance Healthcare Reform"
In the summer of 2009 and amidst a polarized political climate partly created by the
grassroots mobilization of Republicans against the proposed healthcare reform, president
Obama decided to take his party's case for healthcare reform across the country. On 29th
of July 2009, while visiting Bristol, VA, and in a rare move during his summer campaign to
promote the reform eorts, Obama held the town hall meeting at a business establishment
instead of the local high school.
1
There, at the local Kroger supermarket store Obama
accompanied by Kroger's CEO David Dillon praised among else the company for providing
generous health insurance to its employees and pledged \to work with Dave, and others
(italics mine) to see if we can control the costs so that they can keep on providing health
care at an aordable cost to all of you." (Hamstra, 2009). Obama's choice of visiting
Kroger among any other rm is not incidental. Kroger is the second biggest food retailer
1
During a period spanning three months (12th of June to 12th of August), Obama
held 10 town hall meetings in total, nine of them in various states (Wisconsin, Virginia,
Ohio, North Carolina, New Hampshire, Montana, Colorado and Minnesota). The
other took place in the AARP's headquarters in Washington, DC. Out of those only
one meeting was held at a business establishment.
27
in the U.S. (behind Wal-Mart).
2
It is unionized and like the other unionized food retailers
provides wages and benets not typical of the retail sector. More importantly, Kroger
alongside Safeway and Supervalu -the other \Big Three" food retailers- was a member of
the \Coalition to Advance Healthcare Reform" (hereafter CAHR). The coalition, which was
formed in May 2007 (Rau, 2007a), initially comprised of 37 members, included corporate
heavyweights such as Pepsico Inc, Del Monte Foods and Safeway Inc as well as a number
of health related companies such as pharmaceutical giant Merck & Co. Inc and health
insurer Health Net.
3
The \Big Three" played an integral role in the formation of CAHR
as it was from their corporate ranks that the political entrepreneur behind this coalition
emerged. Steven Burd, the founder of CAHR and a vocal supporter of healthcare reform
in the corporate world, is the CEO of Safeway Inc. Burd and CAHR, would play an
active and high prole role in supporting the 2007/08 reform eorts in California along
lines fundamentally similar to the ones embodied in the landmark Patient Protection and
Aordable Care Act of 2010 (hereafter, ACA). The \Big Three's" leading role in health
reform in California was, however, far from an isolated phenomenon.
Only a year before Gov.Schwarzenegger declared 2007 as California's year of health care
reform Maryland's legislature, overriding Gov. Ehrlich's veto, had passed the Fair Share
Health Care Act ( hereafter, FSHA). The FSHA, essentially an employer mandate, would
have required all business employing more than 10000 workers in the state to either spend
at least 8% of their payroll for employee health insurance or to pay the dierence of what
they provide to a state fund. While in principle the FSHA was a standard \pay or play"
law, its structure ensured that the mandate would aect only one rm in the state, namely
2
According to a report prepared for the Department of Agriculture in 2007, Kroger
was 2nd among grocery retailers in terms of total sales, Albertsons, which was acquired
by Supervalu in 2006, 3rd and Safeway 4th (Martinez, 2007).
3
By the time Obama signed the Patient Protection and Aordable Healthcare Act,
CAHR had reached 63 members. Fifteen of CAHR's members were involved in the
health industry either as insurance providers, pharmaceutical companies or manufac-
turers of medical equipment. Out of the remaining 48, more than 70% were involved
in the food marketing system at various stages of production, either as producers,
wholesalers or retailers of groceries.
28
Wal-Mart. Wal-Mart with its stringent anti-union policies has emerged, during the last
two decades, as the ercest competitor of unionized food retailers not only in Maryland
but across the U.S. Competitive pressures therefore faced by unionized grocers such as
Giant Food and Safeway can explain why they would be willing to break ranks with both
the rest of the big business community and umbrella/sectoral business organizations (i.e.
National Federation of Independent Business and the Maryland Retail Association) and
side with labor unions and state health activists in supporting the FSHA.
4
Their rationale
is in clear display in a letter sent to the general assembly by Barry Scher, Giant Food's
vice president of government aairs. In his letter, Scher argues \that all companies should
pay their fair share and not subsidize those companies, especially the major employers in
the state that do not provide adequate medical care coverage for their employees and thus
shift the cost towards companies such as Giant. We will be very out front to support the
Health Care for All Coalition eort" (quoted in Kazee, 2009). Legislative success did not,
however, guarantee actual implementation. On February 2006, the Retail Industry Leaders
Association led, and eventually won, a suit against the FSHA in a U.S. district court in
Maryland. According to the 4th Circuit Court of Appeals, which upheld an earlier decision,
the FSHA was violating the ERISA exemption and could thus not be implemented.
While Maryland never came to implement the FSHA, its example gathered national
attention. In 2006 only close to 30 states considered some version of \Fair Share" law
(Contreras and Lobel, 2006). In more than one case, unionized grocers openly supported
these eorts. In Missouri, labor and business formed the Mid-American Retail Food Joint
Labor-Management Committee. The committee which represented three area supermarket,
and unionized, chains, among them Shop n' Save owned by "Big Three" retailer Supervalu,
and the three union locals introduced legislation which according to its director Michael
Kelly \is very similar to the legislation in Maryland" (Hamstra et al. 2006). In Washington,
Safeway supported a similar bill based on arguments largely re
ecting the rationale behind
4
Safeway and Giant Food were the exceptions, among big business, in supporting
Maryland's \Fair Share" legislation. The major business organizations were openly
hostile to the FSHA.
29
the original bill in Maryland (Ralph, 2006). The proliferation of \Fair Share" legislation
across the country, occasionally with the support of the \Big Three", shows that by 2006
these rms had started to view employer mandates, the most controversial aspect of health
reform for business, under a very favorable light. While it was through their support for \Fair
Share" legislation that the \Big Three" initially signaled their openness to health reform,
it would be in California that this support solidied into something more comprehensive.
As with the landmark federal legislation of 2010, the plan developed in California was
building around individual and employer mandates as well as an expansion of public pro-
grams for individuals below a certain income threshold. It was widely viewed by policy
experts and scholars both as a template for the national eorts as well as \quite simi-
lar" to the ACA, after the latter's passage (Belsh e, 2011, Harbage et al. 2008, Hacker
and Rodgers, 2009). Moreover, even though the California reform eorts eventually failed,
the state's experience with reform was often considered by scholars as more relevant for
the national eorts even when compared to the successful Massachusetts experience. As
Hacker and Rodgers (2009) point out: \Although the state of Massachusetts successfully
enacted a plan for universal coverage in 2006, Massachusetts has many characteristics that
set it apart from other states, including a high percentage of employers oering insurance,
a relatively small uninsured population, an existing uncompensated care pool, a preexisting
set of insurance market regulations that minimized the dislocations of new rules, and a Med-
icaid waiver that provided a ready federal funding stream. By contrast, California-with its
diverse population, strained nances, lack of robust insurance regulations, and large num-
bers of uninsured-provides a better measure of the challenges that national policymakers
will face."
It is likely that the state's less favorable conditions, when compared to Massachusetts's,
and especially it's strained nances put additional pressure on lawmakers in California to
30
design a plan that would include, not a de minimus a la Massachusetts employer contribu-
tion, but a rather robust employer requirement.
5
In any case, in their attempts, reformers
found unexpected support in the face of Safeway and CAHR. Safeway's CEO Burd played a
central role in the formation of CAHR and, as it will be analyzed in detail later, pushed the
Schwarzenegger administration for a higher employer contribution (Kazee, 2009). However,
despite, the unexpected and vocal business support, the reform eorts eventually foundered,
most probably due to the severe budget crisis the state was facing at the time (Hacker and
Jacobs, 2009). As the reform eorts in California were crashing in the shores of the state's
budget crisis, health reform was on its way to occupy, yet again, a central place in national
politics, a place which the election of Obama in November 2008 only helped in solidifying.
Thus, by the time Obama visited Kroger's store in Virginia, the \Big Three" food retailers
had spearheaded some of the most ambitious, state level, reform eorts since the demise of
the Clinton's plan in 1994.
6
Their support for health reform is, however, both puzzling and
consistent if put in a historical context.
On the one hand, it is consistent with business involvement in health reform during
the Clinton administration. As it has been analyzed in Swenson and Greer (2002), for a
time during the early 1990s big business in the form of the \National Leadership Coalition
for Health Care Reform" (hereafter, NLCHCR) openly supported the reform eorts of the
5
In Massachusetts, each rm could pay a
at fee of 295$ per employee (per year)
in lieu of direct health spending on its employees. The employer requirement in
California was set as a percentage of the rm's Social Security payroll and varied
depending on the size of the payroll (the requirement was set to 1% for payrolls below
250000$ and peaked at 6.5% for rms with a payroll larger than 15.000.000$).
6
By ambitious, I mean here reform eorts that were including employer mandates.
The employer mandate is the most controversial issue among business and has tra-
ditionally been an anathema for the major umbrella business organizations. Apply-
ing this criterion, the most important eorts for comprehensive reform were clearly
the Massachusetts (2006) and California (2007) ones. Other states such as Vermont
(2006), New Mexico (2007), Pennsylvania (2007) and Illinois (2007) considered, and
in the case of Vermont passed, comprehensive reforms, including employer mandates.
However, with the exception of Vermont, these eorts did not go far in the legisla-
tive process. In any case, none of these attempts gathered the attention that Mas-
sachusetts and California did. Also, while, Maryland's \Fair Share" legislation was
not aiming at comprehensive reform it gathered nationwide attention as is evidenced
by the introduction of similar legislation in a large number of states.
31
Clinton administration. A few of the unionized grocers, including Safeway, Ralphs (acquired
by Kroger in 1998) and Vons (a Safeway subsidiary since 1997), who joined CAHR in 2007,
were also members of the NLCHCR back in the 1990s. Steven Burd, in all probability echo-
ing the sentiment of other unionized grocers, made a forceful argument for the imposition of
an employer mandate back in 1994, after a joint press conference organized by the Clinton
administration and congressional Democrats. He argued specically that Safeway competes
\with some very large companies that don't oer the same kind of coverage" and, thus, if
health reform doesn't pass with the employer mandate, Safeway might be forced to curtail
its coverage \to level the playing eld" (Rosenblatt and Tumulty,1994). In the end, as it is
well known, even the late show of some business support could not save the Clinton health
plan. With the demise of the Clinton plan, the center of the reform eorts once again moved
to the states.
7
As the events showed, for unionized rms, such as the \Big Three", state
level reform eorts were an opportunity they would often welcome.
On the other hand, the \Big Three's" interest in imposing a meaningful
oor on employer
contributions towards employee health insurance stands in stark contrast with other devel-
opments. During a period in the early 2000s and in their attempt to radically lower labor
costs, the grocers adopted a very confrontational agenda towards the Union of Food and
Commercial Workers (hereafter, UFCW), the main union representing workers in the sec-
tor. Inevitably, this led to intense labor strife, culminating in the largest retail workers'
strike in the U.S. history, the grocery strike of 2003-04 in Southern California.
8
Southern
California alone represents a huge chunk of the chains total sales. In 2003, the year the
strike occurred, it accounted for one-fth of the chains' total sales (Cleeland 2003) and,
7
The Clinton reform eorts were preceded by a wave of activity at the state level
(McDonough et al. 2008).
8
The Encyclopedia of strikes in American History, Aaron Brenner, Benjamin Day,
Immanuel Ness, editors. Armonk, N.Y: M.E. Sharpe, 2009. The strike, at its peak,
involved 67000 workers and lasted for almost 5 months. According to the Bureau of
Labor Statistics the 2003/04 strike was the second in terms of days idle in the U.S
since 1992, when detailed data are available.
32
obviously, any reduction in labor costs there would have been a huge boost to their prof-
itability. However, and this is the crucial point, their goals went far beyond California.
9
Speaking in an investors' conference in 2002, Burd was explicit with regards to Safeway's
long run plans. The goal was, and as the events show it was a goal to which Safeway found
willing allies in the face of the other major unionized grocers, to impose two-tier contracts
across the country by 2007 (Zwiebach, 2002). Two-tier contracts, by sharply increasing
waiting periods for health insurance, were perceived by the \Big Three" as a way to gradu-
ally reduce coverage and eventually bring wages and benets down to the level \enjoyed" by
their non unionized competitors (Ghitelman et al, 2003), the practices of which ironically
they so openly criticized in Maryland and elsewhere.
Table 2.1 which summarizes the main events recounted in this section highlights the
fact that by the time healthcare reform had once again moved to the spotlight of national
politics the \Big Three", who bargain collectively with the UFCW, had been involved in
both the largest industrial con
ict over employee health benets in U.S. labor history and
had been in the forefront of some of the most ambitious government attempts to expand
health insurance coverage since Clinton's failed health plan.
10
These events which form so
strange an alliance constitute the puzzle that lies at the heart of this article. Making sense
of this puzzle is the goal of the analysis that follows.
9
According to Jordan (2004), \The chains wanted to use Southern California, their
largest market, as a national example. What labor cost reductions they could win there
would make it that much easier to win everywhere (italics, mine)."
10
The 2003-04 grocery strike was the largest, in terms of days idle, work stoppage
over employee health benets. Since the mid 1980s, when employee health benets
started to become a highly contentious issue in contract negotiations, the only compa-
rable strike in terms of days idle was the strike between the Communication Workers
of America and Nynex Corp. in 1989. This con
ict was also primarily revolving
around health benets but it involved slightly fewer workers (i.e. around 60000) and
was shorter in duration.
33
2.1 Theoretical Background and the Argument
In the literature on employers and social insurance programs, it is now widely recognized
that business preferences may not be as monolithically opposed to the welfare state as it
was once believed. Business preferences, it is argued, are
uid and are shaped both by
rm/sectoral characteristics as well as by the overall macroeconomic conditions. Employers
may come to embrace, for example, the expansion of social insurance programs if they
increase workers' incentives to invest in rm specic skills (Mares, 2003). In addition,
especially in periods of depressed economic activity and intensifying price competition, rms
which already provide relatively generous benets/wages may welcome the expansion of
programs that force all rms to provide such benets. As Swenson (2002) characteristically
writes \Imposing cost increases on competitors . . . seems a less messy alternative to imposing
cost reductions on workers". A similar rationale has been invoked by students of U.S. health
policy in order to explain variations in business attitudes towards the expansion of health
insurance. It is argued, for example, that large unionized rms, who often provide relatively
generous health benets, may view employer mandates as a way to \level the playing eld"
with competitors providing low or no benets (Gordon, 2003:251-55, Gottschalk, 2000:110-
11, Hacker, 2002:264, Martin, 1995, Swenson and Greer, 2002). For the \Big Three", who
collectively bargain with the UFCW in a system closely resembling what Swenson (2002)
calls negotiated cartelism, employer mandates can thus be seen as an alternative to hardball
bargaining with the unions, a way to \level the playing eld upwards" and thus compete
more eectively with non unionized rms who have been steadily increasing their share
during the last two decades.
11
A fascinating example of how management may come to
perceive legislative (i.e. employer mandates) solutions as an alternative to private (i.e. cost
11
Typically, negotiated cartelism prevails in sectors which are labor intensive, use
simple and inexpensive machinery and produce easily transportable goods. These
factors make entry easy and may prompt employers to collectively bargain with unions
in setting a wage
oor and thus discourage new entrants who may not be able to
aord paying union wages. In the U.S., negotiated cartelism was historically more
prevalent in coal mining, building, the garment industries, retailing and transportation
(Swenson, 2002).
34
shifting to employees) ones is Burd's 1994 threat that Safeway would have to curtail its
coverage if the Clinton Health plan did not include an employer mandate. In anticipation
of the argument, it is important to emphasize the following.
First, that supporting legislative solutions does not imply that management cannot con-
currently pursue cost containment privately. In the early 1990s, big business both experi-
mented with managed care and, at least for some time, supported the reform eorts of the
Clinton administration. At the same time, given that extensive cost shifting could cause
serious industrial con
ict with organized labor, one would expect that, at least, the most
radical plans for cost shifting, such as two-tier contracts, would not be pursued alongside
legislative solutions. Pragmatic managers possibly understand that while some relatively
mild forms of cost shifting to employees would not endanger a cross-class coalition with
labor, a more radical agenda with respect to industrial relations would make the prospect
of a business-labor coalition remote. Second, the choice between legislative and private
solutions, at any given time, would depend on how the relevant actors perceive the rela-
tive costs/benets of each alternative. An illuminating example of this calculation can be
found in the analysis of the evolution of big business support during the Clinton reform
eorts. Specically, Swenson and Greer (2002) attribute the gradual cooling of business
support for the Clinton health plan to the spreading perception, among large employers,
that private solutions (i.e., managed care) were already yielding results by reducing health
premium growth rates. As a result, they argue, business eventually came to perceive gov-
ernment regulation as \super
uous if not downright dangerous". In other words, support
for a legislative solution subsided when it was perceived as inferior to private solutions in a
cost/benet sense. The argument developed in this article is based on these two premises.
I will argue specically that the \Big Three's" interest in healthcare reform waxed and
waned during the rst decade of the 21st century as a result of changes in the industrial rela-
tions between the major unionized grocers and the UFCW. In the early part of the decade,
the \Big Three" showed very little interest in government intervention and attempted to
vastly reduce, and in the long run eliminate, the kind of healthcare benets Obama praised
35
in his meeting at Kroger. Their attempt though led to labor strife and, eventually, to the
great grocery strike in Southern California in 2003/04. Although, eventually, the grocers
in Southern California imposed terms that greatly reduced their labor costs, their victory
came at a high cost and it is exactly this high cost that made their victory in Southern
California a temporary one and, eventually, very dicult to replicate across the country. It
is in this context that their interest in healthcare reform developed in the later part of the
decade.
The analysis is based on the comparative study of the two latest incidents in the long
story that is healthcare reform in California. I specically compare the stance that the
"Big Three" adopted during the reform eorts taking place in 2003 with their stance in
2007, California's latest \Year of Healthcare Reform". The fact that in both instances the
choice between private (i.e. via contract negotiations with the UFCW) and government
solutions was explicitly posed to the \Big Three" renders the comparison methodologically
appropriate. Moreover, while the \Big Three's" support for reform far exceeded the borders
of the state (i.e \Fair Share" legislation), the events in California not only played a decisive
role in shaping their stance towards reform but, also, include some of the most signicant
developments in health policy in the aftermath of the Clinton's plan debacle.
Specically, the 2003-04 con
ict between the UFCW and the \Big Three" in Southern
California was the largest industrial con
ict over employee health benets in the U.S. his-
tory, a con
ict which according to labor experts greatly in
uenced subsequent negotiations
between the UFCW and the \Big Three" across the country. In addition, both the 2003-04
and 2007-08 reform eorts, which feature in the analysis, were signicant events if put into
the national context. As discussed in the previous section, the plan for reform developed
in 2007-08 was widely seen by scholars and policy experts as a template for the national
legislation (Harbage et al. 2008, Hacker and Rodgers, 2009) and gathered the attention
of major stakeholder groups. As Karen Ignagni, president of America's Health Insurance
Plans and a leader in the industry's talks with the White House, acknowledged, as a result
of the California reform eorts \a number of dierent groups began to talk about what are
36
the lessons of California, and how big a problem the rising health cost curve is" (quoted
in Sieb, 2009). While it drew less nationwide attention at the time, the Health Insurance
Act of 2003 was actually the rst \pay or play" legislation enacted in the period after the
failure of Clinton's health plan. It is the politics of this early "pay or play" law which are
discussed in the next section.
2.2 On The Sidelines: Healthcare Reform in 2003
The Health Insurance Act of 2003 (bill SB 2) was an attempt to expand healthcare
coverage among working Californians. The law was making provisions for an employer
mandate that would force all rms (with more than 20 employees) operating in California to
pay a fee so as to fund the State Health Purchasing Program, a purchasing pool operated by
the Managed Risk Medical Insurance Board. The program was to provide health insurance
to those employees who were not receiving health coverage through their employers. Firms
providing coverage to their workers, meeting the minimum requirements set by the law, were
to receive a credit towards the fee. The law was mandating that all employers should pay
at least 80% of the assessed fee with the remaining being picked up by the employee. The
law was to take eect gradually starting with the larger employers (over 200 employees) in
2006 and with medium size rms (between 20 and 199 employees) in 2007, provided that the
appropriate tax credit was in place. Small employers (those with fewer than 20 employees)
were excluded from the mandate.
The legislation was vehemently opposed by business groups, such as the California Cham-
ber of Commerce (which included the bill in its 2003 list of "job killer" bills), on the ground
that it placed unreasonable demands on business and would thus drive jobs out of the state.
While for business the Health Insurance Act was an anathema, labor unions provided strong
support and made it their top legislative priority for 2003 (Jones, 2003). UFCW supported
the bill on the grounds that it would level the playing eld and, therefore, remove the \threat
from Wal-Mart" as a valid argument from the negotiations with the grocers (Cleeland, and
Dickerson, 2003). The bill was nally signed into law in the waning days of Gov. Davis
37
term (October 5th of 2003), just two days before his recall election. The law, however, was
not to take eect. Immediately after its passage, the California Chamber of Commerce led
for a referendum that would decide whether the bill was to take eect or not. The Health
Insurance Act was put into ballot in November 2004 under the title proposition 72 and
was defeated by a very slim margin.
12
Spearheading the campaign against proposition 72
was the \Californians Against Government Run Healthcare", a coalition of business groups
such as the California Restaurant Association, the California Chamber of Commerce and
corporations such as Burger King Inc and Wal-Mart (the major low cost competitor of the
\Big Three"). In fact, Wal-Mart was the second biggest contributor of the repeal campaign
(California Secretary of State website). While Wal-Mart's support for the repeal campaign
is not surprising in light of the notoriously low benets it oers to its employees, the indif-
ference that the \Big Three" showed towards legislation that would have leveled the playing
eld appears intriguing. For one, their cool reaction towards proposition 72 did not stem
from a failure to grasp its regulatory merits. Evidence for this is that four years later, May
2007, and a few days after the announcement of the formation of CAHR, Burd responding
to criticism that his support for the employer mandate originated from a wish to increase
the costs of its low cost rivals argued that \if Safeway was trying to level the playing eld,
it would have supported proposition 72, the 2004 ballot measure that would have required
businesses with 50 or more employees to provide health insurance or pay a fee" (Ainsworth,
2007). Burd's statement is a clear indication of his recognition of the regulatory merits of
the employer mandate for Safeway in 2003. As far as his claim that Safeway did not support
proposition 72 because they were not interested in leveling the playing eld, I will show that
by the time the campaign for proposition 72 was at its height, the grocers had largely solved
the \labor cost" issue by leveling the playing eld downwards. The event linking the cool
reception for healthcare reform in 2003 and the much more receptive stance of the "Big
Three" in 2007 is the great grocery strike that hit Southern California in 2003, the largest
retail workers' strike in the American history.
12
50.9% voted against proposition 72 while 49.1% voted in favor.
38
2.3 The Largest Retail Workers' Strike in the American History
By 2003, the rapid increase in health insurance premiums that had started in the late
1990s reached its peak. California was no exception.
13
Among those employers who were
oering insurance, the problem was more acute for those providing the most generous health
benets. The unionized grocery chains, which were providing benets that according to one
expert were \particularly unusual for the retail industry whose wages and benets tend to
lag behind those of other industries" (Cleeland and Dickerson, 2003), were probably amongst
the hardest hit by the health insurance crisis, at least among large employers. Aggravating
the situation for the unionized grocery sector in California, was the impending expansion
of Wal-Mart supercenters into the state (Cleeland and Vrana, 2005). It is, therefore, not
coincidental that stock analysts were urging the big grocers to act decisively in order to
remain competitive with their non-unionized competitors and especially Wal-Mart (Peltz,
2003). It is in this context that Kroger, Safeway and Albertsons decided to act aggressively
in their negotiations with the UFCW in 2003. That the negotiations with the UFCW
would be particularly dicult was evident as early as 2002, a year before the contract in
Southern California expired when, in a meeting of the four top markets at the union's
headquarters in Washington, Steven Burd of Safeway told the union leaders that their
mutually benecial relationship based on strong prots and generous wages and benets for
store workers was \being jeopardized by low-cost, nonunion competitors such as Wal-Mart
Stores Inc." (Cleeland and Fulmer, 2003). Central in the markets' proposals was the creation
of a two-tier system, where new employees (i.e. those belonging to the second tier) were not
only to receive lower wages and less generous healthcare/pension benets but, also, to face a
much longer waiting period until they could become eligible for health benets (the waiting
period was to increase from a few months to more than a year). After several months of
unfruitful negotiations, the strike began with UFCW striking Vons (owned by Safeway) on
13
Health insurance premiums rose by 15.8% in 2003 in California. Since 2003, pre-
miums are increasing at signicantly lower rates (California Employer Health Benets
Survey, 2008).
39
October 11th 2003. The next day, Albertsons (acquired by Supervalu in 2006) and Ralphs
(owned by Kroger) locked out their union workers in support. The strike which would last
for almost 5 months was the rst in the grocery business in California since 1979 and the
longest supermarket strike in the U.S. history, a strike that ended with the ratication of a
contract that, by contemporary accounts, was (for grocery workers) \certainly inferior, to
their previous contract and, in some respects, inferior to the contract proposal they rejected
just before the strike" (Hiltzik, 2004). Although veteran workers (those hired before the
strike) were to continue receiving generous healthcare benets (without any contribution
to premiums until 2006 and a very low contribution thereafter), the contract was making
provisions for a wage freeze for the rst two years. Crucially, under the new contract a
permanent second tier of workers receiving lower wages and benets was created. These
workers were not only to receive lower wages compared to rst tier workers (with the wage
gap depending on job classication) but would, also, qualify for health insurance only after
a much longer period of time, which ranged from 12 months (for individuals) to 30 months
(for families), a dramatic increase from the 4 months of the pre-strike contract (Dube and
Lantsberg, 2004). Despite their clear win in the labor con
ict, the supermarket chains did
not get out of the con
ict unscathed. A discussion, however, on the costs of the 2003 strike
and their impact on future negotiations between the UFCW and the \Big Three", has to be
postponed for a later section of this paper. First, it is important to examine the reasons for
the largely indierent reaction towards proposition 72 that the \Big Three" exhibited during
the repeal campaign, a few months after the new contract had been ratied (February 29th
2004). A chronology of the relevant events for 2003-04 is presented in Table 2.2.
2.4 The "Big Three's" Stance with Respect to Proposition 72
I have already provided (see Steven Burd's statement above) some evidence that the
supermarket chains were aware of the regulatory role of the Health Insurance Act in 2003.
I will, also, argue here that Burd's claim that the chains were not interested per se in the
40
regulatory benets of the regulation should not be taken at face value. Even a cursory com-
parison between the provisions of the Health Insurance Act and of the post-strike contract
makes it clear that the reason for the non-support of proposition 72 was that the Health
Insurance Act would have annulled the most important gains accruing to the chains as a
result of the new contract. To see this, we need only examine the provisions with respect
to the waiting periods for eligibility for health benets. Under the old contract, employees
were becoming eligible after 4 months, whereas under the new one the time was set to 12
months (for individuals) and to 30 months (for families). The increase in the waiting time
combined with the requirement that new employees had to contribute 20% towards their
premium payments (veterans still did not have to pay premiums) meant that very few work-
ers were to become eligible for health insurance. In fact, according to a recent study (Dube
and Jacobs, 2007), the longer waiting periods account for most of the reduction in coverage
among grocery workers (coverage dropped from 94% to 54%) in Southern California between
2003 and 2006. Given that under the Health Insurance Act workers would become eligible
after only 3 months, it is easy to see why the chains had little incentive to support the legis-
lation after they had signed the new contract. After the strike, the legislation would have
cancelled some of their gains from the new contract even though it would have imposed costs
to their \free riding" competitors. Their post-strike stance can thus be understood if one
compares the benets of the new contract for the employers with the demands the Health
Insurance Act was placing on them. As far as their largely indierent (towards health-
care reform) pre-strike stance is concerned, the following should be noted. Given that the
\Big Three" were oering very generous healthcare benets (with no premium contributions
from the part of their workers) and short waiting periods (4 months), it is clear that the
Health Insurance Act represented an improvement compared to the status quo for the \Big
Three".
14
It was an improvement both because rst, by forcing all employers to provide
14
It is important to note that according to a report prepared by the UC Berkeley
labor center (Dube and Jacobs, 2007) in September 2003, just before the strike started,
employer-sponsored coverage among grocery union workers was close to universal
across California (94% in Southern California and 96% in Northern California). It is
thus highly unlikely that the legislation would have imposed any extra costs on the
41
coverage, it could have reduced the costs of providing insurance to uninsured dependents of
their employees for the \Big Three" and, second, because it would have increased the labor
costs of their main competitors. It is of course not coincidental that among the biggest
contributors of the repeal campaign was Wal-Mart, whose low wages and benets were used
by the chains as one of their prime arguments when pushing the UFCW for concessions.
Thus, the chains' pre-strike stance cannot be attributed to the fact that the legislation was
detrimental to their interests as it possibly was after the new contract had been ratied.
The above suggest that the grocers' pre-strike stance with respect to healthcare reform in
2003 cannot be understood in a static fashion. I will argue that interpreting their stance in
2003 requires a comparison with the much more receptive stance they held during the 2007-
08 campaign for healthcare reform in California. I have already hinted on the role that the
costs of the 2003 strike for business may have played in the post 2003 negotiations between
the grocers and the UFCW. The next section elaborates more on this issue by examining
the circumstances under which the interest in healthcare reform and, then, the support for
an employer mandate developed in 2007-08. Determining why the grocers may have been
more interested in government intervention in 2007-08 goes a long way in explaining their
cool reaction in 2003-04.
2.5 Into the Ring: Healthcare Reform in 2007
The beginning of 2007 found the \Big Three" and the UFCW in Southern California
preparing for negotiations over a new contract which was about to expire in March 2007.
For the union, the main aim in the negotiations was the elimination of the two-tier system
(Hirsch, 2007a) established in their last contract. The \Big Three", for their part, who
had reaped signicant nancial gains from the implementation of the two-tier system (see
discussion above) had no apparent reasons to accept labor's demands (Hirsch, 2007a). In
2007, as in 2003, healthcare reform was at the center of attention in the legislature. Gov.
chains in the form of an increased number of workers deemed eligible for insurance
under SB 2.
42
Schwarzenegger, who had only recently dubbed 2007 as the \Year of Healthcare Reform",
was pushing for a major overhaul through mandates (individual and employer) and an
expansion of eligibility to Medi-Cal. The Democrats had their own proposal (bill AB 8)
which diered crucially from the Governor's plan in that it did not include an individual
mandate and was requiring a larger employer contribution.
15
Ultimately, bill AB 1x1 which
was supported by CAHR would come as a compromise of these earlier proposals.
On the one hand, Democrats' interest in expanding coverage ts well with a long-run
pattern of Democratic interest in healthcare reform in California.
16
On the other hand, the
interest the Republican governor showed in expanding coverage through mandates appears
intriguing for two reasons. First of all, mandates tend to be unpopular among Republicans.
Bill AB 1x1 passed through the assembly in December 2007 without a single Republican
vote. Second, the employer mandate was resented by some of the key business groups
that had been staunch supporters of Schwarzenegger during his tenure, namely the Cal-
ifornia Chamber of Commerce and the California Restaurant Association. Back in 2003,
Schwarzenegger had sided with these groups in their eort to overturn the employer man-
date of and during his tenure he has retained an amicable relationship with both.
17
It is
possible that in getting engaged in the healthcare reform eorts Schwarzenegger, who after
his reelection in 2006 had sided with the Democrats in supporting a few high prole and
15
In bill AB 8, introduced by Assembly Speaker Nuez (D-Los Angeles), the mini-
mum employer contribution was set at 7.5% percent of Social Security base wages.
The employer contribution in Schwarzenegger's plan ranged from 0-4% depending on
payroll size.
16
At least 19 major bills have been introduced in the California legislature with the
aim of expanding health coverage during the last two decades (Dimmitt, 2007). All but
two were sponsored by Democrats. The exceptions are the Brown-Maddy \Aordable
Basic Health Care Act" (1992) which was cosponsored by Assembly Speaker Brown
(Democrat) and Minority Leader Maddy (Republican) and AB 2450 (2006) spon-
sored by Richman (R-Los Angeles). The later, however, was essentially an individual
mandate with no requirements imposed on employers.
17
During his term, Schwarzenegger has vetoed over 90% of the bills that were desig-
nated as \job killers" by the California Chamber of Commerce (California Chamber
of Commerce website).
43
unpopular to the business community bills, such as the \California Global Warming Solu-
tions Act" of 2006, was simply seeking to bolster his political legacy. Although plausible as
an explanation, especially given that he was to be termed out in 2011 and he may, therefore,
had fewer incentives to woo any particular group, I believe that the above explanation tells
only part of the story. A second explanation that, in my opinion, complements the rst
has Schwarzenegger moving in anticipation of some support among business and especially
Safeway and Steven Burd. Safeway and Kroger are signicant nancial contributors to the
governor and in his plan Schwarzenegger included provisions directly inspired from the well-
ness programs that Safeway introduced for non-union employees in 2006 (Rau, 2007c).
18
Thus, it is all but certain that Schwarzenegger was esoteric to the industry's situation, and
the grocers' concerns over escalating health costs. Which of the two reasons mattered most
in the end can only be a matter of speculation.
While the legislature appeared sharply divided over health reform across ideological
lines, both organized labor and business adopted a more
exible stance towards reform. A
signicant part of organized labor, in the form of the Service Employees International Union
(SEIU), the American Federation of State County and Municipal Employees (AFSCME) and
many of the building trades unions supported the bill (Jacobs and Hacker, 2009). However,
the AFL-CIO, although initially supportive of the reform eorts, later withdrew its support
citing concerns over the aordability of the individual mandate. The UFCW which had
welcomed the formation of CAHR as \the rst serious entry of the business community
as full participants into the national healthcare reform debate" and had applauded the
coalition's principles (statement by Joe Hansen, International President, UFCW on May
7th 2007) did not support the nal version of the bill on the grounds that it was not setting
a high enough employer requirement and could thus encourage crowding out. According
to George Landers, director of the United Food and Commercial Workers Western State
Council, the measure would have given the chance to large employers such as \Wal-Mart,
18
According to www.arnoldwatch.org as of June 2007, Kroger (through its California
subsidiary Food 4 Less) and Safeway were amongst the top 100 nancial contributors
to Schwarzenegger.
44
who will not have to pay any more than they presently do" to \allow thousands of their
employees to receive state subsidies for their health care."
19
On the business side, and on the one hand, the major umbrella organizations such as
the California Chamber of Commerce and the California Restaurant Association opposed
the bill. On the other hand, the reform eorts gathered vocal support from a segment of the
business community, most prominently from CAHR and specically Steven Burd who played
a crucial role in bringing the business coalition together (Kazee, 2009). In spite, however, of
open business support, particularly from Safeway, the reform eorts eventually foundered
most likely due to the severe budget crisis the state was facing (Hacker and Jacobs, 2009).
While the governor and the legislature were occupied in drafting a compromise proposal,
the chains were engaging in negotiations for a new contract with the UFCW in Southern
California. After lengthy negotiations, including two strike authorizations, a new contract
was ratied by UFCW members in July 2007. The new contract included slashing waiting
times for health insurance for the second tier workers (down to 6 months from as long as
18 months) and was described as \a win for the UFCW" by labor experts (Hirsch, 2007b).
Table 2.3 presents a chronology of the main events that will feature in the analysis of this
section.
An examination of Table 2.2 and 2.3 brings into attention certain similarities between
the two periods. First, in both cases, negotiations between the UFCW and the \Big Three"
in Southern California took place while healthcare reform was at the center of attention in
the legislature. In 2003-04, the Health Insurance Act was promoted by a strong democratic
majority in both the assembly and the senate with the support of labor unions. In 2007,
both the Republican governor and the Democrats, who were controlling the legislature, were
working on proposals aiming to introduce signicant reforms. Second, while discussions over
healthcare reform started before the initiation of the negotiation between the UFCW and
the supermarket chains, the bills took their nal form after the contracts in Southern
California were signed. Thus, in both cases, the chains enjoyed a window of opportunity
19
Press Conference organized by California Nurses Association, (January 22nd 2008)
45
to support or oppose reform in anticipation of the talks with the UFCW. After the new
contract was ratied by the union, they had in both cases the chance to take a position
over the nal form of the legislation. Crucially, the legislation under discussion had in
both cases obvious benets for the supermarket chains. I have already discussed the merits
that the Health Insurance Act of 2003 would have for the chains. The various proposals
circulating in early 2007 were also building on an employer mandate, although this time the
contribution was set as a percentage of the total payroll (4% according to Schwarzenegger's
plan and 7.5% according to AB 8). In light of the fact that non-unionized competition was
remaining a major threat for the chains, especially as another non-unionized competitor
with vast resources (the British owned Tesco. Inc) was ready to tap into the rich California
market (Chung, 2006), and that larger rms (those with more than 100 employees) that
insure their workers spend on average 10% of their payroll for employee health insurance
(California Health Care Foundation, 2007), it is not surprising that Safeway through Steven
Burd characterized the 4% requirement of the Schwarzenegger plan as \frankly too low and
should be higher." (quoted in Rau, 2007b). Obviously, within certain limits, the more
onerous the employer requirement would be, the larger the cost impact would be on those
"free riding" competitors of the unionized chains. This can explain why Burd and the
grocers may have liked the idea of a higher contribution than the one embodied in the
governor's plan. At the same time that the chains were taking a more supportive stance
towards reform, they were adopting a less combative stance compared to the one in 2003
towards the UFCW. Evidence for a lower degree of preparation for a con
ict is that in
contrast to their practice in 2003, when they hired \thousands of replacement workers well
ahead of the strike deadline" (Milkman, 2005), in 2007 they did not take similar measures
(Hirsch, 2007a). There are two explanations for this change in strategy that lead to radically
dierent conclusions. According to the rst, the lower degree of preparation may stem from
the belief of the chains that UFCW will not strike given the very high costs it incurred
46
in 2003.
20
Essentially, this explanation attributes the lower degree of mobilization to the
weakness of the UFCW. However, this explanation does not bear well with the evidence.
In the aftermath of the 2003 strike and because of the unexpectedly high losses in terms
of lost sales and market share, the chains appeared much less eager to mobilize to the
extent they had done in Southern California prior to the labor con
ict.
21
According to
one labor expert, U.C. Berkeley's Prof. Harley Shaiken, as a result of the costs associated
with the strike, the chains became very reluctant to provoke similar strikes in other places
(Fulmer, 2003). In the same vein, Kent Wong, director of the UCLA Center for Labor
Research and Education, argues that \Management across the country has looked at what
happened in Southern California, and decided that it is not something they want to repeat
(italics mine)." (quoted in Washburn, 2005). As a result of the chains' lower willingness
to engage in such protracted and very costly con
ict, the UFCW managed even as early as
2004, a few months after the Southern California strike, to resist similar concessions in its
negotiations with the \Big Three" in Central and Northern California (Jacobs, 2009). That
the chains were less capable mobilizing against the union does not mean, of course, that
they were willing to accept the union's demands as far as the elimination of the two-tier
system was concerned. It took the UFCW a show of great resolve, the union gained two
strike authorizations, to reverse most of the concessions it had to make in 2003. According
to Ken Jacobs, of the U.C. Berkeley's labor center, in his discussions with UFCW leaders
and sta who participated in the negotiations with the chains, various reasons were cited as
explanations for the outcome. Those ranged from better coordination between UFCW locals
that increased the union's leverage in the negotiations, less pressure from Wal-Mart as a
result of its lower than expected penetration in the California market, high turnover among
grocery employees that was creating problems for store managers, the fact that Steven Burd
had assumed a strong public role in favor of healthcare reform and did not want the bad
20
Workers lost an estimated $800 million in wages due to the 2003-04 strike and
lockout (Hirsh, 2007)
21
According to some estimates, the "2003-04 strike and lockout cost grocers around
$2 billion in lost sales" (Hirsh, 2007).
47
publicity to the fact that by 2007 the chains had already negotiated less stringent contracts
elsewhere and, thus, it was more dicult to defend the provisions of the 2003 contract they
had signed with the union (email correspondence with Jacobs, April 2010). Irrespective of
which factor mattered most for UFCW's victory in 2007 in California, what appears evident
according to labor expert Harley Shaiken is that by 2007 unionized grocers had been unable
to repeat their initial success in Southern California and instead of implementing two-tier
contracts across the country, they had adopted a less combative stance towards the UFCW in
order to avoid a repeat of the 2003-04 work stoppage (Darc e and Davies, 2007).
22
Therefore,
in light of their failure to radically restructure their labor contracts along the two-tier lines
and thus gradually eliminate health coverage, as they did for a time in Southern California,
the chains were much more receptive towards healthcare reform, reform that could, at least
partially, solve the \labor cost" issue without risking a potentially very costly con
ict with
the UFCW. Both their support for comprehensive reform in California and for "Fair Share"
legislation in a number of states can be explained by the eventual failure of the \industrial
con
ict" strategy.
23
2.6 Alternative Interpretations
The argument I have put forward here attributes a large role to the grocers' involvement
in the healthcare reform eorts in California and elsewhere to changes in industrial relations
between the "Big Three" and the UFCW. However, there are at least three alternative
explanations that, if valid, would compromise the explanatory power of the argument I
22
As early as 2005, industry experts such as George Whalin, president of Retail
Management Consultants, were pointing out that while the \Big Three" had managed
to get most of what they wanted in the Southern California negotiations, they had
largely \failed in their larger strategy, which was to implement a two-tiered system
across the country." (quoted in Washburn, 2005).
23
It is interesting to note that Giant Food (owned by Ahold), who provided critical
support for Maryland's "Fair share" law in 2006, had not supported a similar law
a few years earlier (2003-04). The change in Giant Food's stance between 2003 and
2006, in Maryland, is consistent with the argument I advanced here, that is the failure
of the \industrial con
ict" strategy.
48
advanced here. It is, therefore, important at this point to discuss the relevancy of these
arguments for the case at hand.
According to the rst explanation, support for the employer mandate in 2007 may have
re
ected the choice of the \lesser evil" instead of true preferences. This would have been
the case if rms got involved in the debate not with the hope of scoring net gains from the
reforms but, rather, to stave o the passing of even more unfavorable legislation. There is
little evidence, however, that this was the case in 2007. True, there were proposals on the
table such as single payer systems that were (traditionally) an anathema for business, but
these proposals were not as politically salable as proposals including employer mandates
(Jacobs and Hacker, 2009). Moreover, with Schwarzenegger as governor until 2011, it was
more than certain that even if passed through the legislature, they would have been vetoed
by the governor.
24
While the analysis so far has centered around California, it is important
to note that neither in Maryland, where the rst \Fair Share" legislation passed, the \Big
Three's" support for an employer mandate does seem to have re
ected strategic acquiescence
to an adverse political climate. As Kazee (2009) argues, and in contrast to California where
Schwarzenegger had to veto single payer bills twice (in 2006 and 2008), in Maryland there
was little discussion, until 2006, as far as expanding adult health insurance coverage was
concerned. It does therefore seem unlikely that business mobilized in Maryland so as to
avoid the passage of even less favorable legislation, although they appear to have played
an important role in shaping the structure of the employer requirement so as to aect only
Wal-Mart (Kazee, 2009).
An alternative explanation would attribute the increased interest in healthcare reform
in 2007, compared to 2003, to an unusual rise in the costs for health insurance in the
post-strike period. However, neither this explanation appears to be very robust. For while
24
Gov. Schwarzenegger had vetoed single-payer bills in the past such as SB 840
introduced by Senator Kuehl (D-Santa Monica) in 2006.
49
insurance premiums were continuing their upward trend, there is little indication that 2004-
07 was unusual as far as premium increases were concerned.
25
An argument along these
lines could only be salvaged if one could make a case that some (arbitrary?) cost threshold
was exceeded in 2007-08, a formidable task no doubt. Even more importantly, the creation
of the two-tier system had by 2007 dramatically reduced the coverage rates among grocery
workers (Dube and Jacobs, 2007). As a consequence of the drastically lower coverage rates
(coverage dropped from 94% to 54% in Southern California), the costs of employee insurance
for the chains were probably lower in 2007 compared to 2003, despite the increase in health
premiums that took place in the meantime.
26
Finally, it is possible that dierences in the policy proposals themselves sparked interest
in reform in 2007. According to this explanation, the grocers may have been supportive of
reform back in 2003 had they been facing the 2007 proposals. Alternatively, they may have
been reluctant to support reform in 2007 had they been facing the 2003 proposals.
27
I tackle
each possibility in turn.
There can be little doubt that there are non-negligible dierences between the Health
Insurance Act of 2003 and the \Health Care Security and Cost Reduction Act" of 2007.
These dierences pertain to the inclusion of the individual mandate, the expansion of Medi-
Cal and the increase of state subsidies to individuals not qualied for Medi-Cal. All of
the above were part of the later reform package but not of the 2003 legislation which was
essentially an employer mandate. The rst question that arises then is whether, taken
25
After the peak in premium increases in 2003, premiums had been increasing at a
constant rate, around 8% (California HealthCare Foundation, 2008).
26
In assessing the position of the \Big Three" in California in 2007, Andrew Wolf,
who covers the grocery sector for BB&T Capital Markets, argued that the \Big
Three's" \market share isn't back to what it was, but the two-tier (employee sys-
tem) keeps them more protable," (quoted in Darc e and Davies, 2007).
27
Obviously, the shape legislation takes as well as its timing is hardly independent
of the interests of the parties involved. In 2003, SB 2 was \widely seen as a union bill"
whose \language was largely written by a lobbyist for Health Access and other labor
representatives" (Kazee, 2009). Later, in 2007, Schwarzenegger probably pushed for
reform in anticipation of some business support, especially from the unionized grocers.
The \endogeneity", however, of the proposed legislation can be costlessly ignored for
our purposes here.
50
together, these additional provisions could have prompted the "Big Three" to support reform
back in 2003. Although, counterfactual questions are, perhaps, ultimately unanswerable the
following should be noted.
First of all, the individual mandate would not have a direct impact on the \Big Three",
although by reducing cost shifting, it may have slowed premium growth in the future.
28
More importantly, the individual mandate has usually been a major concern for insurance
companies, and not for business in general, when faced with the requirement of providing
guaranteed-issue coverage to all individuals.
29
It is thus reasonable to assume that the
individual mandate would have been an issue of rather secondary importance for the \Big
Three". Second, the expansion of eligibility to public programs (Medi-Cal) would have had
a negligible eect on the unionized chains, since employer sponsored coverage was almost
universal among grocery union workers in California at this point. Moreover, an expansion
of Medi-Cal and of state subsidies could have directly benetted those low wage/benets
competitors such as Wal-Mart, the labor practices of which the \Big Three" were trying so
hard to emulate.
30
Third, one should keep in mind that in 2003 the \Big Three's" goal was
to gradually eliminate insurance among grocery workers. Steven Burd, late in 2002, was
not shy in proclaiming Safeway's long run goals as far as worker benets were concerned. In
his own words, \In the eastern U.S., many states have these features (i.e two-tier contracts,
28
There is a long and inconclusive debate on the existence and size of cost shift-
ing. A recent analysis by the Congressional Budget Oce concluded that \Overall,
the impact of cost shifting on payment rates and premiums for private insurance
seems likely to be relatively small." (Key Issues in Analyzing Major Health Insurance
Proposals, Congressional Budget Oce, December, 2008).
29
Requiring insurers to provide coverage irrespective of preexisting conditions would
create a serious risk of adverse selection if not coupled with an individual mandate.
The reason is that individuals may decide to buy insurance only after they get sick.
30
For a discussion on the benets/costs of the expansion of public programs for
business see Kazee, 2006, pp. 17-23. According to a report prepared by U.C. Berke-
ley's Labor Center in 2004, only half of Wal-Mart workers in California had employer
sponsored coverage (Dube and Jacobs, 2004). This contrasts sharply with the nearly
universal coverage among grocery workers employed in the \Big Three" in the state.
As a result, a much higher percentage of Wal-Mart workers was relying on public
programs such as Medi-Cal.
51
italics mine) as common parts of their contracts, and ve years from now, this will be a
feature of every state with union operators" (quoted in Zwiebach 2002). According to Gregg
Donier, a spokesman for UFCW nationally at the time, the industry's four major employers,
Kroger, Albertsons, Safeway and Ahold, made it clear in their 2002 meeting with the union
at UFCW's headquarters in Washington that they wanted \to destroy health care benets"
and \their long-term objective is to not provide any health care coverage (italics mine).
The bottom line is they want to ratchet wage and benet costs down to Wal-Mart levels"
(quoted in Ghitelman et al. 2003). The above statements clearly point to a very ambitious
agenda by the part of the grocers as far as health benets were concerned. Therefore, and
given this maximalist agenda, it is highly unlikely that the \Big Three" would have been
interested in any form of government intervention in the form of a mandate back in 2003.
If the individual mandate would have entailed few and uncertain benets, the expansion of
public programs would have primarily benetted their main competitors and the employer
mandate would have thwarted their eorts to level health benets downwards, then there
is little that remains that could have led the \Big Three" to overcome business's inherent
distrust for government intervention and come out in support of reform in 2003. In all
probability, had the \Health Care Security and Cost Reduction Act" been presented to
them in 2003, it would have been ignored as blatantly as the ill fated 2003 legislation was.
It can be argued alternatively that had the \Big Three" been presented with the Health
Insurance Act of 2003 in 2007, they may have been much less eager to mobilize in favor
of reform. For example, both during the 2007-08 reform eorts in California and later
nationally, CAHR framed health reform as an issue of shared responsibility among business,
individuals and the government.
31
In contrast, the Health Insurance Act of 2003, a typical
\pay or play" law, was emphasizing only the business responsibility part of the problem.
It is thus conceivable that for ideological reasons the \Big Three" may have been reluctant
in 2007 to support legislation such as the Health Insurance Act, legislation that was placing
31
See, for example, the coalition's statement on September 30th, 2008. Also, the
statement released by CAHR regarding the merging of the Senate HELP and Finance
Committee Healthcare Bills (October 21st, 2009).
52
the onus of responsibility on business. While plausible, this argument does not square well
with the fact that approximately at the same time that reform was dened as a shared
responsibility issue in California, the \Big Three" had little trouble in supporting simple
employer mandates, in the form of \Fair Share" bills, in a number of other states. In other
words, it is hard to imagine that ideological reasons would have prevented support for an
employer mandate in California but would not have done so in states such as Maryland,
Washington and Missouri.
2.7 Conclusion
In this paper, I examined how the largest unionized food retailers gradually came to
embrace healthcare reform in the years preceding the passage of the landmark federal legis-
lation. These rms, I argued, played a catalytic role in supporting, and mobilizing others
to support, some of the most important attempts for health reform in the aftermath of
Clinton's health plan. I, also, argued that the \Big Three's" embrace for reform re
ected
\true" rather than \strategic" preferences to use Hacker and Pierson's (2002) terminology.
Namely, the realization that extensive cost shifting to union workers was too risky a strategy
to be adopted in the pursuit of higher prots.
The present analysis did not touch on the thornier question of how business support may
have aected the politics of healthcare reform. In the literature on employers and the welfare
state, this remains a highly divisive issue and scholars who may agree on how business
preferences are formed often disagree on what role employers have historically played in
the expansion of social insurance programs.
32
With respect to this question, preliminary
evidence suggests that even though big business support seems to have originated in one
sector of the economy, that is food retailing, its role may have been far from negligible. The
following serve as a suggestive illustration.
32
The debate on the employer role in the formulation/passage of the New Deal
legislation is particularly instructive on this issue (e.g. Hacker and Pierson, 2002 and
2004, Swenson, 2004).
53
First, the support of unionized grocers (i.e. Giant Food and Safeway) appears to have
been crucial in the passage of Maryland's \Fair Share" legislation. According to Vincent De
Marco, president of the Maryland's Citizen's Health Initiative (hereafter, MCHI) -the main
coalition of unions, consumer advocates and others that pushed for the legislation, it was
exactly Giant Food's support that \really opened the door" for the passage of the legislation
(Abelson, 2005). De Marco's position as the president of MCHI, a group which according
to Kazee (2009) was central in the state's health politics and which provided \consistent,
vocal, savvy support" to the FSHA, only adds credibility to his claims as he obviously would
have very little, if any, incentive to overstate the role of another party, that is business, in
the passage of legislation that he and the MCHI fought so hard for. Second, even though
California's ambitious reform eorts eventually failed, the vocal support of Safeway appears
to have been valuable to reform proponents. According to Kazee (2009), the \out front"
support by Safeway may have helped the reform eorts in at least three ways. First, by
keeping wavering democrats in line, second by neutering, or at least delaying, business oppo-
sition to the eorts, and nally by creating \the impression of broader business support".
Ultimately, as Kazee (2009) argues \employer support helped maintain momentum behind
a controversial plan, and push it remarkably far along in the process".
Moving to the national stage, two events warrant special attention. First, Obama's
choice of visiting Kroger's store in Virginia in order to pitch the reform eorts. Second, the
inclusion in the ACA of the so called \Safeway" amendment. The amendment allows health
plans to oer higher rewards, in terms of premium reductions, to employees participating
in employer sponsored wellness programs. Safeway, a pioneer in wellness programs, was the
strongest advocate of the amendment, sponsored by senators Carper (D-DE) and Ensign
(R-NV), and it is obviously not a coincidence that the Carper-Ensign amendment came to
be known as the \Safeway" amendment (Wojcik, 2010).
These events suggest the following. First, that the Obama administration, at its highest
echelons, was well aware of the receptive stance towards reform that the \Big Three" had
54
assumed.
33
Second, that the \Big Three" could, at least occasionally, wield signicant
leverage in shaping the nal legislation. As far as the extent of that support and of its
importance for the eventual triumph of the reformers are concerned, these remain questions
for future research.
33
Safeway's Steven Burd was among the CEOs invited to the White House by
Obama in May 2009. Later and during the debate over health reform both Obama
as well as leading politicians referred to the wellness program instituted by Safeway
as a model for the nation (Hilzenrath, 2010).
55
Table 2.1: Chronology of a Puzzle
Year Event
1994 Unionized grocers join NLCHCR, Burd urges the Clinton administration to
include the employer mandate in the nal legislation
2002 Start of the drive to impose two-tier contracts on UFCW nationwide
2003 Largest retail workers strike in the U.S. history, \Big Three" vs UFCW (South-
ern California)
2005-08 Unionized grocers support Fair Share bills in a number of states. CAHR is
formed in California to support the reform eorts that are underway
56
Table 2.2: Key events 2003/04
Date Event
March 19th, 2003 Labor leaders declare passage of SB 2 as labor's top legislative
priority for 2003
May 29th, 2003 The California Chamber of Commerce includes SB 2 in its annual
\Job killer" list
October 5th, 2003 The Health Insurance Act (i.e. SB 2) is signed by Gov. Davis
October 5th, 2003 Contract between the \Big Three" and the UFCW expires in
Southern California
October 10th 2003 The California Chamber of Commerce les for a referendum to
overturn the Health Insurance Act
October 11th, 2003 UFCW strikes Vons and the great grocery strike begins
February 29th, 2004 New contract is ratied by UFCW members, end of the strike
November 2nd, 2004 Proposition 72 is voted down, the Health Insurance Act is over-
turned
57
Table 2.3: Key events, 2007/08
Date Event
December 4th, 2006 AB 8 (Nunez) is introduced
January 8th, 2007 Gov. Schwarzenegger lays down his plan for healthcare reform
March 5th 2007 Contract between the \Big Three" and the UFCW expires in
Southern Calfornia
March 25th, 2007 UFCW gains strike authorization against Albertson
May 7th, 2007 Formation of CAHR is announced
July 22nd, 2007 New contract is ratied by UFCW members in Southern Califor-
nia
October 11th, 2007 Scwarzenegger vetoes AB 8 (Nunez)
December 17th, 2007 Bill AB 1x1 passes through the state Assembly
January 28th, 2008 State Senate Health Committee votes down bill AB 1x1
58
Bibliography
Abelson, Reed. 2005. States and Employers Duel Over Health Care. New York Times. (6
May, 2005) C.1.
Ainsworth, Bill. 2007. Unlike many CEOs, Safeway's Burd backs universal coverage. San
Diego Union Tribune. (10 May, 2007).
Angrist, Joshua and Alan Krueger. 1999. \Empirical Strategies in Labor Economics." In
The Handbook of Labor Economics, Volume III, eds. Orley Ashenfelter and David Card,
North Holland, 1277-1366.
Bailey, Michael A. 2005. \Welfare and the Multifaceted Decision to Move." American
Political Science Review 99: 125-135.
Bell, Winifred. 1965. Aid to Dependent Children,Columbia University Press, New York,
NY.
Belsh e, Kim. 2011. California's Role In Ensuring That The Potential Of Health Reform
Becomes Reality. Health Aairs 30: 71-75.
Boustan, Leah. P. and Robert A. Margo. 2009. \Race, Segregation, and Postal Employ-
ment: New Evidence on Spatial Mismatch." Journal of Urban Economics 65: 1-10.
California HealthCare Foundation. 2007. Employer Health Insurance Costs in the United
States. (July, 2007).
California Healthcare Foundation. 2008. California Employer Health Benets Survey.
(December, 2008).
Chung, Juliet. 2006. Britain's Tesco Plans to Enter U.S. Market With Chain of West Coast
Express Stores. Los Angeles Times, (10 February, 2006). C.1.
Cleeland, Nancy. 2003. Grocers, Clerks Remain Far Apart; As the threat of a strike as
early as Saturday looms, the two sides continued to negotiate over wages and benets. Los
Angeles Times (6 October, 2003). C.1.
59
Cleeland, Nancy and Marla Dickerson. 2003. Rising Health-Care Costs at Heart of Labor
Strife, Los Angeles Times, (14 October, 2003). A.1.
Cleeland, Nancy and Melinda Fulmer. 2003. California; A Question of Economics; The
unrest comes as Wall Street has been pressuring grocers to trim labor costs. Meanwhile,
sales and market share have been slipping. Los Angeles Times (13 October, 2003). C.1.
Cleeland, Nancy and Debora, Vrana. 2005. Wal-Mart CEO Takes His Case to California;
He says the retailer erred in its aggressive pursuit of the crucial market but won't back
down. Los Angeles Times (24 February, 2005). A.1.
Contreras, Julia and Orly Lobel. 2006. Wal-Martization and the Fair Share Health Care
Acts. St. Thomas Law Review 19: 105-136.
Darc e, Keith and Jennifer Davies. 2007. Last grocery strike's eects still felt. San Diego
Union Tribune. (4 March, 2007).
Dimmitt, Michael. 2007. Ninety Years of Health Insurance Reform Eorts in California,
California Research Bureau.
Dube, Arindrajit and Ken, Jacobs. 2004. Hidden costs of Wal-Mart jobs: Use of Safety
Net Programs by Wal-Mart Workers in California, (August, 2004). UC Berkeley Center for
Labor Research and Education.
Dube, Arindrajit and Ken, Jacobs. 2007. Impact of Health Benet Reductions in the
Unionized Grocery Sector in California, (August, 2007). UC Berkeley Center for Labor
Research and Education.
Dube, Arindrajit and Alex Lantsberg. 2004. Wage and health benet restructuring in
Californias grocery industry: Public Costs and Policy Implications, Report (6 July, 2004).
UC Berkeley Center for Labor Research and Education.
Fording, Richard C., Joe Soss, and Sanford Schram. 2007. \Devolution, Discretion, and the
Eect of Local Political Values on TANF Sanctioning." Social Service Review 81: 285-316.
Fording, Richard C., Joe Soss, and Sanford Schram. 2011. \Race and the Local Politics of
Punishment in the New World of Welfare." American Journal of Sociology 116: 1-48.
Fulmer, Melinda. 2003. Grocery Workers in Bay Area Hope to Avoid Strike, Los Angeles
Times, (23 September, 2004). C.1.
Fulmer, Melinda and Ronald White. 2003. Suppliers Caught in the Middle; Vendors shift
products to nonunion stores, but many say there is no way to make up for lost sales. Los
Angeles Times, (8 December, 2003). C.1.
60
Gainsborough, Juliet. F. 2003. \To Devolve or Not To Devolve? Welfare Reform in the
States." Policy Studies Journal 31: 603-623.
General Accounting Oce. 2000. Welfare Reform: State Sanction Policies and Number of
Families Aected. (HEHS-00-44 March 31, 2000)
Ghitelman, David, Hamstra Marka and Elliott Zwiebach. 2003. Strikes bring health care
benets issue to a head. Supermarket News. (20 October, 2003).
Gilens, Martin. 1999. Why Americans Hate Welfare: Race, Media, and the Politics of
Antipoverty Policy, University of Chicago Press,Chicago, IL.
Gordon, Colin. 2003. Dead on Arrival: The Politics of Health Care in Twentieth-Century
America, Princeton University Press, Princeton, NJ.
Gottschalk, Marie. 2000. The Shadow Welfare State: Labor, Business, and the Politics of
Health Care in the United States, Cornell University Press, Ithaca, NY.
Gruber, Jonathan. 2008. Population Movement Estimates for Health Care Reform
Under ABx1-1 with the Voter Initiative Filed December 28, 2007, 11 January
2008,http://www.calhealthreform.org/pdf/GruberAnalysis 011108.pdf (accessed 21 April
2008).
Hacker, Jacob. 2002. The Divided Welfare State: The Battle over Public and Private Social
Benets in the United States, Cambridge University Press, New York, NY.
Hacker, Jacob and Paul Pierson. 2002. Business Power and Social Policy: Employers and
the Formation of the American Welfare State. Politics & Society 30: 277-325.
Hacker, Jacob and Paul Pierson. 2004. Varieties of Capitalist Interests and Capitalist
Power: A Response to Swenson. Studies in American Political Development 18: 186-195.
Hacker, Jacob and Melissa Rodgers. 2009. Prescription for success: Lessons from California
for National Health Reform. Berkeley Center on Health, Economic & Family Security.
Policy Brief. June 2009.
Hamstra, Mark. 2009. Obama Pitches Reform at Kroger. Supermarket News, (3 August,
2009).
Hamstra, Mark, Jon Springer, Maria Tortoreto and Elliot Zwiebach. 2006. Mandatory
health bills get some industry support. Supermarket News, (27 May, 2006).
Handler, Joel. F and Yeheskel Hasenfeld. 2007. Blame Welfare: Ignore Poverty and
Inequality, Cambridge University Press, New York, NY.
61
Harbage, Peter, Len Nichols and Leif Wellington Haase. 2008. Lessons from California's
health reform eorts for the national debate. New America Foundation. March 2008.
Hasenfeld, Yeheskel and Lisa E. Powell. 2004. \The Role of Non-prot Agencies in the
Provision of Welfare-to-Work Services." Administration in Social Work 28: 91-110.
Hasenfeld, Jeheskel, Toorjo Ghose and Kandyce Larson. 2004. \The Logic of Sanctioning
Welfare Recipients: An Empirical Assessment." Social Service Review 78: 304-319.
Hiltzik, Michael. 2004. UFCW Sacrices Workers While Declaring Victory, Los Angeles
Times, (4 March, 2004). C.1.
Hilzenrath, David S. 2010. Misleading claims about Safeway wellness incentives shape
health-care bill. Washington Post. (17 January, 2010).
Hirsh, Lou. 2007. Expert says both sides know pain of grocery strike, Press Enterprise,
(23 June, 2007).
Hirsch, Jerry. 2007a. Grocery con
ict rooted in last strike; The wage deal of three years
ago still rankles workers, and greater competition since then is squeezing the stores. Los
Angeles Times, (23 April, 2007). A.1.
Hirsch, Jerry. 2007b. Grocery strike averted as chains, union reach accord; Workers are
expected to ratify the tentative four-year agreement. Los Angeles Times, (18 July, 2007).
A.1.
Hunt, Matthew O. 2007. \African American, Hispanic, and White Beliefs about
Black/White Inequality, 1977-2004." American Sociological Review 72: 390-415.
Ihlanfeldt, Keith R. and David L. Sjoquist. 1998. \The Spatial Mismatch Hypothesis:
A Review of Recent Studies and Their Implications for Welfare Reform." Housing Policy
Debate 9: 849-892.
Jacobs, Ken. 2009. A Tale of Two Tiers: Dividing Workers in the Age of Neoliberalism,
New Labor Forum 18: 67-77.
Jacobs, Ken and Jacob Hacker. 2009. How to Structure a "Play-or-Pay" Requirement on
Employers: Lessons from California for National Health Reform. Policy Brief (June, 2009).
UC Berkeley Center for Labor Research and Education.
Jones, Gregg. 2003. Labor Leaders Campaign for Health Coverage; Activists push for
passage of bill requiring California employers to provide insurance, Los Angeles Times (19
March, 2003). B.6.
62
Kalil, Ariel, Kristin S. Seefeldt and Hui-chen Wang. 2002. \Sanctions and Material Hardship
under TANF." Social Service Review 76: 642-662.
Kazee, Nicole. 2006. Wal-Mart Welfare: Medicaid and the Business Community, Paper
presented at the Midwest Political Science Association Annual Conference, Chicago, IL,
April 20-23, 2006.
Kazee, Nicole. 2009. Wal-Mart Welfare: Business, Fiscal Regime, and the Politics of
Health Policy in the American States. Ph.D diss. Yale University.
Keiser, Lael R, Peter R. Mueser and Seung-Whan Choi. 2004. \Race, Bureaucratic Discre-
tion, and the Implementation of Welfare Reform." American Journal of Political Science
48: 314-327.
Kennan, John and James R. Walker. 2010. \Wages, welfare benets and migration."
Journal of Econometrics 156: 229-238.
Kim, Byungkyu. 2008. \Second-Order Devolution, Bureaucratic discretion and the imple-
mentation of the Temporary Assistance For Needy Families Program." Ph.D diss. Univer-
sity of Kentucky.
Kinder, Donald and Nicholas Winter. 2001. \Exploring the Racial Divide: Blacks, Whites,
and Opinion on National Policy." American Journal of Political Science 45: 439-456.
Lieberman, Robert. 1998. Shifting the Color Line: Race and the American Welfare
State,Harvard University Press, Cambridge, MA.
Lipsky, Michael. 1980. Street Level Bureaucracy: Dilemmas of the Individual in Public
Services,Russell Sage Foundation, New York, NY.
Lurie, Irene. 2006. At the Front Lines of the Welfare System: A Perspective on the Decline
in Welfare Caseloads,The Rockefeller Institute Press, Albany, NY.
Mares, Isabela. 2003. The Sources of Business Interest in Social Insurance: Sectoral versus
National Dierences. World Politics 55: 229-258.
Martin, J. Cathie. 1995. Nature or Nurture? Sources of Firm Preference for National
Health Reform. American Political Science Review 89: 898-913.
Martinez, Steve. 2007. The U.S. Food Marketing: Recent Developments, 1997-2006, Eco-
nomic Research Report Number 42 (May 2007), United States Department of Agriculture.
McDonough, John, Michael Miller and Christine Barber. 2008. A Progress Report On State
Health Access Reform. Health Aairs 27: 105-115.
63
McKinsey Global Institute. 2001. U.S. Productivity Growth 1995-2000, Washington, DC,
October 2001.
Meyers, Marcia M. 2011. Personal communication with Marcia M. Meyers, School of Social
Work and Public Aairs, University of Washington.
Milkman, Ruth. 2005. Win or Lose: Lessons from Two Contrasting Union Campaigns.
Social Policy 35: 43-47.
Monnat, Shannon M. 2010. \The Color of Welfare Sanctioning: Exploring the Individual
and Contextual Roles of Race on TANF Case Closures and Benet Reductions." Sociological
Quarterly 51: 678-707.
Neubeck, Kenneth J. and Noel A. Cazenave. 2001. Welfare Racism: Playing the Race Card
Against America's Poor,Routledge, New York, NY.
Nicholson, Stephen P., Adrian D. Pantoja and Gary M. Segura. 2005. \Race Matters:
Latino Racial Identities and Political Beliefs." Presented at the Annual Meeting of the
American Political Science Association, Washington, DC.
Piven, Frances F. and Richard A. Cloward. 1971. Regulating the Poor: The Functions of
Public Welfare,Vintage Books, New York, NY.
Peltz, James. 2003. Wall Street Is Chains' Not-So-Silent Partner; Analysts have been
supporting the stores' stance on cost cuts, but not all are touting the companies' stocks.
Los Angeles Times (22 December, 2003). C.1.
Quadagno, Jill. 1994. The Color of Welfare: How Racism Undermined the War on
Poverty,Oxford University Press, New York, NY.
Ralph, Thomas. 2006. Safeway urges quick action on employee health benets. Seattle
Times (14 February, 2006).
Rau, Jordan. 2007a. Universal healthcare gains unlikely backer; A big-business group,
breaking ranks with smaller rms, will lobby Sacramento and D.C. to expand coverage to
all, Los Angeles Times (7 May, 2007), A.1
Rau, Jordan. 2007b. All Californians would be required to carry medical insurance.; $12-
BILLION PROPOSAL; Critics include doctors, hospitals, insurers and some legislators.
Los Angeles Times, (9 January, 2007). A.1.
Rau, Jordan. 2007c. Gov. urges wellness coverage; Administration wants to make insurers
provide preventive programs that oer incentives. Los Angeles Times, (5 January, 2007).
B.1.
64
Riccucci, Norma M. 2005. How Management Matters: Street Level Bureaucrats and Welfare
Reform,Georgetown University Press, Washington, D.C.
Riccucci, Norma M. 2011. Personal communication with Norma M. Riccucci, School of
Public Aairs and Administration, Rutgers University.
Riccucci, Norma M. and Marcia K. Meyers. 2004. \Linking Passive and Active Representa-
tion: The case of Frontline Workers in Welfare Agencies." Journal of Public Administration
Research and Theory 14: 585-597.
Rosenblatt, Robert and Karen Tumulty. 1994. 50 Firms Back Health Mandate; Democrats
Meet, New York Times, (22 July, 1994). pg1.
Samuels, Alana. 2007. Chances of supermarket strike climb; Union leaders quit talks after
learning the big chains signed a pact that could require lockouts. Los Angeles Times, (5
April, 2007). C.1.
Sandfort, Jodi. R. 2000. \Moving beyond discretion and outcomes: Examining public
management from the front lines of the welfare system." Journal of Public Administration
Research and Theory 10 :729-756.
Sieb, F. Gerald. 2009. California Stages Trial Run for U.S. Health-Care Overhaul. Wall
Street Journal, (20 May, 2009). A.2.
Soss, Joe, Linda Houser, Richard Fording and Sanford F. Schram. 2009. \Deciding to
Discipline: Race, Choice, and Punishment at the Frontlines of Welfare Reform." American
Sociological Review 74: 398-422.
Soss, Joe, Sanford F. Schram, Thomas P. Vartanian and Erin O'Brien. 2001. \Setting the
Terms of Relief: Explaining State Policy Choices in the Devolution Revolution." American
Journal of Political Science 45: 378-395.
Stoll, Michael. 2006. \Job Sprawl, Spatial Mismatch and Black Employment Disadvantage."
Journal of Policy Analysis and Management 25: 827-854.
Swenson, Peter. 2002. Capitalists against the Markets,Oxford University Press, New York,
NY.
Swenson, Peter. 2004. Varieties of Capitalist Interests: Power, Institutions, and the Reg-
ulatory Welfare State in the United States and Sweden. Studies in American Political
Development 18: 1-29.
Swenson, Peter and Scott, Greer. 2002. Foul Weather Friends: Big Business and Health
Care Reform in the 1990s in Historical Perspective, Journal of Health Politics, Policy and
Law 27: 605-638.
65
Washburn, David. 2005. Strike's benets realized in Bay Area, San Diego Union-Tribune,
(30 January, 2005).
Watkins-Hayes, Celeste. 2009. The New Welfare Bureaucrats: Entanglements of Race,
Class, and Policy Reform., University of Chicago Press, Chicago, IL.
Winter, Nicholas J. G. 2006. \Beyond Welfare: Framing and the Racialization of White
Opinion on Social Security." American Journal of Political Science 50: 400-420.
Wojcik, Joanne. 2010. Grocery chain stocks health plan services. Business Insurance 44:
3-5.
Wu, Chi-Fang, Marcia Cancian, Daniel R. Meyer and Georey L. Wallace. 2006. \How Do
Welfare Sanctions Work?" Social Work Research 30: 33-50.
Zwiebach, Elliot .2002. Safeways plan to revitalize; Safeways Steve Burd announces a four-
pronged strategy to Infuse health back into the companys earnings and stimulate customer
spending. Supermarket News, (18 November, 2002).
Zwiebach, Elliot .2010. Health Bill Could Add More Costs. Supermarket News, (29 March,
2010).
66

Abstract (if available)

Abstract The dissertation deals with key aspects pertaining to the politics of the Welfare State. In the first part of the dissertation, I examine whether, and how, policy implementation is affected by the personal characteristics of the street level welfare bureaucrats. I then turn my attention to issues surrounding the role of business in the development of progressive social legislation. Specifically, in the second part of the dissertation, I analyse the circumstances under which a segment of the big business community came to adopt a positive stance towards healthcare reform. The aforementioned questions are investigated in the context of the United States. In what follows, I provide an overview of the questions and the key findings of the dissertation. ❧ Chapter 1 examines how the personal characteristics of the street level bureaucrats affect the implementation of welfare policy. While both theory and anecdotal evidence point to the central role of low level bureaucracy in the administration of welfare programs in the U.S., no previous empirical study has examined whether, and how, the personal characteristics of the street level bureaucrats indeed affect how benefits are allocated and sanctions are imposed on welfare clients. This work attempts to fill this gap in the literature. By matching individual level data on welfare clients with county level data on welfare workers, I find that the racial composition of the welfare bureaucracy matters for clients. Specifically, I show that as the presence of whites among welfare workers grows stronger, both black and white clients face an increased likelihood of being cut off from welfare. I argue that these results, which are robust to controlling for the racial composition of the county, reflect the more conservative attitudes held by whites towards welfare and welfare recipients in general. This work provides thus, for the first time, systematic evidence that the racial characteristics of the welfare bureaucracy affect how welfare policy is implemented at the street level. ❧ Chapter 2 investigates the reasons behind big business mobilization in favor of health- care reform in the years preceding the passage of the landmark Patient Protection and Affordable Care Act of 2010. I specifically examine why large unionized firms, such as the ""Big Three"" food retailers, breaking ranks with the major business organizations, spear- headed some of the most important state level reform efforts in the wake of Clinton's plan debacle. I argue that their embrace for comprehensive health reform and employer man- dates in particular reflects their failure to vastly reduce their labor costs through radically restructuring, along two-tier lines, their contracts with the unions. In light of their failure to privately solve the health cost crisis, these firms turned to government for a solution. Initially through the support for ""Fair Share"" legislation in a number of states and, later, through the formation of ad hoc business coalitions, such as the ""Coalition to Advance Healthcare Reform"" in California, the ""Big Three"" signaled their openness to employer mandates, initially, and to more comprehensive solutions, eventually, along lines fundamentally similar to the federal legislation passed in 2010. This work contributes both to the understanding of the politics of healthcare reform in the United States as well as to the larger literature on the role of business in the development of progressive social legislation in capitalist societies.

Linked assets

University of Southern California Dissertations and Theses

Conceptually similar

PDF

Essays on the new economic history of the Middle East

PDF

Perspectives on state capacity and the political geography of conflict

PDF

Three essays on estimating the effects of government programs and policies on health care among disadvantaged population

PDF

Essays on work, retirement, and fostering longer working lives

PDF

The political economy of implementation: intellectual property rights protection across the world

PDF

Essays on health insurance programs and policies

PDF

Essays on macroeconomics and income distribution

PDF

The murky risk of trade protectionism in an interconnected and uncertain global economy: a state, industrial, and regional analysis

PDF

Three essays on the political economy of federal spending

PDF

Three essays on behavioral economics approaches to understanding the implications of mental health stigma

PDF

The impact of trade liberlaization on firm performance in developing countries -- new evidence from Pakistani manufacturing sector

PDF

Essays on health and aging with focus on the spillover of human capital

PDF

Three essays on the microeconometric analysis of the labor market

PDF

Simulation modeling to evaluate cost-benefit of multi-level screening strategies involving behavioral components to improve compliance: the example of diabetic retinopathy

PDF

Adoption of virtual healthcare, self-sufficient wages and paid neighbors concept will ensure optimal living for vulnerable people and their paid caregivers

PDF

The political psychology of scarcity in American society: a study of culture, discourse, and public opinion

PDF

Essays on improving human interactions with humans, algorithms, and technologies for better healthcare outcomes

PDF

Essays on the economics of climate change adaptation in developing countries

PDF

The stars and the state: astronomy, astrology, and the politics of natural knowledge in early medieval Japan

PDF

Three essays on heterogeneous responses to macroeconomic shocks

Asset Metadata

Creator Pipinis, Dimitris (author)

Core Title Essays on the politics of the American welfare state

School College of Letters, Arts and Sciences

Degree Doctor of Philosophy

Degree Program Economics

Publication Date 07/09/2012

Defense Date 07/09/2012

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

Tag business,healthcare reform,OAI-PMH Harvest,Patient Protection and Affordable Care Act,public opinion,Temporary Assistance for Needy Families,welfare bureaucracy

Language English

Contributor Electronically uploaded by the author (provenance)

Advisor Nugent, Jeffrey B. (committee chair), Easterlin, Richard A. (committee member), Wong, Janelle (committee member)

Creator Email pipinis@usc.edu,pipinisd@gmai.com

Permanent Link (DOI) https://doi.org/10.25549/usctheses-c3-54295

Unique identifier UC11288292

Identifier usctheses-c3-54295 (legacy record id)

Legacy Identifier etd-PipinisDim-928.pdf

Dmrecord 54295

Document Type Dissertation

Rights Pipinis, Dimitris

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