Perceptions of Program Abuse and Support for Social Insurance

Date01 January 2021
AuthorAlbert H. Fang,Gregory A. Huber,Scott E. Bokemper
Published date01 January 2021
DOI10.1177/1532673X20924784
Subject MatterArticles
https://doi.org/10.1177/1532673X20924784
American Politics Research
2021, Vol. 49(1) 59 –75
© The Author(s) 2020
Article reuse guidelines:
sagepub.com/journals-permissions
DOI: 10.1177/1532673X20924784
journals.sagepub.com/home/apr
Article
“Over half the people on disability are either anxious or their
back hurts. Join the club. Who doesn’t get up a little anxious for
work every day and their back hurts?”
Senator Rand Paul (R-KY), January 14, 20151
The claim that program abuse is rampant in the nation’s dis-
ability insurance program, as illustrated by the epigraph
above, has become commonplace in contemporary American
political discourse.2 Do perceptions that disability benefits
are being abused by people who could work decrease support
for disability insurance? More generally, do perceptions of
abuse in public social insurance programs undercut support
for those programs?
These questions have grown in importance as government
social insurance programs have become targets for retrench-
ment despite both their important function in a market econ-
omy and their design, in which benefits are available only to
those who contribute to a program’s financing. Mandatory
contributions allow individuals to collectively pool their
resources to address individual risks. Because program benefi-
ciaries subsidize their own care and benefits are limited to
defined shocks, these risk pooling programs are often popular
vis-à-vis more traditional redistributive welfare programs. At
the same time, social insurance benefits raise the possibility of
abuse if individuals claim benefits despite not needing them
for the reasons specified by the program. This abuse may in
turn undercut support for social insurance by leading citizens
to believe beneficiaries are not deserving of assistance.3
An extensive observational survey literature documents
the correlation between beliefs about program abuse and the
deservingness of program beneficiaries and support for both
social insurance (e.g., government health care) and govern-
ment benefits more generally (e.g., welfare) (Aarøe &
Petersen, 2014; Petersen et al., 2011; van Oorschot, 2000).
While this literature has documented important associations
between beliefs about beneficiaries and support for govern-
ment programs, existing work is unable to demonstrate a
causal relationship between perceiving program abuse and
program support because perceptions of program abuse may
originate from many sources. Indeed, a key finding from
prior research is that individual-level characteristics, like
ideology, often explain program support. In addition, fea-
tures of a particular beneficiary, like their race, often affect
people’s perceptions of that beneficiary’s deservingness, but
that effect could arise for multiple reasons, for example,
924784APRXXX10.1177/1532673X20924784American Politics ResearchBokemper et al.
research-article2020
1Yale University, New Haven, CT, USA
2Independent Scholar, Seattle, WA, USA
Corresponding Author:
Gregory A. Huber, Forst Family Professor of Political Science,
Department of Political Science and Institution for Social and Policy
Studies, Yale University, 77 Prospect Street, PO Box 208209, New Haven,
CT 06520, USA.
Email: gregory.huber@yale.edu
Perceptions of Program Abuse and
Support for Social Insurance
Scott E. Bokemper1, Albert H. Fang2, and Gregory A. Huber1
Abstract
Do perceptions of abuse in social insurance programs undercut program support? Answering this causal question is difficult
because perceptions of program abuse can arise from multiple potential causes. Examining the case of disability insurance, we
circumvent this challenges using laboratory experiments to study the interplay between program abuse and program support.
Specifically, we test whether participants vote to reduce benefit levels when they observe program abuse, even if that abuse
is not directly costly to them. We use a labor market shock to induce program abuse and show that the observation of a
healthy worker receiving benefits causes workers who are unaffected by the shock to vote to lower benefits. This effect
arises only when reducing benefit levels also reduces taxes. Our results demonstrate a causal link between program abuse
and diminished support for social insurance, validating accounts that stress how violations of cooperative norms can undercut
socially beneficial government programs.
Keywords
political economy of social insurance, program abuse perceptions, support for social insurance, labor market shocks, free
riding, laboratory experiment
60 American Politics Research 49(1)
beliefs about program abuse, affect or prejudice toward the
potential beneficiary, and so on. Individuals may therefore
report perceptions of abuse (or non-deserving beneficiaries)
not because they have observed abuse (or undeserving
behavior), but because they are already opposed to the pro-
gram or have beliefs and attitudes about its beneficiaries.
More generally, any analysis that relies on as-given expressed
attitudes or beliefs about beneficiaries (e.g., whether they are
abusing the program) to predict program support will pro-
duce biased estimates unless all factors that explain those
beliefs and affect program support are accounted for.
The notion that deservingness is a key source of program
support originates in two distinct potential causal pathways.
The first is an account emphasizing self-interest—low
deservingness indicates people are using a program despite
not actually needing it or not having merited care (e.g.,
because they had not previously contributed), and that abuse
reduces the chance that a person’s own subsequent meritori-
ous claims will be met. For example, in Bokemper and
DeScioli (2018), abuse of a “public fund” depletes the fund
such that players with valid claims may not receive benefits.
Program abuse might therefore undercut program support
because as with any cooperative task, free-riding upsets an
equilibrium of shared public goods production. The second
causal pathway emphasizes instead an intrinsic motivation to
enforce appropriate norms. In this account, individuals seek
to punish those, or prevent abuse by, individuals who take
advantage of others because such behavior violates core
ideas about appropriate behavior (e.g., reciprocity), regard-
less of whether it affects one’s own self-interest in this par-
ticular case. For example, third party observers in public
goods games may sanction other players who violate expec-
tations about public goods production despite not having a
stake in that particular public good (Fehr & Gächter, 2002).
In this article, we examine whether this second mecha-
nism, an intrinsic concern for beneficiary deservingness,
causes reduced program support. To do so, we construct an
experimental setting that remove any direct material effect of
program abuse on player welfare and instead ask whether
observing program abuse reduces program support. This par-
ticular setting is a “hard case” for testing the intrinsic norms
account because program abuse is costless, but the program
supports players’ material self-interest. While prior experi-
mental work has examined reactions to norm violations
when players can sanction a specific individual, it is unknown
whether reduced deservingness will lead individuals to
reduce support for the overall program that is being abused.4
To answer this question, we designed and implemented a
series of laboratory experiments where subjects participate
in a novel disability insurance game. The structure of the
game first allows us to induce program abuse by some play-
ers and then randomly exposes other players to information
about that abuse. We then assess whether observing abuse
reduces support for disability insurance, even in a setting
where all players in expectation benefit from the social
insurance program and abuse does not endanger access to
benefits. Our baseline experiment demonstrates that by
reducing the market wage some workers receive, we are able
to induce program abuse, operationalized as claiming bene-
fits when one could in fact participate in the labor market.
Our core political experiments reveal that unaffected work-
ers who are randomly informed about other players’ abusive
behavior support lower benefits when funding those benefits
requires higher taxes. This work therefore isolates the causal
effect of perceiving program abuse on support for social
insurance programs.
Our design necessarily abstracts from many of the details
of real labor markets and social insurance programs. Nor do
we incorporate features like beneficiary characteristics or
social norms that may explain baseline differences in support
for social insurance, although for more universal social
insurance programs most individuals are eligible to receive
benefits conditional on prior contributions. Our simplifica-
tion, however, is an important advantage because it allows us
the control necessary to isolate the causal effect of program
abuse on program support. Furthermore, while our design is
stylized, it retains the core logic that prior work has sug-
gested is at play—individuals turning to social insurance to
counteract diminished labor market returns (e.g., Autor &
Duggan, 2003; Black et al., 2002) and observation of this
abusive behavior undercutting subsequent support for social
insurance. In addition, because our game uses monetary
incentives, we can create baseline preferences over out-
comes, thereby providing a degree of control not feasible
outside the lab setting (see, e.g., Woon, 2012). Likewise,
choices in this game are costly decisions, which reduces con-
cerns that survey responses are pure “cheap talk” that do not
correspond with actual behaviors.
While the experimental setting with costly choices pro-
vides researchers with a great deal of control, we remain sen-
sitive to issues of external validity. For example, our subjects
are undergraduates who participate in laboratory studies for
modest amounts of money. How to extrapolate our results to
other subjects and outside the laboratory setting is an issue we
discuss in greater detail below. Nonetheless, given the impor-
tance of the question of how to sustain citizen support for
social insurance and government spending more generally, as
well as the key theoretical role that prior work has ascribed to
ideas about deservingness and program abuse, we believe that
isolating this causal relationship is of great theoretical and
empirical importance, a key advantage of our approach. We
return to issues of external validity in the conclusion.
Program Abuse, Deservingness, and
Support for Social Insurance
A core theoretical concept for understanding public support
for government spending programs is the deservingness heu-
ristic. Differences in perceptions of deservingness across
recipients, as well as differences across programs (Cook &

To continue reading

Request your trial

VLEX uses login cookies to provide you with a better browsing experience. If you click on 'Accept' or continue browsing this site we consider that you accept our cookie policy. ACCEPT