Effects of welfare reform on women's voting participation.
Corman, Hope ; Dave, Dhaval ; Reichman, Nancy E. 等
Effects of welfare reform on women's voting participation.
I. INTRODUCTION
The broad goal of the landmark welfare reform legislation in the
United States in the 1990s was to reduce dependence on government
benefits by promoting work, encouraging marriage, and reducing
nonmarital childbearing. The legislation represented a convergence of
dissatisfaction with the welfare system on both sides of the political
spectrum, with welfare participation becoming viewed by many as a cause
of dependence rather than a consequence of disadvantage. The key
strategy for reducing dependence was to promote employment by imposing
work requirements as a condition for receiving benefits as well as time
limits on receipt of cash assistance. The basic argument was that labor
force participation would break a "culture of poverty" by
increasing self-sufficiency and reconnecting members of an increasingly
marginalized underclass to the mainstream ideals of a strong work ethic
and civic responsibility (Katz 2001).
In terms of increasing employment of low-skilled women and
decreasing welfare caseloads, welfare reform was deemed a great success,
at least before the Great Recession hit at the end of 2007. Temporary
Assistance to Needy Families (TANF) caseloads declined by 50% between
1997 and 2001 alone, and over half of TANF cases are now "child
only," meaning that adults in the household are not eligible
(Loprest 2012). Employment rates of low-skilled mothers rose
dramatically after the early 1990s, and there is strong consensus that
welfare reform played a major role (Schoeni and Blank 2000; Ziliak
2006).
A handful of studies have found that welfare reform reduced
undesirable behaviors that have often been ascribed to "welfare as
we knew it," providing some support for the mainstreaming argument.
Kaestner and Tarlov (2006) found that welfare reform reduced
women's binge drinking. Corman et al. (2013) found that welfare
reform led to declines in illicit drug use among women at risk for
relying on welfare, with some evidence indicating that the effects
operate, at least in part, through work incentive policies. Corman,
Dave, and Reichman (2014) found that welfare reform led to reductions in
women's property crime. This emerging literature supports the
widely embraced argument that welfare reform discourages antisocial
behavior and suggests that disenfranchised women have been brought from
the margins to the mainstream. However, as far as we know, the only
studies that have directly tested the widely held assumption that
welfare reform encourages mainstream behavior (other than work, which is
required) have focused on marriage or non-marital fertility and have
generally revealed weak or ambiguous effects (e.g., Blank 2002;
Gennetian and Knox 2003; Grogger and Karoly 2005). To test the
"culture of poverty" argument that making welfare much less of
an option encourages personal and civic responsibility, it is necessary
to go beyond marital status by considering direct measures of mainstream
behavior.
In this article, we investigate the effects of welfare reform on
voting, which is an important form of civic participation in democratic
societies but a fundamental right that many citizens do not exercise.
Exploiting changes in welfare policy across states and over time, and
comparing relevant population subgroups within an econometric
difference-in-differences (DD) framework, we use the November Current
Population Survey (CPS) with the added Voting and Registration
Supplement (VRS) to estimate the causal effects of welfare reform on
women's voting registration and voting participation from 1990 to
2004, the period during which welfare reform unfolded. We explore the
extent to which effects appear to have operated through employment as
well as the extent to which effects varied by state political
orientations and welfare strictness.
The findings from this study provide important information that
promises to inform culture of poverty debates and provide a more
complete picture of the effects of a major policy shift in the United
States that is still very much in effect today. The findings also
contribute to the political science literature on the determinants of
voting by providing a strong test of the effects of work incentives
(and, by inference, employment) on voting behavior.
II. BACKGROUND
A. Welfare Reform in the United States
The Personal Responsibility and Work Opportunity Reconciliation Act
(PRWORA) of 1996, often referred to as welfare reform, ended entitlement
to welfare benefits under Aid to Families with Dependent Children (AFDC)
and replaced the AFDC program with TANF block grants to states. Features
of the legislation were time limits on cash assistance, work
requirements as a condition for receiving benefits, stricter sanctions
for noncompliance with work requirements and other program rules,
stronger child support enforcement, and family caps that limited
benefits for additional children. The broad goals of PRWORA were to
reduce dependence on government benefits by promoting work, encouraging
marriage, and reducing nonmarital childbearing.
Although welfare reform is often dated to the landmark 1996 PRWORA
legislation, reforms actually started taking place in the early 1990s
when the Clinton Administration greatly expanded the use and scope of
"welfare waivers." Many policies and features of state waivers
were later incorporated into PRWORA. However, PRWORA departed from its
waiver precursors by imposing a "work first" approach that was
designed not only to reduce welfare dependence, but also to reconnect
members of an increasingly marginalized underclass to the mainstream
ideals of a strong work ethic and civic responsibility (Katz 2001).
PRWORA granted considerable discretion to states in establishing welfare
eligibility and program rules. As a result, there is substantial state
policy variation within the broad national regime of time-limited cash
assistance for which work is required.
B. Employment, Welfare, and Voting
Glaeser, Laibson, and Sacerdote (2002) examined individual
investments in activities that create "social capital,"
defined broadly as connections within social networks such as community
organizations and religious institutions. Individuals choose to engage
in such behaviors if the benefits outweigh the costs. By increasing
employment, welfare reform may increase civic participation (including,
perhaps, voting) by shifting women from the individualistic job of
homemaker to more socially interactive occupations and increasing their
participation in unions, but it could also decrease civic participation
through an increase in the opportunity cost of time. Welfare reform may
also increase civic participation as a result of the new normative
climate of increased personal and civic responsibility. In terms of
voting in particular, Feddersen (2004) offered theoretical reasons why
individuals vote even though doing so imposes a cost and is unlikely to
affect the outcome, one of which involves belonging to a social network
that has a stake in the election's outcome.
The political science literature makes it very clear that
education, employment, and income--all aspects of socioeconomic
status--are strongly associated with political participation and voter
turnout in particular (Verba, Schlozman, and Brady 1995). Most studies
testing for causal relationships between socioeconomic status and
political participation have focused on education, and there has been
considerable debate about whether and in what contexts education causes
political participation, sometimes operationalized as voting (e.g.,
Berinsky and Lenz 2011; Dinesen et al. 2016; Kam and Palmer 2008, 2011;
Milligan, Moretti. and Oreopoulos 2004; Pelkonen 2012; Sondheimer and
Green 2010). Moreover, studies of effects of education cannot be used to
make inferences about effects of employment or income, as different
aspects of socioeconomic status may operate in distinct ways--for
example, education may impart knowledge about politics or a sense of
civic obligation, employment may provide networks of information about
politics and the importance of engagement, and income may foster
investment in political outcomes. Strong evidence about effects of
employment or income on political participation is scarce, and existing
studies tend to focus on specific mechanisms rather than identify causal
effects. For example, findings by Schur (2003). using data from two
nationally representative U.S. household surveys conducted by the
Rutgers Center for Public Interest Polling following the November
elections in 1998 and 2000, suggest that being employed increases an
individual's political activities through increased income, civic
skills, political efficacy, and recruitment at work.
Few studies have attempted to estimate causal effects of welfare or
cash assistance on voting or any other form of civic participation.
However, a strand of political science literature suggests that
citizens' experiences with social welfare programs can affect
political participation by providing lessons in how citizens and
government relate, giving recipients a stake in maintaining or enhancing
program benefits, or providing resources that facilitate political
action (Soss 1999; Bruch. Ferree. and Soss 2010). According to Soss
(1999), clients experience a lack of power when dealing with welfare
agencies, which translates to powerlessness in the face of government
more generally and serves as a disincentive to vote. Recent findings by
Sugie (2015)--that partners of incarcerated men are less likely to
register and vote than similar partners of men who have not been
incarcerated--are consistent with the scenario of marginalization
through feelings of powerlessness in interactions with public
institutions. This research suggests that welfare reform would increase
women's political participation, including voting, by disconnecting
them from a system that fosters feelings of powerlessness--unless there
are offsetting effects (i.e., reduced stake in maintaining or enhancing
program benefits or reduced resources that facilitate political action).
Based on the above framework. Swartz et al. (2009) investigated
voter turnout among young women in Minnesota from the Youth Development
Study in 1996 and 2000 and found that welfare recipients were less
likely to vote in both elections than nonrecipients, suggesting that
reducing welfare participation increases voting. However, the Youth
Development study did not target poor women, and did not explicitly
address the possibility that unobserved factors confounded the
associations between welfare participation and voting. Thus, the
observed negative associations between welfare participation and
likelihood of voting cannot be interpreted as causal effects.
Baez et al. (2012) found that new recipients of a conditional cash
transfer program in Colombia were more likely to vote than matched
individuals whose incomes were just above the income threshold (and
therefore did not receive the transfers) and to vote for the incumbent
(under whose regime the cash transfer program was implemented) in the
2010 presidential election. The findings from this quasi-experimental
study suggest that low-income women are more likely to vote when
legislation has been passed that affects their cash benefits. However,
it is not clear whether these findings would apply to the United States,
where Edlund and Pande (2002) found that party preference is much more
responsive to economic changes for middle-income women than for poor
women.
Watson (2015) used British panel survey data to estimate the
effects on civic participation of several pieces of legislation
involving public income support for unmarried parents. The first
legislative change imposed quarterly work-focused interviews for
unmarried parents whose youngest child was 14 years old or more.
Subsequent legislative changes made unmarried parents with children 5
years of age or older ineligible for cash assistance as single parents,
and required them to enter the unemployed pool with obligations of
attendance at fortnightly interviews. Using DD analyses, Watson found
that the requirement that parents of older children engage in a
quarterly work-focused interview had no effect on voting, but that the
requirement of frequent interviews for parents of children 5 years old
or more had a negative effect on voting. It is not clear whether these
findings, which suggest that stricter welfare policies (such as work
requirements) would decrease voting among welfare recipients who do not
leave the rolls, would apply in the United States.
A qualitative study of community leaders found that the time
constraints of complying with TANF requirements reduce
community-building activities on the part of women (Jennings 2001). This
finding stands in contrast with the literature on socioeconomic status
and voting participation, which suggests that employment increases
voting but has not focused on the important but specific population of
women at risk for relying on welfare. On the other hand, Anderson,
Curtis, and Grabb (2006) found that civic participation of American
women decreased during the 1990s while it increased in other developed
countries, and speculated that increasing time commitment to paid work
alongside declining levels of public support may be responsible.
However, the links between welfare, employment, and civic participation
in the United States were not empirically established in that study.
Moreover, voting--which the authors did not study--is likely to impose
fewer time constraints than would other forms of civic engagement.
III. DATA
We use data from the November CPS with the added VRS for the years
1990 through 2004, which span the implementation of welfare reform. The
general CPS is a monthly nationally representative survey of over 50,000
households that collects detailed information on labor force
participation as well as sociodemographic characteristics of each
household member.
The VRS takes place bi-annually at the end of November, in
even-numbered years when Congressional elections occur. Thus, every
second supplement takes place during a presidential election, when
voting turnout is higher (DeSilver 2014). The survey asks household
members if they are eligible to vote, whether they had registered to
vote by the election that occurred that month, and whether they had
voted in that election.
The VRS has several key strengths and is particularly well-suited
for addressing our research question. First, it was designed and
conducted to produce comprehensive national data on voter
characteristics, participation, and trends in the United States. In
contrast, administrative datasets include no information on individuals
who are not registered to vote, contain limited demographic information
on those who are registered, are not available for all states, and often
are not available for long enough time periods to ascertain trends.
Second, the VRS has a high response rate and the largest sample of any
survey used to measure voting participation, allowing us to maximize
statistical power. Other national surveys (e.g.. General Social Survey,
American National Election Studies--ANES, and Cooperative Congressional
Election Study) either have much smaller samples or are limited to
periods subsequent to the implementation of welfare reform. Third, while
it has long been recognized that individuals may overstate their voting
participation, this bias is minimal in the VRS compared to other
surveys. For example, the 2012 VRS turnout rate was 61.8%, compared to
58.6% from actual ballots, while the ANES reported a turnout rate of 78%
in 2012 (U.S. Election Project 2016a). Although the VRS has never been
subjected to a voter validation study, which compares respondents'
reports about whether they voted with local registration and voting
records, it consistently produces more accurate estimates of turnout
than other surveys, including ANES (Pew Center on the States 2012),
which reportedly has been unable to locate the voting records of a
sizable percentage of its respondents (McDonald 2007). In this study, we
focus on women who are at least 21 years old and up to age 49, a group
that is both eligible to vote and likely to have minor children living
in their household. Of those, we only include women who are citizens and
thus eligible to register to vote.
We initially follow the standard in the welfare reform literature
by using state-specific and time-varying indicators for both AFDC
waivers and TANF (Blank 2002). Twenty-nine states enacted AFDC waivers,
across various months, from 1992 to 1996 (see Appendix, Table A1). We
use a dichotomous variable that equals 1 if a statewide waiver was in
place before November of that year that substantially altered the nature
of AFDC with respect to time limits, sanctions, or work requirements. We
also use a dichotomous indicator for whether, before November of that
year, the state had implemented TANF. The data on whether states had
waivers and when they enacted TANF come from U.S. Department of Health
and Human Services (1999). Although the new welfare regime is very much
in effect today, our observation window coincides with, and exploits,
maximum policy and implementation change.
We separately consider voting in any election, which can take place
every 2 years, and voting in presidential elections, which can take
place every 4 years. The latter is the more relevant measure of
engagement for our study because (1) as indicated earlier, voting rates
are much lower in congressional elections than in presidential
elections, and (2) within congressional elections, voting rates are
highest for individuals who are age 65 years and older. non-Hispanic
white, have high levels of education, and have relatively high incomes
(File 2015)--individuals not at all likely to rely on welfare. Thus, the
more relevant margin for potential effects of welfare reform on voting
would be presidential elections.
Because presidential elections are more relevant for our study than
off-year elections and occur only every 4 years, separate measures of
AFDC waivers and TANF are based on very few elections. We thus also use
a single measure of Any Welfare Reform implementation that has also been
used in the relevant literature (e.g., Dave et al. 2011) and is defined
as either a major AFDC waiver or TANF in a given state and year. Most
notably, only two states (Michigan and New Jersey) implemented waivers
before the 1992 presidential election and these were implemented in
October 1992 when the election process was well underway, while 28
states implemented waivers before the 1996 presidential election (see
Appendix, Table Al). As such, separate waiver effects would almost
exclusively reflect the 1996 presidential election (Clinton's
re-election), which had a historically low voter turnout (U.S. Census
Bureau 2012). Thus, the single measure of Any Welfare Reform is our
preferred measure for this study. Nevertheless, we also show some
estimates using the separate measures of AFDC waivers and TANF in order
to map the study to the existing literature but urge the reader to
interpret the separate effects, particularly those for AFDC waivers,
with caution. Other advantages of the single measure of Any Welfare
Reform are that it facilitates ease of discussion and comparison across
models and preserves statistical power for subsequent stratification
analyses.
IV. METHODS
We employ a quasi-experimental research design--akin to a pre- and
post-comparison with treatment and control groups--in conjunction with
multivariate regression methods, broadly referred to as DD models, to
estimate the effects of welfare reform on women's voting behavior.
We conduct numerous specification checks and tests to investigate the
validity of the identification assumptions underlying our methodology,
the robustness of our results, and patterns across subgroups of mothers.
The basic model can be expressed as follows, where Y refers to the
outcome for the ;th individual residing in state s at time t:
(1) [Y.sub.ist] = [[alpha].sub.1] + [[pi].sub.1] (AFDC
[Waiver.sub.st]) + [[pi].sub.2] ([TANF.sub.st]) + [X.sub.ist] [beta] +
[Z.sub.st][delta] + [State.sub.s][lambda] + [Year.sub.t] [phi] +
[[epsilon].sub.ist].
AFDC Waiver and TANF are dichotomous variables indicating whether a
major waiver or TANF had been implemented. (1) X represents a vector of
individual characteristics (age, age-squared, race, Hispanic origin,
marital history including widowed/divorced/separated and never married,
number of children in the household, number of adults in the household,
and metropolitan statistical area--MSA residence), and Z represents a
vector of state-level characteristics (current and 1-year lagged state
unemployment rate, current and 1-year lagged state personal income per
capita, log female population, state poverty rate, state minimum wage,
1--and 2-year lagged welfare caseloads, percentages of the state
legislature that were Democrat/Republican, Republican governor,
Democratic governor, and registration/voting rates of males). (2) We do
not control for education, employment, or family income in the baseline
specification since prior work has shown that education and employment,
which are associated with income, were affected by welfare reform (e.g.,
Dave, Corman, and Reichman 2012; Ziliak 2006) and are potential pathways
by which welfare reform may affect voting behaviors. A full set of state
(State) and year (Year) fixed effects are included to capture unobserved
time-invariant state-specific factors, as well as overall national
trends. In the above specification, [[alpha].sub.1], [[pi].sub.1],
[[pi].sub.2], [beta], [delta], [lambda], and [phi] represent the vector
of parameters to be estimated.
Identification in this DD framework comes from comparing changes in
women's voting participation in states that have implemented
welfare reform to changes in states that have not yet done so, with the
implicit assumption being that the latter are a valid counterfactual for
the former (i.e., in the absence of welfare reform, trends in
women's voting participation would be parallel across the
"treatment" and control states). However, this "parallel
trends" assumption is unlikely to hold owing to a multitude of
time-varying state-specific factors that may be related to voting
behavior. That is, the coefficients of [[pi].sub.1], and [[pi].sub.2]
may represent not only effects of welfare reform, but also effects of
other state-specific trends that affect voting behavior over the time
period in which welfare reform was implemented.
We address the issue of potentially confounding state-specific
trends by adding another "difference"--comparing women at risk
for relying on welfare (target group) and similar women who are much
less at risk for relying on welfare (comparison group)--and estimating
difference-in-difference-in-differences (DDD) models as specified in
Equation (2) below. Target represents a dichotomous indicator equal to
one if the individual is in the target group and zero if the individual
is in the comparison group. (3)
(2) [Y.sub.ist] = [alpha]' + ([[alpha].sub.1] -
[[alpha].sup.*.sub.1]) [Target.sub.i] + ([[pi].sub.1] -
[[pi].sup.*.sub.1]) (AFDC [Waiver.sub.st] * [Target.sub.i]) +
([[pi].sub.2] - [[pi].sup.*.sub.2]) ([TANF.sub.st] * [Target.sub.i]) +
[[pi].sup.*.sub.1] (AFDC [Waiver.sub.st]) + [[pi].sup.*.sub.2]
([TANF.sub.st]) + [X.sub.ist] [beta]' + [Z.sub.st] [delta]' +
[State.sub.s] [lambda]' + [Year.sub.t] [phi]' +
[[eta].sub.ist].
The choice of target and comparison groups is integral to a valid
implementation of the DDD methodology. Identifying the target
group--individuals who are at risk of relying on public assistance--is
relatively straightforward; welfare recipients have traditionally come
from low-educated unmarried-parent households. The assumption necessary
for the DDD effect to represent an unbiased estimate is that in the
absence of welfare reform, unobserved state-varying factors would affect
the target and comparison groups similarly. If this assumption is valid,
then [[pi].sup.*.sub.1], and [[pi].sup.*.sub.2] will capture the effects
of any unmeasured trends that coincided with welfare reform and were
also related to voting. In other words, the basic welfare reform
indicators do not provide (and are not intended to provide) estimates of
the effects of welfare reform implementation on voting outcomes among
women exposed to welfare reform. The estimates of interest, which do
this job, are those for the interaction terms between the welfare reform
indicators and Target group indicator.
In our main analyses, we compare unmarried women ages 21-49 that
have a high school education or less and live with children (target
group) to two alternative comparison groups: (1) unmarried women with no
children in the same age and education groups, and (2) married mothers
in the same age and education groups. These comparison groups have been
validated in prior work on the effects of welfare reform on women's
education (Dave et al. 2011; Dave, Corman. and Reichman 2012), drug use
(Corman et al. 2013), and crime (Corman, Dave, and Reichman 2014) and in
the broader welfare reform literature. We assess the sensitivity of our
findings to the use of these two different comparison groups, as well as
to the use of three additional alternative comparison groups consisting
of unmarried mothers ages 21-49 with higher levels of education.
After establishing baseline estimates from DDD specifications, we
consider potentially confounding factors that may be correlated with
both the implementation of welfare reform and voting behaviors. In
particular, we consider voting legislation that coincided, to some
extent, with the implementation of welfare reform. The National Voter
Registration Act (NVRA) of 1993 was designed to ease the process of
registering to vote and maintaining registration by requiring states to
allow individuals to register when they applied for or renewed a
driver's license and when they visited an office providing public
assistance to the poor or disabled (U.S. Department of Justice 2015).
Because women in our target group were more likely than women in the
comparison groups to rely on welfare, the NVRA could have
disproportionately affected their voting behavior through more
interactions with public assistance offices.
Six states already had liberal registration practices in effect
when the NVRA was passed and the other states were required to implement
the law in 1995 (with 1996 representing the first national election
under this new regime). Although the NVRA was designed to increase voter
turnout, there is mixed evidence on its effect. Knack (1995) found a
lagged and positive effect of motor voter duration for states that were
early adopters. However, Knack (1999), who focused on the 1996 election,
and Brown and Wedeking (2006), who focused on elections through 2004
found no effects of these laws when controlling for past state election
turnout. This evidence suggests that the NVRA legislation would not
confound our estimated effects of welfare reform on voting.
Nevertheless, we assess the sensitivity of our estimates to controlling
for when each state implemented the NVRA. (4)
We further control for state-specific linear trends and, more
flexibly, include State*Year indicators instead of NVRA implementation.
The State*Year indicators account for all observed and unobserved
state-specific factors and thus also address concerns about policy
endogeneity--the possibility that the timing of welfare reform
implementation in states is a function of time-varying state-specific
factors. We also address the issue of potentially different underlying
trends in outcomes between the target and comparison groups by including
an interaction between a linear trend and the target group prior to
policy implementation.
Next, we assess differential effects across relevant state-level
margins--namely, the dominant political party of the state legislature
and strictness of the state's welfare regime. Poor unmarried women
tend to favor Democrats (Pew Research Center 2015) and may be more
invested in the political process in states with party dominance that
aligns with their political leanings. The latter analyses indirectly
inform the extent to which the policy effects are driven by shifts in
employment by considering a dose--response check--that is, whether the
observed policy effects are larger in states that had a more stringent
push toward employment as part of their welfare regimes. We supplement
these models by directly assessing the role of employment and family
income in mediating the observed effects of welfare policy on voting
participation.
In supplementary models, we allow for potential lagged effects, as
welfare reform may not affect behavior immediately. For example, if
employment is the main mechanism through which welfare reform affects
voting, women need time to find jobs and adjust their behavior in
response to the new situation. We thus estimate models that allow for 6-
and 12-month lags between welfare reform implementation and voting
responses.
The relative timing of elections and the rollout of welfare reform
suggest a number of additional specification checks. Specifically, we
assess sensitivity of our estimates to sequentially dropping the 1996,
2000, and 2004 elections, to ensure that our findings are not being
driven by any one presidential election, as well as to excluding African
American women from the sample, to address potential confounding effects
of sharp increases in black voter turnout since the 1996 presidential
election (Taylor 2013), which occurred just months after the enactment
of PRWORA. As a placebo test, we estimate effects of welfare reform on
voting outcomes among women in states that implemented welfare reform
relatively late (post September 1996), for whom we would not expect
welfare reform implementation to affect voting behaviors during the 1992
and 1996 presidential elections. As another placebo test, we estimate
effects of welfare reform on voting in the 1996 election among women in
the seven states that implemented welfare reform in September and
October of 1996, for whom the welfare reform happened too late to have
any effects on voting.
We also assess the robustness of our estimates to alternative
comparison groups based on education. In addition to our two main
comparison groups (unmarried women with no children and a high school
education or less, and married mothers with a high school education or
less), we alternatively compare how welfare reform affected voting
participation of the target group (unmarried mothers with a high school
education or less) relative to each of the following comparison groups:
unmarried mothers with some college but not a college degree, mothers
with any college education (degree or not), and mothers with at least a
college degree.
V. RESULTS
Table 1 shows mean percentages that registered and voted by year
for the target group and our two primary comparison groups--(1)
unmarried childless women ages 21-49 with a high school education or
less, and (2) married mothers in the same age and education groups. For
(1), the average difference between the target and comparison group in
the percent registered to vote in the pre-welfare reform years (1990 and
1992) was about 4 percentage points, while the average difference in the
post-welfare reform years (1998 to 2004) was zero. Thus, the difference
in the percent registered decreased. The corresponding differences for
voting were about 7 percentage points in the pre-welfare reform period
and about 3 percentage points in the post-welfare reform years. For (2),
the average difference between the target and comparison groups in the
percent registered to vote in the pre-welfare reform years was about 13
percentage points, and this difference decreased to 10 percentage points
in the post-welfare reform period. The corresponding difference in
percentage that voted was about 2.5 percentage points. These mean
differences are suggestive of "positive" effects of welfare
reform on voting behavior of women at risk for welfare reliance (i.e.,
smaller declines for the target group). However, these differences may
be confounded by other changes occurring over the period, and the
multivariate DDD models in subsequent tables address this concern.
Similar patterns are evident in Figures la and 1 b and Figures 2a
and 2b, for our target group (unmarried mothers age 21-49 with at most a
high school education) and the two primary comparison groups. The time
series in these figures correspond to numbers of years before and after
welfare reform was implemented in the woman's state of residence.
(5) Before welfare reform, women in the target group were less likely to
register and to vote than those in the comparison groups. However, as
welfare reform was implemented, the target group appeared to exhibit
behavior more like that of the comparison groups (particularly the
first). (6) This possible narrowing of the differences in registration
and voting patterns between women most at risk of welfare receipt and
similar women who were unlikely to be impacted by welfare policy, even
without conditioning on any other factors, is suggestive that welfare
reform may have attenuated decreases in voting among low-educated
unmarried mothers. However, as for the differences in means, these
trends may be confounded by other changes.
A downward trend in voting participation is also evident in Figures
2a and 2b over the pre-policy period. While voting turnout generally
follows national trends, there is considerable variation across states
in the magnitudes (and sometimes the direction) of the trend (U.S.
Election Project 2016b). For example, national trends indicate a
decrease of 6.4 percentage points between 1992 and 1996, years during
which welfare reform was implemented. This average trend masks
considerable heterogeneity across states--for example, the decline
during the same period was more than twice as high in Utah (13.8
percentage points) and close to zero in Wyoming (0.6 percentage points).
Based on our CPS sample, national trends account for 40% of the observed
variation in voting participation. Thus, there is considerable variation
in voting trends across states, driven by observable and unobservable
state-specific factors. The multivariate DDD models in subsequent tables
address the concern that the effects of welfare reform may be confounded
by such state trends.
From our November CPS data, we also found that employment increased
much more substantially for our target group than for either of the
comparison groups after the implementation of welfare reform.
Specifically, we compared the percentages of each group that were
employed before and after welfare reform was implemented in their state
of residence. In unadjusted results, the employment rate of the target
group grew 7.5 percentage points relative to comparison group 2 and 18.2
percentage points relative to comparison group l. (7) This difference,
which was expected based on past literature, provides further validation
of our target and comparison groups. Also validating, more generally, is
that in Table 1, we can see that for all low-educated women in the
relevant age range, both registration and voting were higher in
presidential election years than in nonpresidential election years.
The top panel of Table 2 presents regression results from linear
probability models predicting voting behavior using our first comparison
group (unmarried women with no children). The bottom panel of the same
table presents corresponding estimates using the second comparison group
(married mothers with at most a high school education). In both panels,
the estimates in the first two columns are for "registered to
vote" as the dependent variable; those in the next two columns are
for "voting in any even-numbered year," those in the fifth
column are for "voting in a presidential election year," and
those in the last column are for "voting in a non--presidential
even-numbered year" (congressional election). For the first two
outcomes there are two sets of regressions. The first corresponds to
Equation (2), wherein AFDC waivers and TANF are allowed to have separate
effects. The second combines AFDC waivers and TANF into one
indicator--Any Welfare Reform. For Models 5 and 6. we present results
only for Any Welfare Reform, our preferred measure for this article for
reasons discussed earlier.
These DDD estimates are consistent with the trends presented in
Table 1 and the figures, and indicate that welfare reform was associated
with smaller decreases in registration and voting among women exposed to
welfare reform compared to those who had similar characteristics but
were less exposed. Model 1 in Table 2 (top panel) suggests that TANF
attenuated declines in registration and voting by 3.5 and 2.7 percentage
points, respectively, but that the AFDC waivers had no effect. However,
as indicated earlier, the separate estimates for AFDC waivers and TANF.
particularly the former, should be interpreted with caution because they
are based on few elections. Using the same comparison group and the
single measure of Any Welfare Reform (our preferred measure for this
study), we find that welfare reform attenuated decreases in registration
and voting by 2.9 and 1.9 percentage points, respectively, and that, as
expected, the effects on voting were much stronger in presidential
election years (2.7 percentage points). These estimates translate to
effect sizes of 6.0, 7.0. and 6.4% relative to the baseline means for
registration, voting in any election, and voting in a presidential
election, respectively, for the target group. (8)
The results are similar when using the other comparison group
(Table 2, bottom panel), although the AFDC waivers as well as TANF were
associated with registration and voting (and statistically significant
for the former). Results using the single measure of Any Welfare Reform,
our preferred measure for this study, suggests that welfare reform
attenuated voting registration decreases by about 2.7 percentage points
(Model 2), voting in any election by 2.3 percentage points (Model 4),
and voting in a presidential election by 3.9 percentage points, with the
last translating to about a 9.2% effect relative to the baseline mean.
The magnitudes in Table 2 are plausible given that similar
specifications using both comparison groups find increases in employment
on the order of 6-10 percentage points among the target group of women
(reported in footnote 7). If we assume that employment is the main
channel by which welfare reform affected voting, this would suggest a
marginal "treatment-on-the-treated" effect of employment on
voting of 0.19 to 0.36, (9) which is the imputed instrumental variables
(IV) estimate using the indicator for Any Welfare Reform as an IV for
employment. While employment is likely the primary channel of effect,
welfare reform may also impact other determinants of voting behavior,
such as educational investments, income, labor supply at the intensive
margin, and the normative climate. To the extent to which this is the
case, this imputed IV effect would overstate the impact of employment on
voting. Nevertheless, we present it here as an upper bound estimate of
the effect of employment on voting, and also to gauge the credibility of
the magnitude. This estimated marginal effect of employment on voting
compares to the observed average propensity to vote, conditional on
employment, of 0.40 among the target group prior to welfare reform. The
similarity between the two magnitudes is validating and also suggests
that employment is a primary channel of effect underlying the link
between welfare reform and voting. (10)
As indicated earlier, a key concern underlying the estimates in
Table 2 relates to unobserved time-varying state factors that may be
correlated with the timing of welfare reform and which may also impact
voting participation. A related concern is that the comparison groups
may not be perfect counterfactuals for the target group, and thus may
not fully purge the effects of all unobserved time-varying
state-specific factors. We address these concerns in the models shown in
Table 3.
Results shown in the first column of Table 3 (Model 1) are the
"baseline estimates" from Model 5 in Table 2 which use the
single measure of Any Welfare Reform. One potentially important policy
confounder, as described in the previous section, is the NVRA, which
most states implemented in 1995 though several states had implemented
similar provisions earlier. Model 2 controls for the implementation of
the law or whether the state had implemented similar (or more liberal)
provisions in the past. While 1996 was the first national election year
under the NVRA, 14 states implemented welfare reform after the 1996
election, and 13 states had provisions similar to (or more liberal than)
the NVRA prior to 1996. Thus, there is substantial variation in welfare
reform implementation even after controlling for provisions of the NVRA.
The estimated effects of Any Welfare Reform on voting participation are
unaffected when introducing this control. Model 3 additionally controls
for unobserved time-varying state factors by including state-specific
linear trends, and the estimates are slightly smaller than those in
Models 1 and 2. Model 4 controls for all observed and unobserved
time-varying state factors through the inclusion of a full set of
State*Year fixed effects instead of indicators for NVRA or
state-specific linear trends. (11) The estimated effects of welfare
reform are somewhat smaller in magnitude but remain statistically
significant. As indicated earlier, saturating the model with State*Year
fixed effects also addresses concerns about policy endogeneity.
Another empirical concern related to state trends is that within a
given state, there may be differential outcome trends between the target
and comparison groups. In Model 5, we control for any such differential
pre-policy trends by including an interaction between a linear trend and
the target group. When using the first comparison group (top panel), the
estimated effect of Any Welfare Reform is smaller than that in the other
specifications and marginally significant. When using the second
comparison group (bottom panel), the estimated effect becomes larger
than that in any of the other specifications and suggests a 5 percentage
point relative increase (12) in voting participation; this compares to
an estimated 4 percentage point effect in the baseline specification
(Model 1) that does not control for the State*Year indicators or
differential pre-policy trends. Given the general robustness of the
estimates across all of these specifications and to preserve degrees of
freedom, we use Model 1 from these tables as the baseline for comparison
in subsequent specifications.
The results in Tables 2 and 3 consistently suggest that welfare
reform is associated with a relative increase in voting participation,
especially in presidential election years. These are mean effects
realized over all states, weighted by the target population in each
state. Table 4 assesses heterogeneity in these effects across relevant
state margins. Models 1 and 2 (using comparison group 1) and 5 and 6
(using comparison group 2) stratify the sample based on whether the
state's legislature had a Republican or a Democratic majority. As
expected, we find that the relative increase in voting participation
associated with welfare reform was driven primarily by Democratic
states; in states with a Republican majority in the legislature, welfare
reform is not associated with any significant change in voting
behaviors. Models 3 and 4 (and Models 7 and 8) stratify the sample into
strict/moderate versus weak work incentives based on the typology
developed by Blank and Schmidt (2001) that incorporates various features
of states' TANF programs (benefit generosity, earnings disregards,
sanctions, and time limits) and categorizes states as strong, weak, or
mixed in overall work incentives.
If our estimated effects of welfare reform on voting represent
causal links and operate through employment (the first-order effect of
welfare reform), we would expect the effects to be stronger in states
that enacted stricter pro-employment policies under welfare reform.
Owing to limited sample sizes upon stratification, standard errors
inflate and render some of the estimates imprecise. However, the
estimates are clearly larger in states with stronger employment-based
incentives, suggesting that the effects of welfare reform operate, at
least in part, through employment.
Table 5 directly assesses the mediating effects of employment as
well as those of education and family income. Models 1 and 5 add
employment status (an indicator for whether the woman is currently
employed) to the baseline specifications (Model 5 in Table 2). The
effects of welfare reform decline in magnitude, and as expected, current
employment is positively associated with voting participation. Models 2
and 6 instead include categorical indicators for weekly hours worked,
with nonemployment as the reference category. Models 3 and 7 further
include the natural log of family income, and the final specifications
(Models 4 and 8) add an indicator for high school completion. Comparing
the estimated effects of welfare reform in Models 4 and 8 with those
from the corresponding baseline models, we find that a large part of the
effects of welfare reform on voting appears to be mediated by
employment, income, and education--factors that prior studies found or
suggest were affected by welfare reform. The estimates in Models 4 and 8
again suggest that employment and family income encourage voting
participation, although conditional on employment, more working hours
reduce voting participation, likely owing to time constraints. Higher
educational attainment is also associated with voting participation. The
mediation analyses in Table 5, which are consistent with the
hypothesized effects, should be interpreted with caution since the
mediators are endogenous and constitute what Angrist and Pischke (2009)
refer to as "bad controls." (13) However, they provide a
useful first look at hypothesized pathways.
Finally, we consider a number of different specification checks and
placebo tests. We assess potential lagged effects in Table 6.
Specifically, Models 1 and 2 (using comparison group 1) and 5 and 6
(using comparison group 2) allow for welfare reform to impact voting
behavior with 6and 12-month lags, respectively. If there are no strong
lagged effects, we would expect the DDD coefficients in these models to
be substantially diminished in magnitude. However, the estimates remain
relatively robust, and with respect to the 12-month lag. increase
somewhat in magnitude (suggesting a 3.4 to 4 percentage point relative
increase in voting participation), which suggests some persistence and
some delay between implementation and changes in voting behavior.
Second, seven states implemented welfare reform in September and
October of 1996 (Appendix, Table Al) and it would not be reasonable to
expect any substantial effect of welfare reform on voting for these
states during the 1996 presidential election. Particularly, if we were
to find a substantive or significant effect of welfare reform for these
states, this would suggest that our treatment effects are biased and
reflective of differential trends in the reform-implementing states
between the target and comparison groups. We perform this check in
Models 3 and 7 (for comparison groups 1 and 2, respectively) in Table 6,
and it is validating that we do not find any such effects for this group
of states.
Third, for the 22 states and DC, which implemented welfare reform
relatively late (post September 1996), it would not be plausible for
welfare reform to affect voting during the 1992 and 1996 presidential
elections. This placebo test, shown in Models 4 and 8 in Table 6, is
akin to testing for significance of lead effects and it also informs if
there are differential pre-policy trends between the treatment and
comparison groups in late-implementing versus early-implementing states.
It is again validating that the coefficients are statistically
insignificant and close to zero. These results are consistent with
Models 4 and 5 in Table 3, which also showed that our estimates were not
sensitive to controlling for state-specific unobserved trends and
controlling for differential (parametric) pre-policy trends between the
target and control individuals.
In additional specification tests, we excluded individual
presidential election years (to be sure that the results were not driven
by any particular election) and excluded African Americans from the
sample (to address potential confounding effects of sharp increases in
black voter turnout since the 1996 presidential election). Appendix
Table A2 reproduces the results from Model 1 in Table 3 for the full
sample, as well as from specifications that sequentially drop
observations from 1996, 2000, and the 2004 election cycles (presidential
election years that took place after the implementation of welfare
reform). While these results are suggestive of some heterogeneity in the
magnitude of the welfare reform effects for the target group across
elections, the estimates are relatively robust to the exclusion of the
various presidential elections despite a loss of statistical power and
we cannot reject the null that the effects are the same across
specifications. These findings alleviate concerns that we do not have
enough time variation in the implementation of welfare reform to look at
presidential voting behavior, which only happens every 4 years.
Appendix Table A3 reproduces the results from Model 1 in Table 3,
along with corresponding results after excluding African Americans from
the sample. Though the decrease in sample sizes slightly reduces
precision, the DDD estimates remain stable in terms of both direction
and magnitude; they range from 2.0 to 2.9 and 2.2 to 5.6 percentage
points for comparison groups 1 and 2, respectively, when we exclude
African Americans, which compare to effects ranging from 2.0 to 2.7 and
2.5 to 5.1 percentage points, respectively, for the full sample. These
results confirm that the relative increases in black voter turnout
starting in 1996 did not confound our main estimates.
Finally, we assess the sensitivity of our estimates to additional
comparison groups consisting of unmarried women with higher levels of
education. These estimates are presented in Appendix Table A4. Models 1
and 2 reproduce the estimates from Model 1 in Table 3 based on our two
main comparison groups--low-educated unmarried women with no children
and low-educated married women with children (comparison groups
generally ineligible for welfare). In Models 3 through 5, we alternately
compare how welfare reform affected voting participation among
low-educated unmarried mothers (target group) relative to unmarried
mothers with some college though not a college degree (Model 3), mothers
with any college education (degree or not) (Model 4), and mothers at
least a college degree (Model 5). The coefficient for Any Welfare Reform
remains negative and in the range of 5-6 percentage points in these
models, indicative of the general downward trend in voting participation
among higher-educated unmarried mothers. The DDD estimates all fall in
the range of those of our two main comparison groups, suggesting a 3.1
to 3.4 percentage point relative increase in voting for the target group
relative to these alternate comparison groups.
VI. CONCLUSION
This study found robust evidence that welfare reform in the United
States in the 1990s, which coincided with ongoing declines in voter
turnout, led to relative increases in women's voting on the order
of 2 percentage points (for any even-year election) to 3-4 percentage
points (presidential election), which translates to about a 6 to 9%
effect relative to the baseline means. These robust findings suggest
that welfare reform had pro-social effects on civic participation, as
characterized by voting. We took great care to account for the
methodological challenges inherent in our study, including the potential
confounding effects of the NVRA of 1993, which coincided, to some
extent, with the implementation of welfare reform. The effects were
largely confined to presidential elections, were stronger in states in
which the majority of state legislators were Democrats, were stronger in
states with stronger work incentive policies, and appeared to operate
through employment, education, and income.
Although we ruled out the possibility that the NVRA confounded our
estimates, it is possible that this legislation played an indirect role
in affecting voting behavior of women at risk of relying on welfare.
Specifically, the "Motor Voter" feature of the NVRA required
state motor vehicle agencies to incorporate voter registration into
their driver's license, renewal, and change-of-address
applications. The substantial increases in women's employment as a
result of welfare reform may have increased voting registration, and
potentially voting participation, through increased demand for cars (for
which motor vehicle licenses are required) as a means of transportation
to work. Interestingly, the percentage of female householders under the
age of 35 in the United States who owned cars increased from 67.8% to
73.8% between 1993 and 2004, coinciding with the unfolding of welfare
reform. Unfortunately, we are unable to directly test this mechanism
with our data.
The findings from this study inform culture of poverty debates by
providing a rigorous test of the widely embraced argument, on both sides
of the political spectrum, that welfare reform brings women from the
margins to the mainstream and encourages pro-social behavior. It
complements previous studies finding that welfare reform reduced
women's binge drinking, illicit drug use, and property crime
(Corman, Dave, and Reichman 2014; Corman et al. 2013; Kaestner and
Tarlov 2006). As far as we know, no previous population-based studies
have investigated the causal effects of welfare policy in the United
States on voting, on any other form of civic participation, or on any
nontargeted mainstream behaviors more generally (targeted behaviors
being working and marriage), and no such studies of which we are aware
have found undesirable effects on nontargeted behaviors. As such, this
study adds to the growing evidence that welfare reform encourages
mainstream behavior and helps to provide a more complete picture of the
effects of a major policy shift in the United States that is still very
much in effect today and under which the next generation has been
raised.
The findings from this study also contribute to the political
science literature, in two ways. First, as indicated earlier, strong
evidence about effects of employment or income on political
participation is scarce or nonexistent. Our study contributes to this
literature by exploiting a large-scale policy shift that resulted in
substantial exogenous increases in employment and providing perhaps the
strongest evidence yet that employment improves voter turnout. That
said, the results cannot be generalized to groups other than women at
risk for reliance on welfare in the United States and we are not able to
explore potential mechanisms such as information networks at work.
Second, our finding that welfare reform (which would reduce interactions
with welfare agencies because it reduced welfare reliance) led to
relative increases in voting among women at risk of relying on welfare
than among similar women at much lower risk of relying on welfare
(defined multiple ways) is consistent with the scenario proposed by Soss
(1999) in which women's experiences with the welfare system
discourage political participation. However, this evidence is indirect
and thus only suggestive.
Future research is needed to further elucidate the micro-level
processes linking welfare policy, welfare participation, employment, and
voting.
APPENDIX
TABLE A1
Implementation Dates of Welfare Reform by State, United States
October 1992 September 1996 October 1992
to February to January to January
1997 1998 1998
AFDC Waiver TANF Any Welfare
Reform
Alabama November 96 November 96
Alaska July 97 July 97
Arizona November 95 October 96 November 95
Arkansas July 94 July 97 July 94
California December 92 January 98 December 92
Colorado July 97 July 97
Connecticut January 96 October 96 January 96
DC March 97 March 97
Delaware October 95 March 97 October 95
Florida October 96 October 96
Georgia January 94 January 97 January 94
Hawaii February 97 July 97 February 97
Idaho July 97 July 97
Illinois November 93 July 97 November 93
Indiana May 95 October 96 May-95
Iowa October 93 January 97 October 93
Kansas October 96 October 96
Kentucky October 96 October 96
Louisiana January 97 January 97
Maine November 96 November 96
Maryland March 96 December 96 March 96
Massachusetts November 95 September 96 November 95
Michigan October 92 September 96 October 92
Minnesota July 97 July 97
Mississippi October 95 July 97 October 95
Missouri June 95 December 96 June 95
Montana February 96 February 97 February 96
Nebraska October 95 December 96 October 95
Nevada December 96 December 96
New Hampshire October 96 October 96
New Jersey October 92 July 97 October 92
New Mexico July 97 July 97
New York November 97 November 97
North Carolina July 96 January 97 July 96
North Dakota July 97 July 97
Ohio July 96 October 96 July 96
Oklahoma October 96 October 96
Oregon February 93 October 96 February 93
Pennsylvania March 97 March 97
Rhode Island May 97 May 97
South Carolina October 96 October 96
South Dakota June 94 December 96 June 94
Tennessee September 96 October 96 September 96
Texas June 96 November 96 June 96
Utah January 93 October 96 January 93
Vermont July 94 September 96 July 94
Virginia July 95 February 97 July 95
Washington January 96 January 97 January 96
West Virginia February 96 January 97 February 96
Wisconsin January 96 September 97 January 96
Wyoming January 97 January 97
Source: U.S. Department of Health and Human Services (1999).
TABLE A2
Estimated Effects of Welfare Reform on Voting in Presidential
Elections. Overall and Excluding Specific Presidential Election
Years
Specification 1 2 3 4
Excluding Excluding Excluding
Sample All 1996 2000 2004
Comparison Group Comparison Group 1
Any Welfare -0.049 ** -0.057 ** -0.065 *** -0.050 **
Reform (0.019) (0.026) (0.021) (0.023)
Welfare 0.027 ** 0.047 *** 0.023 * 0.026 *
Reform*Target (0.011) (0.013) (0.013) (0.013)
N 15,669 11.826 12,323 12,034
Specification 5 6 7 8
Excluding Excluding Excluding
Sample All 1996 2000 2004
Comparison Group Comparison Group 2
Any Welfare -0.046 *** -0.023 -0.046 *** -0.041 ***
Reform (0.015) (0.022) (0.016) (0.015)
Welfare 0.039 *** 0.052 *** 0.052 *** 0.025 *
Reform*Target (0.013) (0.017) (0.014) (0.014)
N 25,796 19,126 20,476 20.374
Notes: Estimates in columns 1 and 5 are from specification I in panels
a and b. respectively, in Table 3. Corresponding estimates excluding
specific presidential elections from the sample are presented in the
other columns.
TABLE A3
Estimated Effects of Welfare Reform on Voting in Presidential
Elections, Excluding African Americans from the Sample
Specification 1 2 3
Comparison Group 1
Full sample
Welfare Reform * Target 0.027 ** 0.027 ** 0.026 **
(0.011) (0.011) (0.011)
N 15,669 15,669 15,669
Comparison Group 1
Excluding African Americans
Welfare Reform * Target 0.029 ** 0.029 ** 0.029 **
(0.013) (0.013) (0.013)
N 11,919 11,919 11,919
Comparison Group 2
Full sample
Welfare Reform * Target 0.039 *** 0.039 *** 0.039 ***
(0.013) (0.013) (0.013)
N 25,796 25,796 25,796
Comparison Group 2
Excluding African Americans
Welfare Reform * Target 0.038 ** 0.038 ** 0.039 **
(0.016) (0.016) (0.016)
N 22,078 22,078 22,078
Motor Voter law No Yes Yes
State linear trend No No Yes
State * Year indicators No No No
Target * Linear pre-policy trend No No No
Specification 4 5
Comparison Group 1
Full sample
Welfare Reform * Target 0.023 * 0.020 *
(0.011) (0.011)
N 20,682 20,682
Comparison Group 1
Excluding African Americans
Welfare Reform * Target 0.023 * 0.020 *
(0.013) (0.012)
N 15,407 15,407
Comparison Group 2
Full sample
Welfare Reform * Target 0.025 ** 0.051 ***
(0.011) (0.016)
N 35,482 35,482
Comparison Group 2
Excluding African Americans
Welfare Reform * Target 0.022 0.056 ***
(0.014) (0.019)
N 30,121 30,121
Motor Voter law No No
State linear trend No No
State * Year indicators Yes Yes
Target * Linear pre-policy trend No Yes
Notes: Each cell provides DDD estimates from a separate model
corresponding to specifications in Table 3.
TABLE A4
Estimated Effects of Welfare Reform on Voting in Presidential
Elections. Using Alternate Higher-education Comparison Groups
1 2
Target Group Unmarried Mothers with High School
Education or Less
Unmarried Women Married Mothers
with No Children with High School
and High School Education or Less
Comparison Education or Less (comparison
Group (comparison group 1) group 2)
Any Welfare -0.049 ** -0.046 ***
Reform (0.019) (0.015)
Welfare Reform 0.027 ** 0.039 ***
* Target (0.011) (0.013)
N 15,669 25,796
3 4 5
Target Group Unmarried Mothers with High School
Education or Less
Unmarried Unmarried Unmarried
Mothers with Mothers Mothers with
Some College with Any College a College
Comparison but No Degree (degree or not) Degree
Group
Any Welfare -0.57 ** -0.050 ** -0.059 **
Reform (0.024) (0.020) (0.027)
Welfare Reform 0.031 0.031 0.034 *
* Target (0.023) (0.019) (0.020)
N 15,142 17.236 11,309
Notes: Estimates in the first two columns are from Model 1 in panels
a and b, respectively, in Table 3. Corresponding estimates
using higher-education comparison groups of women ages
21-49 are presented in specifications 3-5.
ABBREVIATIONS
AFDC: Aid to Families with Dependent Children
ANES: American National Election Studies
CPS: Current Population Survey
DD: Difference-in-Differences
DDD: Difference-in-Difference-in-Differences
EDR: Election Day Registration
IV: Instrumental Variables
MSA: Metropolitan Statistical Area
NVRA: National Voter Registration Act
PRWORA: Personal Responsibility and Work Opportunity Reconciliation
Act
TANF: Temporary Assistance to Needy Families
VRS: Voting and Registration Supplement
doi: 10.1111/ecin.12433
Corman: Professor, Department of Economics. Rider University,
Lawrenceville, NJ 08648. Phone 609-895-5559, Fax 609-896-5387, E-mail
corman@rider.edu
Dave: Professor, Department of Economics, Bentley University, NBER
& IZA, Economics, Waltham, MA 02452. Phone 781-891-2268. Fax
917-426-7015. E-mail ddave @ bentley.edu
Reichman: Professor, Department of Pediatrics, Rutgers
University--Robert Wood Johnson Medical School, New Brunswick, NJ 08903,
and University of Toronto, Institute of Health Policy, Management &
Evaluation, Dalla Lana School of Public Health. Phone 732-235-7977, Fax
925-522-3345, E-mail reichmne@rwjms.rutgers.edu
REFERENCES
Anderson. R., J. Curtis, and E. Grabb. "Trends in Civic
Association Activity in Four Democracies: The Special Case of Women in
the United States." American Sociological Review, 71(3), 2006,
376-400.
Angrist, J., and J. S. Pischke. Mostly Harmless Econometrics: An
Empiricist's Companion. Princeton, NJ: Princeton University Press,
2009.
Baez, J. E., A. Camacho, E. Conover, and R. D. Zarate.
"Conditional Cash Transfers, Political Participation, and Voting
Behavior." World Bank Policy Research Working Paper No. 6215, 2012.
Berinsky, A. J., and G. S. Lenz. "Education and Political
Participation: Exploring the Causal Link." Political Behavior,
33(3), 2011, 357-73.
Blank, R. M. "Evaluating Welfare Reform in the United
States." Journal of Economic Literature, 40(4), 2002, 1105-66.
Blank, R. M., and L. Schmidt. "Work, Wages, and Welfare,"
in The New World of Welfare, edited by R. M. Blank and R. Haskins.
Washington, DC: Brookings Institution. 2001.
Brown. R. D., and J. Wedeking. "People Who Have Their Tickets
but Do Not Use Them: 'Motor Voter,' Registration, and Turnout
Revisited." American Politics Research, 34(4), 2006, 479-504.
Bruch, S. K., M. M. Ferree, and J. B. Soss. "From Policy to
Polity: Democracy, Paternalism, and the Incorporation of Disadvantaged
Citizens." American Sociological Review, 75(2), 2010, 205-26.
Corman, H., D. Dave, D. Das, and N. E. Reichman. "Effects of
Welfare Reform on Illicit Drug Use of Adult Women." Economic
Inquiry, 51(1), 2013, 653-74.
Corman, H., D. Dave, and N. E. Reichman. "Effects of Welfare
Reform on Women's Crime." International Review of Law and
Economics, 40. 2014. 1-14.
Dave, D., N. E. Reichman, H. Corman. and D. Das. "Effects of
Welfare Reform on Vocational Education and Training." Economics of
Education Review, 30(6), 2011, 1399-415.
Dave, D., H. Corman, and N. E. Reichman. "Effects of Welfare
Reform on Education Acquisition of Adult Women." Journal of Labor
Research, 33(2), 2012, 251-82.
DeSilver, D. "Voter Turnout Always Drops Off for Midterm
Elections, but Why? Facttank--News in the Numbers." Pew Research
Center. 2014. Accessed February 7, 2016.
http://www.pewresearch.org/fact-tank/2014/
07/24/voter-turnout-always-drops-off-for-midtermelections-but-why/
Dinesen, P. T., C. T. Dawes, M. Johannesson, R. Klemmensen, P.
Magnusson, A. S. N0rgaard. I. Petersen, and S. Oskarsson.
"Estimating the Impact of Education on Political Participation:
Evidence from Monozygotic Twins in the United States, Denmark and
Sweden." Political Behavior, 38, 2016, 579-601.
Edlund, L., and R. Pande. "Why Have Women Become Left-Wing?
The Political Gender Gap and the Decline in Marriage." Quarterly
Journal of Economics, 117(3), 2002, 917-61.
Feddersen, T. J. "Rational Choice Theory and the Paradox of
Not Voting." Journal of Economic Perspectives, 18(1), 2004, 99-112.
File, T. "Who Votes? Congressional Elections and the American
Electorate: 1978-2014, Population Characteristics, P20-577." U.S.
Census Bureau. Washington, DC, 2015. Accessed March 10, 2016.
https://www.census.gov/
content/dam/Census/library/publications/2015/demo/ p20-577.pdf.
Gennetian. L. A., and V. Knox. Staying Single: The Effects of
Welfare Reform Policies on Marriage and Cohabitation. New York:
MDRC.399, 2003.
Glaeser, E. L., D. Laibson, and B. Sacerdote. "An Economic
Approach to Social Capital." Economic Journal, 112(483), 2002,
F437-58.
Grogger, J., and L. A. Karoly. Welfare Reform: Effects of a Decade
of Change. Cambridge, MA: Harvard University Press, 2005.
Jennings, J. "Welfare Reform and Neighborhoods: Race and Civic
Participation." The Annals of the American Academy of Political and
Social Science, 577, 2001, 94-106.
Kaestner, R., and E. Tarlov. "Changes in the Welfare Caseload
and the Health of Low-Educated Mothers." Journal of Policy Analysis
and Management, 25(3), 2006.623-43.
Kam, C. D.. and C. L. Palmer. "Reconsidering the Effects of
Education on Political Participation." Journal of Politics, 70(3),
2008, 612-31.
--. "Rejoinder: Reinvestigating the Causal Relationship
between Higher Education and Political Participation." Journal of
Politics, 73(3), 2011, 659-63.
Katz, M. B. The Price of Citizenship: Redefining the American
Welfare State. New York: Metropolitan Books, Henry Holt & Co., 2001.
Knack, S. "Does 'Motor Voter' Work? Evidence from
State-level Data." Journal of Politics, 57(03), 1995, 796-811.
--. "Drivers Wanted: Motor Voter and the Election of
1996." PS: Political Science and Politics, 32(2), 1999, 237-43.
Loprest, P. J. "How Has the TANF Caseload Changed Over
Time?" 2012. Accessed February 8. 2015. http://
www.urban.org/UploadedPDF/412565-
How-Has-theTANF-Caseload-Changed-Over-Time.pdf
McDonald, M. "The True Electorate: A Cross-Validation of Voter
Registration Files and Election Survey Demographics." Public
Opinion Quarterly, 71(4), 2007, 588-602.
Milligan. K., E. Moretti. and P. Oreopoulos. "Does Education
Improve Citizenship? Evidence from the United States and the United
Kingdom." Journal of Public Economics, 88(9-10), 2004, 1667-95.
National Council of State Legislatures. "Same Day Voter
Registration." 2015. Accessed February 8, 2106. http://
www.ncsl.org/research/elections-and-campaigns/
same-day-registration.aspx
Pelkonen, P. "Length of Compulsory Education and Voter
Turnout--Evidence from a Staged Reform." Public Choice, 150, 2012,
51-75.
Pew Center on the States. "Election Administration by the
Numbers: An Analysis of Available Datasets and How to Use Them."
2012. Accessed June 19, 2016. http://
www.pewtrusts.org/~/media/legacy/uploadedfiles/pcs_
assets/2012/pewelectionsbythenumberspdf. pdf
Pew Research Center. "The Politics of Financial Insecurity: A
Democratic Tilt. Undercut by Low Participation." 2015. Accessed
February 7, 2016. http://www.peoplepress.org/files/20
15/01/1-8-15-Financial-.securityrelease.pdf
Schoeni, R. F., and R. M. Blank. "What Has Welfare Reform
Accomplished? Impacts on Welfare Participation, Employment, Income.
Poverty, and Family Structure." National Bureau of Economic
Research Working Paper No. 7627, 2000.
Schur, L. "Employment and the Creation of an Active
Citizenry." British Journal of Industrial Relations, 41(4), 2003,
751-71.
Sondheimer, R. M., and D. P. Green. "Using Experiments to
Estimate the Effects of Education on Voter Turnout." American
Journal of Political Science, 54(1), 2010, 174-89.
Soss, J. "Lessons of Welfare: Policy Design. Political
Learning, and Political Action." American Political Science Review,
93(02), 1999, 363-80.
Sugie, N. F. "Chilling Effects: Diminished Political
Participation among Partners of Formerly Incarcerated Men." Social
Problems, 62, 2015, 550-71.
Swartz, T. T., A. Blackstone, C. Uggen, and H. McLaughlin.
"Welfare and Citizenship: The Effects of Government Assistance on
Young Adults' Civic Participation." The Sociological
Quarterly, 50(4), 2009. 633-65.
Taylor, P. "The Growing Electoral Clout of Blacks Is Driven by
Turnout, Not Demographics." Pew Social & Demographic trends,
2013. Accessed April 10, 2016. http://
www.pewsocialtrends.org/files/2013/01/2012_Black_
Voter_Project_revised_1-9.pdf.
U.S. Census Bureau. "Statistical Abstract of the United
States: 2012." Section 7: Elections, 2012. Accessed February 7,
2016 http://www.census.gov/prod/ 2011 pubs/12statab/election.pdf.
U.S. Department of Health and Human Services. "State
Implementation of Major Changes to Welfare Policies, 1992-1998."
Washington. DC: Office of the Assistant Secretary for Planning and
Evaluation. 1999. Accessed Febraury 9, 2016.
https://aspe.hhs.gov/basicreport/state-implementation-major
-changes-welfare-policies-1992-1998.
U.S. Department of Justice. "About the National Voter
Registration Act." 2015. Accessed February 8, 2016. http://
www.justice.gov/crt/about-national-voter-registration-act
U.S. Election Project. "CPS Vote Over-Report and NonResponse
Bias Correction." 2016a. Accessed June 19, 2016.
http://www.electproject.org/home/voter-turnout/ cps-methodology
--. "Voter Turnout National Turnout Rates, 1787-2012."
2016b. Accessed October 22, 2016.
http://www.electproject.org/home/voter-turnout/voterturnout-data
Verba, S., K. L. Schlozman, and H. E. Brady. Voice and Equality:
Civic Voluntarism in American Politics. Cambridge, MA: Harvard
University Press, 1995.
Watson, S. "Does Welfare Conditionality Reduce Democratic
Participation?" Comparative Political Studies, 48(5), 2015, 645-86.
Ziliak, J. P. "Taxes, Transfers, and the Labor Supply of
Single Mothers." Unpublished Working Paper No. 2006, 2006. Accessed
February 5, 2016. http://www.nber.org/~confer/2006/URCf06/ziliak.pdf
(1.) We present the general model with separate estimates for AFDC
waivers and TANF, as some of our models are specified that way. However,
as indicated earlier, our preferred specifications use a single
indicator of whether the state had implemented an AFDC waiver or TANF.
(2.) The legislature data are from several volumes of the
Statistical Abstract of the United States. Composition of State
Legislatures by Political Party Affiliation. Data on the Governor's
party were obtained from the National Governors' Association. Male
registration rates by state and year (from our CPS data) were included
in models predicting women's registration, and male voting rates
were included in models of women's voting.
(3.) We implemented the reduced-form DDD-based analysis rather than
estimating IV models using the welfare reform measures as instruments
for employment, because if welfare reform had any effect on voting
through channels other than employment, welfare reform measures would be
inappropriate IVs. That is, welfare reform would be correlated with the
error term in the structural model linking employment to voting and it
would thus violate the exclusion restriction. The reduced form DDD
estimates, on the other hand, capture any and all channels, reinforcing
or counteracting, through which welfare reform may affect the voting
outcomes. That said, we do later impute and discuss the implied IV
estimate as a reference point for the reduced form estimates.
(4.) In contrast, both Knack (1999) and Brown and Wedeking (2006)
found that election day registration (EDR). another voting-related
policy implemented by some states that was not required by the NVRA but
was consistent with its intent, did increase voter turnout. Only three
states implemented EDR during the rollout of welfare reform--Idaho and
Wyoming in 1994 and New Hampshire in 1996 (National Council of State
Legislatures 2015). Excluding these states from our analyses does not
materially affect any of the estimates presented or referred to in this
article.
(5.) A value of negative I corresponds to the 12-month period
preceding the implementation of welfare reform in the woman's
state, and a value of 0 corresponds to the first 12 months after welfare
reform implementation.
(6.) Note that there could be up to a 2-year lag between the
implementation of welfare reform and a congressional election, and there
could be up to a 4-year lag between the implementation of welfare reform
and a presidential election.
(7.) Some of this difference may be due to unobserved trends and
economic conditions. The conditional DDD estimates on employment
(discussed later), based on Equation (2). are 6.4 percentage points and
10.0 percentage points for groups 2 and 1, respectively (/; value =
0.000). which are consistent with the literature of the effects of
welfare reform on employment.
(8.) The coefficient of the main indicator for Any Welfare Reform
is negative and significant for presidential elections (Model 5 in Table
2), confirming the various unmeasured factors driving the general
downward trend in voting participation among women coinciding with
welfare reform that was evident in the figures (for non-presidential
elections, the trends are generally positive). It is this presence of
confounding trends that motivates our use of various control groups
(women not at risk of being on public assistance) and the DDD framework
to net out such state-specific time-varying unobserved characteristics.
That is. in the absence of plausible control groups and the third
"D." a DD model, which would just include the indicator for
Any Welfare Reform and not the interaction, would lead to biased
estimates of changes in voting behavior. That is, the estimates would,
to a large extent, reflect the secular downward trend in voting among
low-educated women. The fact that the estimate for Any Welfare Reform is
negative and significant when considering presidential years but not
significantly different from zero when considering all elections is not
particularly surprising. and differences in fact are to be expected, as
that estimate (which is not the estimate of interest for our study)
reflects changes in a constellation of factors (including, but not
limited to welfare reform) over different sets of years. The variability
of estimates for this indicator when considering different election
years does not compromise our inference that the estimates of the
interaction term (which are net of all trends coinciding with welfare
reform during the specified time period) were driven almost entirely by
presidential year elections.
(9.) To get these numbers, we divide the coefficients in
specification (5) in Table 2 by the growth in employment relative to the
comparison group. That is .19 is obtained by dividing .019 by .10 and
.36 is obtained by dividing .023 by .064. This is the Wald estimator for
deriving the IV estimate in the case of a binary instrument.
(10.) We further assess the role of employment and other mechanisms
below in Tables 4 and 5.
(11.) Note that all state/year variables drop out of the equation
when state/year dummies are included.
(12.) For ease of exposition, we use the term "relative
increase" to refer to a reduced decrease.
(13.) That is, labor supply, education, and income are themselves
outcomes of welfare reform, and are also potentially correlated with
unobserved factors in Equation (2). This may lead to a form of sample
selection bias (Angrist and Pischke 2009) since, with the additional
mediators, the effect of welfare reform is being estimated within
sub-populations with the same level of the mediators (e.g.. the same
education level or employment status) and sorting into these
sub-populations is not random but likely a function of welfare reform.
Caption: FIGURE 1 Registered to Vote by Years Since Welfare Reform
in Target Group Compared to Two Different Comparison Groups
Caption: FIGURE 2 Voted in Election by Years Since Welfare Reform
in Target Group Compared to Two Different Comparison Groups
TABLE 1
Annual Rates of Voting Registration and Voting, 1990-2004
Target Group Comparison Group 1
Unmarried Mothers Unmarried Women with No
with High School Children and High School
Education or Less Education or Less
Registered Voted Registered Voted
1990 0.49 0.26 0.53 0.33
1992 0.56 0.42 0.60 0.49
1994 0.48 0.23 0.52 0.28
1996 0.57 0.35 0.59 0.41
1998 0.53 0.23 0.53 0.26
2000 0.59 0.39 0.58 0.41
2002 0.55 0.24 0.54 0.27
2004 0.63 0.46 0.62 0.48
Comparison Group 2
Married Mothers
with High School
Education or Less
Registered Voted
1990 0.62 0.41
1992 0.68 0.60
1994 0.61 0.38
1996 0.66 0.50
1998 0.65 0.36
2000 0.69 0.55
2002 0.65 0.37
2004 0.71 0.59
Note: All groups consist of women ages 21-49 who are
U.S. citizens.
TABLE 2
Estimated Effects of Welfare Reform on Women's Registration
and Voting, Using (a) Comparison Group 1 and (b) Comparison Group 2
(a) Target Group:
Unmarried Mothers
with High School
Education or Less Registered Voted
Comparison Group 1: Any Any Any
Unmarried Women with Election Election Election
No Children and High Year Year Year
School Education (1) (2) (3)
or Less
AFDC Waiver 0.013 0.023
(0.022) (0.014)
TANF -0.016 -0.016
(0.018) (0.016)
AFDC Waiver* -0.003 -0.024
Target (0.016) (0.015)
TANF * Target 0.035 *** 0.027***
(0.010) (0.009)
Any Welfare -0.008
Reform (0.014)
Any Welfare 0.029 ***
Reform * Target (0.008)
N 32,234 32,234 32,790
(a) Target Group:
Unmarried Mothers
with High School
Education or Less Voted
Comparison Group 1: Any Presidential Nonpresidential
Unmarried Women with Election Election Election
No Children and High Year Year Year
School Education (4) (5) (6)
or Less
AFDC Waiver
TANF
AFDC Waiver*
Target
TANF * Target
Any Welfare -0.005 -0.049 ** 0.037
Reform (0.012) (0.019) (0.025)
Any Welfare 0.019 ** 0.027 ** 0.010
Reform * Target (0.008) (0.011) (0.011)
N 32,790 15,669 17,121
(b) Target Group:
Unmarried Mothers
with High School
Education or Less Registered Voted
Comparison Group 2: Any Any Any
Married Mothers with Election Election Election
High School Education Year Year Year
or Less (1) (2) (3)
AFDC Waiver -0.012 -0.018
(0.012) (0.012)
TANF 0.001 < 0.001
(0.015) (0.01)
AFDC Waiver * Target 0.042 ** 0.023
(0.016) (0.016)
TANF * Target 0.024 ** 0.023 **
(0.010) (0.010)
Any Welfare Reform -0.006
(0.011)
Any Welfare 0.027 ***
Reform * Target (0.01)
N 54,470 54,470 55.048
(b) Target Group:
Unmarried Mothers
with High School
Education or Less Voted
Comparison Group 2: Any Presidential Nonpresidential
Married Mothers with Election Election Election
High School Education Year Year Year
or Less (4) (5) (6)
AFDC Waiver
TANF
AFDC Waiver * Target
TANF * Target
Any Welfare Reform -0.013 -0.046 *** 0.026 *
(0.010) (0.015) (0.015)
Any Welfare 0.023 *** 0.039 *** 0.015
Reform * Target (0.010) (0.013) (0.010)
N 55,048 25,796 29,252
Notes: AFDC = Aid to Families with Dependent Children. TANF =
Temporary Assistance to Needy Families. All groups consist of women
ages 21-49 who are U.S. citizens. Coefficients from ordinary least
squares models are reported. State-clustered standard errors are
reported in parentheses. All models control for state and year
fixed effects and include the following covariates: age,
age-squared, race (black, other race), Hispanic, marital history
(widowed/divorced/separated, never married), number of children
<18, number of household members age 18+ years, MSA residence,
current and 1-year lagged state unemployment rate, current and 1
-year lagged state personal income per capita, log female
population, state poverty rate, state minimum wage, 1- and 2-year
lagged welfare caseloads, % state legislature Democrat/Republican,
Republican governor. Democrat governor, and registration/voting
rates of males. * p [less than or equal to]. 10; ** p [less than or
equal to]. 05; *** p [less than or equal to] .01.
TABLE 3
Estimated Effects of Welfare Reform on Voting in Presidential
Elections, Accounting for Differential Trends, Using
(a) Comparison Group 1 and (b) Comparison Group 2
(a) (1) (2) (3)
Any Welfare -0.049 ** -0.053 ** -0.053 **
Reform (0.019) (0.020) (0.023)
Welfare Reform* 0.027 ** 0.027 ** 0.026 **
Target (0.011) (0.011) (0.011)
Motor Voter law No Yes Yes
State linear No No Yes
trend
State * Year No No No
indicators
Target*Linear No No No
pre-policy trend
N 15,669 15,669 15,669
(a) (4) (5)
Any Welfare
Reform
Welfare Reform* 0.023 ** 0.020 *
Target (0.011) (0.011)
Motor Voter law No No
State linear No No
trend
State * Year Yes Yes
indicators
Target*Linear No Yes
pre-policy trend
N 20,682 20,682
(b) (1) (2) (3)
Any Welfare -0.046 *** -0.049 *** -0.052 ***
Reform (0.015) (0.014) (0.015)
Welfare Reform 0.039 *** 0.039 *** 0.039 ***
* Target (0.013) (0.013) (0.013)
Motor Voter law No Yes Yes
State linear No No Yes
trend
State * Year No No No
indicators
Target* Linear No No No
pre-policy trend
N 25,796 25,796 25.796
(b) (4) (5)
Any Welfare
Reform
Welfare Reform 0.025 ** 0.051 ***
* Target (0.011) (0.016)
Motor Voter law No No
State linear No No
trend
State * Year Yes Yes
indicators
Target* Linear No Yes
pre-policy trend
N 35,482 35,482
Notes: See Table 2. Target group consists of unmarried mothers with
a high school education or less, (a) Comparison group consists of
unmarried women with no children and a high school education or
less (comparison group 1). (b) Comparison group consists of
married mothers with a high school education or less (comparison
group 2). In each panel. Model 1 is identical to the corresponding
Model 5 in Table 2 and Models 2 through 5 add covariates to that
model as indicated. The coefficient for Any Welfare Reform is not
presented in Models 4 and 5 since it is perfectly collinear with
the State*Year fixed effects.
TABLE 4
Estimated Effects of Welfare Reform on Voting in
Presidential Elections, Heterogeneous Effects
Along Relevant State Margins
Comparison Group 1: Unmarried
Women with No Children and a
High School Education or Less
(1) (2) (3) (4)
State State Work Work
Legislature Legislature Incentives Incentives
Sample Republican Democratic Strong or Weak
Mixed
Any Welfare -0.057 * -0.062 * -0.050 ** 0.087
Reform (0.031) (0.031) (0.022) (0.051)
Welfare 0.015 0.034 ** 0.033 *** -0.019
Reform*Target (0.018) (0.015) (0.011) (0.055)
N 5.669 9.144 13,852 1.817
Comparison Group 2: Married
Mothers with a High School
Education or Less
(5) (6) (7) (8)
State State Work Work
Legislature Legislature Incentives Incentives
Sample Republican Democratic Strong or Weak
Mixed
Any Welfare -0.015 -0.089 *** -0.056 *** 0.098 **
Reform (0.018) (0.022) (0.016) (0.034)
Welfare 0.005 0.047 *** 0.051 *** -0.051
Reform*Target (0.035) (0.016) (0.013) (0.042)
N 9.448 14.926 22,591 3.205
Notes: See Table 2. Target group consists of unmarried mothers with
a high school education or less. Specification for all models in
this table correspond to Mode! 5 in Table 2. Models 1 and 5 include
state/year observations in which the percentage of state
legislators that were Republican was greater than 50%. Models 2 and
6 include state/year observations in which the percentage of state
legislators that were Democratic was greater than 50%. For
Nebraska, we use the corresponding percentages of national
legislators from that state, since no Nebraska state legislators
have official party affiliation (i.e.. they are all nonpartisan).
TABLE 5
Estimated Effects of Welfare Reform on Voting in Presidential
Elections and Potential Mediating Effects of Employment,
Income, and Education
Comparison Group 1: Unmarried
Women with No Children and a
High School Education or Less
(1) (2) (3) (4)
Any Welfare -0.047 ** -0.046 ** -0.049 *** -0.049 ***
Reform (0.019) (0.019) (0.018) (0.018)
Welfare Reform 0.016 0.015 0.020 * 0.017
* Target (0.011) (0.011) (0.012) (0.012)
Employed 0.138 ***
(0.011)
Weekly hours 0.052 * 0.058 * 0.048
< 10 (0.028) (0.030) (0.030)
Weekly hours 0.085 *** 0.070 *** 0.060 ***
10-20 (0.019) (0.019) (0.019)
Weekly hours 0.108 *** 0.075 *** 0.056 ***
20-30 (0.016) (0.015) (0.016)
Weekly hours 0.151 *** 0.010 *** 0.078 ***
30-40 (0.011) (0.011) (0.011)
Weekly hours 0.139 *** 0.078 *** 0.057 ***
> 40 (0.018) (0.017) (0.017)
Ln family 0.075 *** 0.062 ***
income (0.005) (0.005)
High school 0.129 ***
graduate (0.012)
N 15,669 15.669 15,669 15,669
Comparison Group 2: Married
Mothers with a High School
Education or Less
(5) (6) (7) (8)
Any Welfare -0.044 *** -0.044 *** -0.036 ** -0.037 **
Reform (0.015) (0.015) (0.016) (0.016)
Welfare Reform 0.033 ** 0.033 ** 0.023 * 0.018
* Target (0.013) (0.013) (0.013) (0.013)
Employed 0.085 ***
(0.008)
Weekly hours 0.104 *** 0.095 *** 0.085 ***
< 10 (0.022) (0.022) (0.023)
Weekly hours 0.100 *** 0.077 *** 0.063 ***
10-20 (0.01 1) (0.011) (0.011)
Weekly hours 0.078 *** 0.052 *** 0.039 ***
20-30 (0.011) (0.011) (0.010)
Weekly hours 0.091 *** 0.045 *** 0.030 ***
30-40 (0.009) (0.008) (0.008)
Weekly hours 0.075 *** 0.027 ** 0.012
> 40 (0.013) (0.012) (0.012)
Ln family 0.102 *** 0.083 ***
income (0.005)
High school 0.157 ***
graduate (0.010)
N 25,796 25,796 23,740 23,740
Notes: See Table 2. Target group consists of unmarried mothers
with a high school education or less. Specification for all
models in this table correspond to Model 5 in Table 2, with
the addition of mediators indicated.
TABLE 6
Estimated Effects of Welfare Reform on Voting in Presidential
Elections, Lagged Effects and Placebo Checks
Comparison Group 1 Unmarried Women with No
Children and a High School Education or Less
All States States that All States
Implemented
Welfare
Reform
post-August
1996
Sample Elections: Elections: Elections:
1992, 1996, 2000, 1992 & 1996 1992 & 1996
2004
(1) (2) (3) (4)
Any Welfare 0.027 *
Reform-lagged (0.015)
6-months *
Target
Any Welfare 0.034 **
Reform-lagged (0.016)
12-months *
Target
Any Welfare -0.011
Reform * (0.041)
Target
Post-1996 -0.001
Welfare Reform (0.021)
implementation *
Target
N 15,669 15,669 3,469 8,688
Comparison Group 2 Married Mothers with a
High School Education or Less
All States States that All States
Implemented
Welfare
Reform
post-August
1996
Sample Elections: Elections: Elections:
1992, 1996, 2000, 2004 1992 & 1996 1992 & 1996
(5) (6) (7) (8)
Any Welfare 0.039 ***
Reform-lagged (0.014)
6-months *
Target
Any Welfare 0.040 ***
Reform-lagged (0.015)
12-months *
Target
Any Welfare -0.032
Reform * (0.037)
Target
Post-1996 -0.004
Welfare Reform (0.019)
implementation *
Target
N 25.796 25,796 5,921 15.094
Notes: Coefficients from ordinary least squares models are
reported. State-clustered standard errors are reported in
parentheses. All models control for state and year fixed effects
and include the following covariates: age. age-squared, race
(black, other race), Hispanic, marital history
(widowed/divorced/separated, never married), number of children
<18. number of household members age 18+ years. MSA residence,
current and 1-year lagged state unemployment rate, current and
1-year lagged state personal income per capita, log female
population, state poverty rate, state minimum wage, 1- and 2-year
lagged welfare caseloads, % state legislature Democrat/Republican,
Republican governor. Democrat governor, and registration/voting
rates of males. * p [less than or equal to]. 10; ** p [less than or
equal to] .05; *** p [less than or equal to] .01.
COPYRIGHT 2017 Western Economic Association International
No portion of this article can be reproduced without the express written permission from the copyright holder.
Copyright 2017 Gale, Cengage Learning. All rights reserved.
