Preprints and early-stage research may not have been peer reviewed yet.
Preprint

Entrepreneurship as a Response to Labor Market Discrimination for Formerly Incarcerated People

Abstract and Figures

This paper examines entrepreneurship as a way to overcome labor market discrimination. Specifically, we examine entrepreneurship as a career choice for formerly incarcerated individuals, a group of individuals who face substantial discrimination in the labor market. Using the United States National Longitudinal Survey of Youth 1997 data, we find that formerly incarcerated people are more likely to become entrepreneurs compared to individuals without a criminal record. We take advantage of an exogenous state and county level policy shock "Ban-the-Box" in the United States to further disentangle the underlying mechanism of how labor market discrimination affects formerly incarcerated individuals in their entrepreneurial choices. The findings suggest that entrepreneurship is a viable alternative career choice for formerly incarcerated people, yielding both higher income and lower recidivism rates. In addition to reporting robustness checks and addressing alternative explanations, we discuss theoretical, empirical, and policy implications.
Content may be subject to copyright.
1
Entrepreneurship as a Response to Labor Market Discrimination
for Formerly Incarcerated People
1
Kylie Jiwon Hwang
2
Damon J. Phillips
Columbia Business School
Last Updated March 6th, 2020
ABSTRACT
This paper examines entrepreneurship as a way to overcome labor market discrimination.
Specifically, we examine entrepreneurship as a career choice for formerly incarcerated individuals,
a group of individuals who face substantial discrimination in the labor market. Using the United
States National Longitudinal Survey of Youth 1997 data, we find that formerly incarcerated people
are more likely to become entrepreneurs compared to individuals without a criminal record. We
take advantage of an exogenous state and county level policy shock “Ban-the-Box” in the United
States to further disentangle the underlying mechanism of how labor market discrimination affects
formerly incarcerated individuals in their entrepreneurial choices. The findings suggest that
entrepreneurship is a viable alternative career choice for formerly incarcerated people, yielding
both higher income and lower recidivism rates. In addition to reporting robustness checks and
addressing alternative explanations, we discuss theoretical, empirical, and policy implications.
1
This research was generously supported by the Ewing Marion Kauffman Foundation, the Wheeler Institute for
Business and Development, and Columbia Business School. We also thank Bruce Western, Bruce Kogut, and Dan
Wang for their comments and suggestions. Previous drafts of this manuscript have greatly benefited from
participants of the seminar at American Sociological Association Annual Conference, Academy of Management
Strategy Doctoral Consortium, Strategic Management Society Annual Conference, Wharton People and
Organizations Conference, East Coast Doctoral Conference, and Transatlantic Doctoral Conference. The contents of
this publication are solely the responsibility of the authors.
2
Corresponding Author: Kylie Jiwon Hwang, 3022 Broadway Uris Hall 7 East, NY, NY, 10027,
jkh2134@columbia.edu
2
INTRODUCTION
The United States is the world's leader in incarceration with 2.2 million people in prisons
and jails as of 2016, marking a 600% increase of the penal population over the last 40 years
(Bureau of Justice Statistics 2018). This increase has also led to more than 600,000 people per
year reentering society from incarceration (Bureau of Justice Statistics 2019). Successful reentry
is the exception however, with over two-thirds of formerly incarcerated people rearrested within
three years of reentry (Alper et al. 2018).
One of the key factors thwarting successful reentry is the severe discrimination in the
labor market faced by formerly incarcerated people. As of 2018, formerly incarcerated people
suffer from an unemployment rate of 27.3 percent compared to 5.8 percent for the general public
(Prison Policy Initiative 2018). Formerly incarcerated individuals especially those who are
African American are not only less likely to be hired by employers, but those who are hired
earn lower wages and experience less wage mobility (Petit and Western 2004, Western and Petit
2005, Western and Beckett 1999, Pager 2003, Western 2002). The evidence by scholars,
combined with first-person accounts, has given rise to a stylized fact which is capstured by an
unemployment-recidivism narrative: formerly incarcerated individuals face labor discrimination
that leads to unemployment or underemployment, and as a consequence their likelihood of
returning to prison substantially increases. In informing this narrative, scholarly research has
also fueled our understanding that until labor market discrimination is substantially reduced, the
U.S. will continue to be faced with poor reentry outcomes.
3
In this paper we offer a modification to this narrative by drawing attention to formerly
incarcerated people who become entrepreneurs.
3
Here, entrepreneurship is examined as a
response to labor market discrimination, where one not only overcomes this discrimination by
earning an income through starting a new business, but in doing so also increases the overall
chance of successful reentry. Through this examination we seek to complement long-run efforts
to reduce labor market discrimination with an evaluation of entrepreneurship as an alternative
trajectory. We also contribute to the important yet challenging question of how stigmatized
individuals can overcome persistent labor market discrimination without solely depending on
employers’ shifting priorities and public policy changes. Thus, while we maintain the existing
narrative of unemployment and recidivism for formerly incarcerated individuals, this paper
broadens the focus on employment to include entrepreneurship.
Specifically, we investigate whether formerly incarcerated individuals are more likely to
become entrepreneurs compared to those without criminal records. Entrepreneurial entry
decisions are made as a function of a worker’s calculus of whether to pursue paid employment or
launch a new venture (Sørensen and Sharkey, 2014, Kacperczyk, 2012, Yang and Kacperczyk
2018, Hellmann, 2007). As labor market discrimination decreases opportunities for paid
employment (Pager, Western, and Bonikowski, 2009, Pager 2003), such lack of employment
opportunities may push individuals into starting their own businesses (Light 1972). Thus,
compared to individuals who have never been to prison, formerly incarcerated individuals may
be more likely to seek entrepreneurship as a route to secure work and income.
3
Following prior studies on entrepreneurship, we define entrepreneurship conceptually as launching a
new business (Evans 1989, Sørensen and Sharkey 2014, Yang and Aldrich 2014) and operationally as
self-employment (Greenfield et al 1979, Fairlie 1999, Aldrich 1990, Hegde and Tumilson 2018). For
robustness, we also examine incorporated self-employment and self-employment with employees.
4
Importantly, we unpack an underlying causal mechanism of why formerly incarcerated
individuals become entrepreneurs. The decision to become an entrepreneur is driven by many
characteristics correlated with incarceration such as education, poverty, and risk-preference. We
parse out the diverse underlying mechanisms driving entrepreneurial entry of formerly
incarcerated individuals, by taking advantage of the staggered implementation of an exogenous
state and county-level policy shock in the United States, the “Ban-the-Box” policy. The “Ban-
the-Box” policy, which has been implemented in 35 states and over 150 counties as of 2019, bars
employers from checking criminal backgrounds until later in the hiring process (Avery and
Hernandez 2019). The policy aims to mitigate discrimination in the employment process and
increase employment opportunities for formerly incarcerated individuals. We use the
introduction of this policy to unpack whether entrepreneurial entry of formerly incarcerated
individuals is driven by the lack of alternative employment opportunities in the labor market.
We find that formerly incarcerated individuals are more likely to become entrepreneurs
than those who have never been incarcerated. More critical to our thesis however, the likelihood
of entrepreneurship for those who have been to prison varies by whether or not they were
residing in a locale where the “Ban-the-Box” policy was implemented. In particular, formerly
incarcerated individuals are less likely to enter into entrepreneurship when “Ban-the-Box” policy
is implemented in the state or county of residence, indicating that a key reason that individuals
with criminal records pursue entrepreneurship is to overcome labor market discrimination. We
also find this effect is strongest for African American formerly incarcerated individuals, who
experience the greatest labor market discrimination before the implementation of “Ban-the-Box”.
These findings help us gain insight into the underlying mechanism of how the change in the
severity of labor market discrimination affects entrepreneurship rates.
5
We further investigate whether entrepreneurship can be a valid route to overcome labor
market discrimination for formerly incarcerated individuals by examining their income and
recidivism rates. The negative relationship between incarceration and income is commonly
attributed to the stigma of a criminal record, as employers undervalue formerly incarcerated
individuals as employees and impose income penalties (Holzer, Raphael, and Stoll 2003,
Western and Beckett 1999, Western 2002). We argue that entrepreneurship will mitigate the
income inequality that formerly incarcerated individuals experience in paid employment, as
entrepreneurship provides an opportunity to earn based on one’s productivity rather than an
employer’s valuation (Hegde and Tumilson 2018). We find supportive evidence that
entrepreneurship provides formerly incarcerated individuals with higher income compared to
their earnings from employment, helping individuals with criminal records to close the income
gap.
Concerning recidivism, we argue that entrepreneurship offers both greater economic
incentives and social incentives, such as responsibility and work satisfaction, that desist formerly
incarcerated individuals from reengaging in criminal activity. We find evidence that
entrepreneurship decreases the likelihood of returning to prison beyond the effect of paid
employment, supporting the view of entrepreneurship as a way for formerly incarcerated
individuals to successfully reenter the society. These findings on income and recidivism suggest
that entrepreneurship may not only help discriminated individuals improve their economic well-
being, but also to improve social integration.
We test the propositions of this paper using data drawn from the United States National
Longitudinal Survey of Youth (NLSY97) from 1997 to 2015. We combine this dataset with data
we collected from the National Employment Law Project on the implementation of “Ban-the-
6
Box” policies in the U.S. at the state and county level (Avery 2019). We address endogeneity
concerns by taking advantage of a quasi-experimental design, which uses the exogenous state
and county level policy shock, “Ban-the-Box.” For further robustness, we also conduct three
different types of matching procedures to restrict the analysis to a comparison between formerly
incarcerated individuals and non-formerly incarcerated individuals with similar characteristics.
Our results are consistent throughout.
Our paper provides implications for theory, empirical work, and policy. We offer a
modification to the current narrative on unemployment and recidivism for formerly incarcerated
individuals by introducing entrepreneurship as an alternative route to successful reentry. We find
that entrepreneurship provides an opportunity for formerly incarcerated individuals to not only
find work and decrease economic inequality, but also to decrease recidivism rates. We contribute
to empirical work by using a quasi-natural experiment design to establish causality and providing
stronger evidence for the underlying mechanism between incarceration and entrepreneurship.
Lastly, we provide policy implications by highlighting entrepreneurship as a route for formerly
incarcerated individuals to overcome labor market discrimination without depending on
policymakers or employers.
THEORY
Incarceration, Labor Market Discrimination, and Recidivism
In the past two decades, a growing number of studies have investigated the effect of
incarceration on labor market outcomes. Scholars have found consistent evidence that contact
7
with the criminal justice system leads to reduction in economic opportunities (e.g. Petit and
Western 2004, Pager 2003). Research has documented that individuals coming out of prisons or
jails experience significantly lower employment, with the formal employment rate in the first
year after release ranging from 40% to 64% (Western and Beckett 1999, Pager 2003, Freeman
1991, Waldfogel 1994, Petit and Lyons 2007). Scholars have further found that even when
formerly incarcerated individuals do find employment, they are often relegated to jobs of lower
quality (e.g. Western 2002). Formerly incarcerated individuals experience earnings penalties of
10% to 30% after release from prison, relative to their earnings prior to incarceration (Waldfogel
1994, Western 2002). Harding, Morenoff, and Wyse (2019) have found that formerly
incarcerated individuals are often sorted into jobs that are characterized not only by lower wages,
but also greater turnover, poor working conditions, and irregular work schedules. Other studies
have emphasized longer term adverse effects of incarceration on economic opportunities, such as
penalties in wage growth, upward job mobility, and job stability (Western 2002, Nagin and
Waldfogel 1995, Bushway 1998).
Many studies on labor market outcomes have also observed racial disparities in the
negative impact of incarceration on employment outcomes. Racial minorities, especially African
Americans, are not only more likely to be incarcerated but also face greater penalties for
incarceration (e.g. Pager 2003, Western 2002, Lyons and Petit 2011). Pager (2003) shows that
the adverse effect of a criminal record is 40% larger for blacks than for similar white candidates
looking for employment. Lyons and Petit (2011) document greater wage penalties for black
formerly incarcerated individuals 10% lower earnings and 21% slower wage growth
compared to similar white formerly incarcerated individuals.
8
One of the main reasons for such negative employment outcomes for formerly
incarcerated people is discrimination based on the stigma associated with a criminal record (e.g.
Pager 2003). Employers consider a criminal record as a “negative credential” signaling low
worker quality (Grogger 1992), untrustworthiness (Waldfogel 1994), and lack of honesty (Lott
1992), and thus are less likely to hire formerly incarcerated individuals than comparable job
applicants without a criminal record (Western 2002, Uggen et al 2014, Holzer, Raphael, and
Stoll 2003). Particularly, the stigma of incarceration is intensified toward African Americans,
where employers are more likely to discriminate when race and incarceration are compounded
(Pager 2003). A series of recent audit experiments have found that employers discriminate based
on criminal records, reducing the likelihood of a “callback” by 50% for white applicants with a
record and 60% for black applicants with a record (Pager 2003, 2007, Pager, Western, and
Bonikowski 2009, Uggen et al. 2014). While there are additional individual level mechanisms
through which incarceration may negatively affect subsequent employment, such as erosion of
human and social capital (Kling 1999, Hagan 1993, Harding, Morenoff, and Wyse 2019), the
institutional effect of employer discrimination remains the main driver of adverse employment
opportunities.
Researchers have underlined the significance of employment barriers associated with
incarceration because employment is one of the strongest predictors of successful reentry.
Studies show evidence that unemployment or job instability following release from prison
increases the chances of reoffending (e.g. Shover 1996, Sampson and Laub 1997). Employment
primarily provides formerly incarcerated individuals with the economic means for basic needs,
reducing the material motivations for crime and increasing the costs of recidivism (Sampson and
Laub 1997). Yet, beyond offering economic opportunities, employment is also a key foundation
9
for social reintegration and commitment (Sampson and Laub 1997, Harding, Morenoff, and
Wyse 2019). Uggen (1999) elaborates that employment, diminishes recidivism rates even
controlling for wages. Commitment to work and work satisfaction itself is a positive transition in
the life course of formerly incarcerated individuals, reducing their motivation to commit another
crime (Uggen 1999, Uggen 2000, Sampson and Laub 1997, Uggen and Staff 2001).
The adverse reentry consequences of unemployment or underemployment associated
with labor market barriers for formerly incarcerated individuals has led to a stylized fact that is
often represented by an unemployment-recidivism narrative. In this narrative, formerly
incarcerated individuals face discrimination from employers that leads to unemployment or
underemployment, and as a consequence their likelihood of successful reentry substantially
decreases. The narrative and the evidence it represented has helped to focus researchers and
policymakers on improving the employment prospects for formerly incarcerated people. Through
efforts such as “Ban-the-Box”and Fair Chance Hiring, policymakers and employers have
induced slight reduction in employment barriers for formerly incarcerated individuals.
Incarceration & Entrepreneurship
Despite this focus on improving employment prospects, labor market discrimination
persists for formerly incarcerated individuals. Their unemployment rate remains 5 times higher
than individuals without a criminal record, and employment for formerly incarcerated people
continues to be limited to short-term, unstable, and lower paying jobs (Sugie 2018). Meanwhile,
while labor markets discrimination persists, there is increasing anecdotal evidence of formerly
incarcerated individuals carving out their own careers by becoming entrepreneurs. Yet research
on understanding entrepreneurship as a reentry route for formerly incarcerated individuals has
10
been sparse in both sociology and economic literature. This paper, to our knowledge, is one of
the first to examine how entrepreneurship can be an alternative labor market choice for formerly
incarcerated people.
In particular, we offer a modification to the unemployment-recidivism narrative by
considering entrepreneurship as an important alternative labor market choice for formerly
incarcerated individuals. We argue that formerly incarcerated individuals may choose to launch
their own business in response to barriers to employment. We suggest that entrepreneurship may
provide formerly incarcerated individuals with not only economic opportunities but help their
successful reentry into society. Thus, while gainful employment remains a more common path to
successful reentry, we introduce entrepreneurship as an alternative route that formerly
incarcerated individuals may pursue and benefit from when employment opportunities are
scarce.
There are four primary questions we seek to address in this paper. First, we examine the
extent to which formerly incarcerated people engage in entrepreneurship compared to people
who have not been to prison. Second, we investigate the underlying mechanism as to why people
who have been to prison choose to become entrepreneurs. Third, we assess whether
entrepreneurship helps formerly incarcerated individuals overcome economic inequality, by
examining earnings. Lastly, we address the question of how entrepreneurship affects successful
reentry by investigating the recidivism rates of formerly incarcerated entrepreneurs.
Incarceration & Entrepreneurial Entry
The first question this study investigates is the extent to which formerly incarcerated
individuals take part in entrepreneurship, compared to individuals who have never been to
11
prison. The decision to become an entrepreneur is a function of the set of opportunities available
in the labor market (Burton, Sørensen, and Beckman 2002, Sørensen and Sharkey 2014). Müller
and Arum (2004) note that individuals “actively decide” to take part in entrepreneurship after
“considering the perceived relative costs and benefits attached to distinct paths” in the labor
market. Career opportunities in the labor market shape the decision to transition into
entrepreneurship (Amit, Muller, Cockburn 1995), making entrepreneurial entry more appealing
when alternative labor market choices are relatively less compelling (Sørensen and Sharkey
2014, Kacperczyk and Marx 2016, Yang and Kacperczyk 2018, Castellaneta, Conti, and
Kacperczyk, Hellmann 2007). As the labor market opportunity structure has a profound
influence on an individual’s transition to entrepreneurship, discrimination in the labor market
increases the likelihood of disadvantaged individuals to become entrepreneurs.
While entrepreneurship entails a risk of failure and uncertainty, labor market
discrimination can mean that the expected returns from employment opportunities for formerly
incarcerated individuals are lower than the returns from pursuing entrepreneurship. This implies
that, compared to similar individuals without a criminal record, formerly incarcerated individuals
are more likely to become entrepreneurs because they face lower expected returns from
employment and thus lower opportunity costs of exiting employment and pursuing
entrepreneurship. Accordingly, we formally test the extent to which formerly incarcerated
individuals choose entrepreneurship as a labor market outcome.
Explanatory Mechanism of Incarceration & Entrepreneurship
Second, we unpack the underlying mechanism of why formerly incarcerated individuals
enter entrepreneurship, compared to similar individuals who have not been incarcerated. This
12
focus allows us to parse out the effect of labor market discrimination from the individual level
effect that drives formerly incarcerated individuals to transition into entrepreneurship.
We argue that the main underlying mechanism driving formerly incarcerated individuals
into entrepreneurship is labor market discrimination. As shown by many scholars, the stigma
associated with the mark of a criminal record negatively impacts employment opportunities for
formerly incarcerated individuals, by increasing unemployment, work-related insecurity, and
income penalties (Holzer, Raphael, and Stoll 2006; Pager 2007; Stoll and Bushway 2008; Pager,
Western, and Bonikowski 2009, Sugie 2018). Such discrimination from employers decreases the
opportunity cost of leaving employment and increases the expected returns from
entrepreneurship relative to employment, and may push formerly incarcerated individuals to seek
entrepreneurship as an alternative choice.
At the same time, individual level mechanisms such as preferences or human capital may
make entrepreneurship more appealing than employment for formerly incarcerated individuals.
Scholars have found suggestive evidence that individuals who take part in criminal activity are
similar to entrepreneurs, in terms of low risk aversion and preference for autonomy (e.g.
Gottschalk 2009, Lockwood et al 2006, Rieple 1998). Other studies also find that formerly
incarcerated individuals possess human capital more fit for entrepreneurship than employment,
such as higher entrepreneurial ability and lower employment-related ability and skills (Fairlie
2002, Sonfield, Lussier, and Barbato 2001).
While both labor market discrimination and individual level mechanisms may jointly
drive the decision of formerly incarcerated individuals to transition into entrepreneurship, our
focus is on understanding the mechanisms associated with labor market discrimination while
using a research design to control for individual level mechanisms. Thus we examine whether
13
labor market discrimination drives formerly incarcerated individuals to choose entrepreneurship
instead of employment, notwithstanding individual level mechanisms such as preferences or
human capital. We are able to disentangle the two mechanisms by utilizing an exogenous change
to one of the two drivers - labor market discrimination for formerly incarcerated individuals. As
we argue that entrepreneurship is a response to labor market discrimination for formerly
incarcerated individuals, we expect to find that the exogenous decrease of labor market
discrimination for formerly incarcerated individuals will subsequently diminish their likelihood
to enter into entrepreneurship. In particular, we expect to see the greatest reduction of
entrepreneurial transitions for formerly incarcerated individuals who face the most labor market
discrimination before the exogenous change African American (black) formerly incarcerated
individuals. Specifically, we expect to find that an exogenous decrease of labor market
discrimination has a larger negative impact on the entrepreneurial entry rates of black formerly
incarcerated individuals. By using the exogenous change of labor market discrimination, this
study effectively isolates and verifies the effect of labor market discrimination on entrepreneurial
transitions of formerly incarcerated people, beyond any individual effects.
Entrepreneurship & Income
Third, this study investigates whether formerly incarcerated individuals are able to
overcome economic barriers through entrepreneurship, by examining earnings. Formerly
incarcerated individuals experience considerable income penalties in employment because
employers discriminate based on their criminal record (Western 2002). In other words,
independent of the underlying ability of formerly incarcerated individuals, employers are more
likely to undervalue their ability based on the stigma associated with a criminal record such as
low worker quality (Grogger 1992), untrustworthiness (Waldfogel 1994), and lack of honesty
14
(Lott 1992). Yet, unlike employees, who receive their earnings based on the employer’s
valuation of potential ability, entrepreneurs are residual claimants of their own ability rather than
noisy perceptions of it (Hegde and Tumilson 2018). Thus, employees that are undervalued by
employers are more likely to increase their earnings by exiting employment and engaging in
entrepreneurship.
These arguments suggest that formerly incarcerated people will be able to increase their
earnings as entrepreneurs compared to their earnings as employees, as they no longer rely on the
discriminatory perceptions of employers. While founding a new business entails risk of failure
and higher variance in earnings, the elimination of discrimination and stigma that formerly
incarcerated individuals face from employers will, on average, improve their earnings from
entrepreneurship compared to that from employment. Thus, we expect entrepreneurship to
increase earnings, compared to their employment income, for formerly incarcerated individuals.
We also expect to find that the income gap between individuals with and without a criminal
record will be smaller for entrepreneurs, compared to employees. By removing the barriers to
competitive earnings, entrepreneurship will provide formerly incarcerated individuals the
opportunity to mitigate economic inequality.
Entrepreneurship & Recidivism
Finally, the fourth objective of this study is to assess how entrepreneurship impacts
successful reentry, by examining recidivism rates. Employment, or securing legitimate work, is
one of the strongest predictors of desistance to crime (Glaser 1969, Farrington et al. 1986,
Trasler 1979, Shover 1996, Sampson and Laub 1993, Uggen 2000, Bushway and Reuter 2002).
Employment not only provides immediate financial support but also increases future expected
15
earnings, significantly increasing the opportunity cost of criminal behavior and consequentially
reducing the likelihood of reoffending (Pezzin 1995). Thus, we expect that entrepreneurship,
which we predict to yield greater economic opportunities than employment for formerly
incarcerated individuals, will further reduce the likelihood of recidivism.
Moreover, entrepreneurship may also provide formerly incarcerated individuals social
and psychological incentives to refrain from further criminal activities. Entrepreneurship entails
being responsible for your business and often times the livelihood of fellow employees,
increasing the sense of responsibility and commitment for formerly incarcerated entrepreneurs.
Also, anecdotal evidence from formerly incarcerated entrepreneurs suggest that entrepreneurship
enhances work satisfaction, self-esteem, and commitment.
4
Studies have found that commitment
to work and work satisfaction diminishes the likelihood of recidivism, beyond the financial
aspect of work itself (Uggen 1999). Thus, such social and psychological incentives of
entrepreneurship may further help individuals to stay out of prison. While it is beyond the scope
of this paper to fully parse out these mechanisms, we expect entrepreneurship to decrease
recidivism for formerly incarcerated individuals beyond paid employment, supporting the view
of entrepreneurship as a way for formerly incarcerated individuals to successfully reenter
society.
By investigating the four research questions, this paper seeks to present a modification to
the narrative of unemployment and recidivism. We complement the current narrative focused on
employment by introducing entrepreneurship as an alternative route for formerly incarcerated
individuals to overcome both economic and social barriers to successful reentry. In the following
4
Defy Ventures: https://defyventures.org/blog/entrepreneurship-as-a-tool-for-social-change-reflecting-on-my-time-
at-defy-ventu
16
sections, we describe the data and methods, and show our empirical analyses that allows us to
address the above research questions.
DATA AND MEASURES
In order to investigate our research question, we merge data from the 1997 cohort of the
National Longitudinal Survey of Youth (NLSY 97) with hand-coded data on “Ban-the-Box”
policy changes for all states and counties of the United States. The NLSY97 follows the lives of
a representative cohort of 8,984 men and women, who were 12-18 years old when first surveyed
in 1997. These individuals were interviewed annually from 1997 through 2011 and biennially
thereafter. The restricted NLSY97 Geocode data provide identifying information about yearly
state and county level residence by survey respondent, thus allowing us to utilize county and
state-level policy shocks to address the causal mechanism. As shown by prior research (e.g.
Western 2002, Western and Petit 2010) the NLSY data is suitable for research on incarceration
because it reports detailed data on youth detention and adult incarceration. Moreover, the NLSY
dataset includes a comprehensive range of variables on entrepreneurship and employment,
allowing us to examine entrepreneurial and employment transitions as well as related earnings.
Our analyses use the NLSY 97 from years 1997 to 2015. Our final estimation sample is a
balanced panel with 170,696 individual-year observations on 8,984 unique individuals.
We merged the NLSY 97 data with a hand-coded database on “Ban-the-Box” policy
changes for all states and counties of the United States. We generated this data by combining
details of “Ban-the-Box” policies from the National Employment Law Project (NELP) (Avery
and Hernandez 2018). We supplemented the data from NELP by hand-coding details from
17
legislative bills and executive orders of states and counties on the implementation of the “Ban-
the-Box” policy or the Fair Chance Act. We collected data on the effective date of the policy and
whether the policy included public, private, and/or contract employers, as shown in Table 1.
5
Our
analyses consider “Ban-the-Box” policies for public employers effective by December 2015,
which results in 18 states and 133 counties that implement the “Ban-the-Box” policy.
[Insert Table 1 Here]
We measured entrepreneurship by examining whether a survey respondent reported their
job as “self-employed”. The NLSY97 surveys explicitly define self-employment as: “self-
employed jobs are where you own your own business (for example, a lawn service) or where you
do the same type of task for many different people (designing web sites, for instance). In self-
employed jobs, you are your own boss”.
6
This definition of self-employment is consistent with
those used in surveys such as the Current Population Survey (CPS), the official source of data on
employment and unemployment in the United States, as well as previous studies of
entrepreneurship (e.g., Light 1972, Portes and Zhou 1996, Yang and Aldrich 2014, Evans and
Leighton 1989, Hegde and Tumlinson 2015). In order to exclude short-term self-employment
stints or freelancing that are unlikely to be actual entrepreneurship, we only include self-
employment spells that last more than 4 weeks to measure entrepreneurship. As a result, on
average 9.23 percent of the survey respondents were self-employed each year and 32.62 percent
of the survey respondents had at least once experienced self-employment throughout the survey
rounds. The statistics from our sample are comparable with CPS, which reports self-employment
5
When information about a policy’s effective date was available, we used that date as the start date of the policy;
otherwise we used the date the policy was announced or passed by the legislature. If only the year (month) of
implementation was available, we used January 1 of that year (the first of that month) as the start date.
6
NLSY variable codebook on self-employment can be found:
https://www.nlsinfo.org/sites/nlsinfo.org/files/attachments/17036/NLSY97R17Employment.html
18
rates in the United States were approximately 10 11 percent during this period (Hippel 2010).
For robustness, we also measured entrepreneurship as (a) the subset of self-employed individuals
who report owning an incorporated business and (b) the subset of self-employed individuals who
report having employees. Our results are consistent throughout.
Prior incarceration is measured by whether the survey respondent served time in a
correctional institution. The NLSY97 documents monthly status of whether the respondent was
incarcerated or not in each month of the year, collected yearly from 1992 to 2015. The prior
incarceration variable is 1 if the respondent responded as previously incarcerated in any months
of year t-1 or earlier, and 0 otherwise. This prior incarceration variable provides the key
information needed to estimate the effect of incarceration after release. We find that on average
1.59 percent of the survey respondents were currently incarcerated each year and 9.34 percent
have been incarcerated at least once during the years 1992 to 2015. This is similar to statistics
from other studies and samples, where the percentage of those that have been previously
incarcerated range from 7.8 percent to 9.2 percent (Western 2002, Bonczar and Beck 1997). The
accuracy of incarceration as measured in the NLSY has been further assessed by Western (2002)
that shows comparable incarceration trends between the NLSY survey data and aggregate data
from the CPS and BJS administrative data. Individuals were excluded from analyses in the years
when they were currently incarcerated as they are unable to participate in employment or
entrepreneurship. Current incarceration status can also be controlled for by adding a variable
measuring whether respondent was incarcerated in current year t. This approach yields
substantively identical results.
We measured recidivism by examining whether a formerly incarcerated individual is
reincarcerated after release from prison or jail. There has been active discussion on how to
19
effectively measure recidivism, as recidivism has been a “fruit salad concept” measured in a
variety of ways by different scholars (Beck 2001). While recidivism is conceptually defined as
“reengaging in criminal behavior after receiving a sanction or undergoing an intervention for a
previous crime” (National Institute of Justice 2014, Johnson 2017), there is considerable
disagreement on how to operationalize this concept in terms of the scope of criminal behavior
and the time frame. Some scholars broadly define recidivism by counting any new contact with
the criminal justice system, including minor offenses and rearrests for technical violations
(Bureau of Justice Statistics 2016, United States Sentencing Commission 2016). Yet other
scholars argue that recidivism should be more narrowly defined as the commission of a new
serious offense, resulting in a new sentence (Administrative Office of the U.S. Courts, 2015).
Our study follows this second school of thought and measures recidivism as the re-incarceration
of formerly incarcerated individuals for a new offense after release from their previous sentence.
We find that 32.2 percent of formerly incarcerated individuals fall into recidivism under this
measure. For robustness, we also measure recidivism as the re-arrest of formerly incarcerated
individuals for a new offense after release, which yields 44.28 percent recidivism rates. Both
measures are consistent with the average recidivism rate in the United States found by other
scholars (e.g. Bureau of Justice Statistics 2018), and we find identical results with both measures
of recidivism.
Earnings is measured as yearly income and logged yearly income for robustness. The
NLSY97 surveys report income after checking the information against the individuals’
information gathered from Employer Surveys and Current Population Surveys (CPS). It is
possible that entrepreneurs’ incomes and wealth are higher, not because they are compensated
more for their work, but because they work more (Hegde & Tumilson 2018). Therefore, we
20
conducted robustness checks by measuring individual earnings through their log hourly pay rates
and find consistent results. Finally, according to some scholars, entrepreneurs under-report their
income by as much as 30 percent (Sarada 2010). Therefore, we also use the reported net worth of
respondents as a measure of their overall wealth and find consistent results.
The NLSY 97 categorizes race and ethnicity as Black, Hispanic, Asian, White, and mixed
race. Our sample holds 25.99 percent Blacks, 21.16 percent Hispanics, 1.78 percent Asians, and
50.19 percent Whites, and 0.92 percent mixed race. We control for each race and ethnicity by
adding dummy variables for all race categories except Whites, the omitted category. For most of
our analyses, we include individual-level fixed effects that control for race. We also examine
sub-samples by race, as past scholarship on labor market discrimination and incarceration
suggests that the main effects for each race may differ (Western 2002, Pager 2003). Indeed, we
observe substantially different likelihoods of incarceration by race: 13.10 percent of Blacks, 9.26
percent of Hispanics, 4.38 percent of Asians, 7.42 percent of Whites, 13.25 percent of mixed
race were ever incarcerated. We also control for gender. In our sample 48.80 percent are female.
Female respondents and male respondents also show different likelihood of incarceration, as 3.76
percent of the female respondents were ever incarcerated, compared to 14.66 percent of the male
respondents.
We account for individual-level human capital differences that may affect either the
likelihood of being incarcerated or selection into entrepreneurship and employment. Specifically,
we include variables on individual educational attainment and cognitive ability. We measured
educational attainment by the log years of total education completed. We find similar results
when measuring educational attainment as the highest educational degree attained. Cognitive
ability of individuals is measured by the percentiles generated from the Armed Services
21
Vocational Aptitude Battery (ASVAB) Test scores. The ASVAB Test or the composite
percentile generated from this test (Armed Forces Qualification Test Score), has been used to
measure the cognitive ability of individuals in the setting of both incarceration and
entrepreneurship (Western 2002, Fairlie 2002, Hegde and Tumilson 2018). The ASVAB Test
measures the respondent’s knowledge and skills in the topical areas of Arithmetic Reasoning,
Math Knowledge, Word Knowledge, and Paragraph Comprehension. The NLSY respondents
took the ASVAB from the summer of 1997 through the spring of 1998 when they were 12 to 18
years of age. We use the age-adjusted percentiles of the ASVAB test scores, which were
generated using the procedures described in the NLSY 97 Codebook Supplement Appendix 10.
7
We also control for lagged yearly individual and family income, as scholars have found
wealth to drive decisions to engage in entrepreneurship by providing resources that facilitate
both the founding and management of a business (Renzulli, Aldrich, and Moody 2000, Evans
and Jovanovic, 1989; Evans and Leighton, 1989). We measure lagged yearly individual and
family income through annual survey questions that address respondents’ own and total family
income in the previous year. We also control for the number of months worked, including both
self-employment and employment, in the previous year. We also control for the local
unemployment rate at the MSA-level, as the unemployment rate of the local area of residence
may affect the respondent’s employment and entrepreneurship opportunities in addition to
incarceration. Finally, we include year and county (MSA) level fixed effects.
7
NLSY 97 Codebook Supplement Appendix 10 can be found here:
https://www.nlsinfo.org/content/cohorts/nlsy97/other-documentation/codebook-supplement/appendix-10-cat-asvab-
scores
22
Table 2 shows the descriptive statistics of the main variables from year 1997 to 2015. The
first section of Table 2 provides the statistics for the full individual-year sample with 170,696
observations, while the second section of Table 2 provides the statistics for variables that apply
to only formerly incarcerated individuals with 7,369 observations. Table 3 further provides
illustrative statistics for one year, 2010. This table shows the individual-level statistics for all
individuals, individuals who are never incarcerated, and those who have been formerly
incarcerated, sorted by race. The raw statistics show that entrepreneurship rates among formerly
incarcerated individuals is higher than for those never incarcerated, supporting our theory.
[Insert Table 2 Here]
[Insert Table 3 Here]
METHODS
To study incarceration and entrepreneurship, we conduct a series of OLS regression analyses.
First, we estimate the probability that an individual engages in entrepreneurship as a function of
former incarceration, race, and other control variables such as cognitive ability, education, and
prior income through the main model:
        
where  represents whether survey respondent i engaged in entrepreneurship
at year t for the period 1997-2015,  measures whether respondent i was
formerly incarcerated at time t,  is a vector of other individual level control variables, and
23
is an error term. For this model, incarceration produces a shift in the probability of
entrepreneurial engagement by percent.
Yet, it is difficult to interpret the results from this model as the causal effect of
incarceration on entrepreneurship, because of the nonrandom selection of individuals into
incarceration. Preexisting characteristics of formerly incarcerated individuals that place them at
high risk of incarceration may also affect their likelihood of engaging in entrepreneurship. The
increase of entrepreneurship for formerly incarcerated individuals may be a function of
preexisting traits of formerly incarcerated individuals (e.g. risk preference, entrepreneurial
ability, poor interpersonal skills) instead of the “treatment” effect of incarceration itself.
Furthermore, it is difficult to formally identify explanatory mechanisms through the analysis of
incarceration and subsequent entrepreneurship outcomes. Pager (2003) raises similar limitations
of the difficulty of parsing out underlying mechanisms for studies on incarceration and
subsequent employment outcomes. While researchers have offered numerous mechanisms that
may explain the observed relationship between incarceration and employment such as the
influence on social networks (Hagan 1993), the loss of human capital (Becker 1975),
institutional trauma (Parenti 1999), and legal barriers to employment (Dale 1976), it has been
difficult to discern which of these causal mechanisms are at work. Therefore, in order to identify
a causal relationship between incarceration and entrepreneurship and to establish the explanatory
mechanism, this paper utilizes several research designs and a unique empirical setting.
Firstly, we conduct fixed effects regression models. We include individual-level fixed
effects to absorb time-invariant, observed and unobserved, individual traits such as cognitive
ability, personality, impulsivity, risk preferences, and fixed demographic characteristics such as
race and gender (Caspi et al. 1998). By including the individual fixed effects, we are able to
24
observe within individual changes of entrepreneurship and employment choices, before and after
being incarcerated. We also include year fixed effects to capture any time trends and MSA
(county) fixed effects to control for differences between regions.
In addition to fixed effects, we exploit a quasi-natural experiment provided by an
exogenous policy shock and a triple-difference regression analysis. Specifically, we use the
staggered enactment of a policy widely known as “Ban-the-Box”, a county and state-level law
barring employers from examining job applicants’ criminal records until later in the hiring
process. As of 2019, the “Ban-the-Box” policy has been staggeredly adopted in 35 states,
Washington D.C., and 170 cities and counties in the United States for public employers,
spanning 21 years. Among these localities, 13 states, Washington D.C., and 18 cities and
counties have extended their “Ban-the-Box” policy to private employers as well. Due to the
limited adoption of “Ban-the-Box” policy for private employers before 2015, our analysis in this
paper focuses primarily on the effects of implementing at least a public “Ban-the-Box” policy.
Table 1 shows the list of states, cities, and counties that have adopted the “Ban-the-Box” policy
and the effective dates.
We are able to leverage the staggered adoption of “Ban-the-Box” to identify the causal
mechanism of incarceration on entrepreneurship for several reasons. Firstly, the implementation
of this policy exogenously increases employment opportunities for formerly incarcerated
individuals in impacted localities. This initiative provides formerly incarcerated job applicants a
better chance at employment by removing the conviction history question from job applications
and allowing employers to judge applicants on their qualifications without the stigma of a record
(Craigie 2020, Avery and Hernandez 2018, Shoag and Veuger 2016). By comparing the level of
entrepreneurship engagement in localities with and without the “Ban-the-Box” policy, we are
25
able to identify whether labor market discrimination plays a role in formerly incarcerated
individuals engaging in entrepreneurship. Furthermore, we are able to identify the differences by
race, as Ban-the-Box policy is often presented as an important tool for reducing racial disparity
by improving access to employment for formerly incarcerated black men (Pinard 2014, Clarke
2012). We find supportive evidence from our sample that the enactment of the “Ban-the-Box”
policy does in fact increase employment for formerly incarcerated individuals, particularly black
formerly incarcerated individuals, as will be discussed in the Results section of the paper (refer
to Table 5 and Figure 1).
Moreover, the implementation of the “Ban-the-Box” policy offers a unique setting to
tease apart the causal mechanism of the effect of incarceration on entrepreneurship. The adoption
of “Ban-the-Box” policy has no direct correlation with entrepreneurship, other than through the
policy’s impact on labor market discrimination for formerly incarcerated individuals. The “Ban-
the-Box” policy does not affect other possible causal mechanisms such as social networks, loss
of human capital, or institutional trauma that are known to influence formerly incarcerated
individuals in their labor market choices. In other words, the adoption of the “Ban-the-Box” law
serves as a unique proxy for the change (i.e. decrease) in the level of employer discrimination for
formerly incarcerated individuals in a given locality after the adoption of the ban.
We also exploit the variation in the timing of “Ban-the-Box” implementation in a triple
difference design, to address the concern of whether the adoption of the “Ban-the-Box” policy is
a function of a locality’s entrepreneurship rate, employment rate, or other unobservable traits.
Our triple difference regression model mitigates such concerns, as we compare the changes in
entrepreneurship for formerly incarcerated individuals relative to the changes in entrepreneurship
for non-formerly incarcerated individuals, in “Ban-the-Box” localities versus non “Ban-the-Box”
26
localities, after “Ban-the-Box” policies go into effect. We further validate that the treatment
(“Ban-the-Box” implemented) and control (non “Ban-the-Box” implemented) localities had no
preexisting trends in entrepreneurship or employment for formerly incarcerated and non-
formerly incarcerated individuals before the year of the policy adoption, suggesting that adoption
of “Ban-the-Box” policy was not an endogenous choice based on entrepreneurship or
employment trends (refer to Figure 1 & Figure 2).
Specifically, we run a triple difference regression by using the model:

        
      
where  represents whether survey respondent i engaged in entrepreneurship
at year t for the period 1997-2015,  measures whether respondent i was
formerly incarcerated at time t,  is 1 when the county or state of residence for
respondent i at time t has adopted the “Ban-the-Box” policy and 0 otherwise,
   is the interaction of prior incarceration and the enactment
of “Ban-the-Box” policy,    are fixed effects at the individual-level, year-level, and
MSA (county) level respectively,  is a vector of time-variant control variables, and is an
error term.
8
Although our methodology of utilizing the triple difference method before and after an
exogenous policy shock addresses most endogeneity problems, for robustness we utilize a
8
The triple difference specification follows Imbens and Woodridge (NBER, 2007).
27
matching method in order to address remaining concerns about non-random selection into
incarceration. As formerly incarcerated individuals may be very different from the comparison
group of individuals without an incarceration record, our matching method helps to restrict the
control group to people similar to the treatment group (formerly incarcerated individuals),
except that this “control group” has not been “treated” (incarcerated) (Western 2002). We utilize
three different matching methods, (1) coarsened exact matching (CEM), (2) propensity score
matching (PSM), and (3) matching the treatment group of formerly incarcerated individuals to
formerly arrested yet not incarcerated individuals. The three types of matching are different in
methods, but all aim to address the issue of non-comparable treatment and control group.
Through CEM and PSM matching, we use an extensive set of variables such as demographics,
education, family background, and residence area in order to create an observationally similar
comparison group. For our third matching, we match the treatment group of formerly
incarcerated individuals to the subset of the control group who have an arrest record but have not
been incarcerated. All three of our matching methods show that after matching, the treatment
group and control group are similar in terms of the extensive set of variables used for matching.
(Refer to Appendix)
RESULTS
Table 4 shows the results on how incarceration affects entrepreneurship. Our results show
that formerly incarcerated individuals are more likely to become entrepreneurs compared to
individuals without a criminal record. The results show that having been formerly incarcerated
increased one’s likelihood of becoming an entrepreneur by 5.9 percent (Table 4 Model 1). This is
28
true controlling for education, cognitive ability test scores, family and individual income (t-1),
number of months employed (t-1), gender, and race (Table 4 Model 2). While the average
individual without a criminal record has 7.09 percent likelihood of becoming an entrepreneur,
similar individuals with a criminal record are more than 50 percent more likely to choose
entrepreneurship with a 12.69 percent likelihood of becoming an entrepreneur (Table 4 Model
2). Table 4 Model 3 shows similar results when adding individual fixed effects. We find
consistent results when measuring entrepreneurship as incorporated self-employment (Table 4
Model 4) and self-employment with employees (Table 4 Model 5).
[Insert Table 4 Here]
Table 5 and Table 6 show the triple difference analyses with the staggered
implementation of the “Ban-the-Box” policy. Through the results we are able to make a causal
claim, and also unpack an underlying mechanism of why formerly incarcerated individuals
choose entrepreneurship at higher rates compared to individuals who have not been to prison or
jail. First, in Table 5 we test the effects of the “Ban-the-Box” policy on employment for formerly
incarcerated individuals, by examining the number of months employed per year in the paid-
employment sector. While the average individual without a criminal record works in paid
employment for 7.18 months in a year, past incarceration decreases employment by 0.52 months
(7.2 percent) after release (Table 5 Model 1). However, the negative impact of incarceration on
employment is mitigated when “Ban-the-Box” policy is implemented, as employment for
formerly incarcerated individuals increase by 0.4 months in states or counties where the policy is
adopted (Table 5 Model 1). Formerly incarcerated individuals still face discrimination from
employers when residing in states or counties with “Ban-the-Box” policy, but they are more
likely to be employed compared to before the implementation of “Ban-the-Box” policy.
29
[Insert Table 5 Here]
Figure 1 summarizes the effect of “Ban-the-Box” policy on employment for formerly
incarcerated individuals. The solid horizontal line represents the baseline employment of
formerly incarcerated individuals in states and counties that never implement “Ban-the-Box”
policy. The dashed line shows employment for formerly incarcerated individuals in states and
counties that implement “Ban-the-Box” policy at time T (labeled “Ban-the-Box”). The figure
shows that the two lines overlap during T-2 and T-1, confirming that “Ban-the-Box” policy was
indeed exogenous without any pre-trends between the counties and states that implemented and
did not implement “Ban-the-Box” policy. Also, Figure 1 shows that at time T, when “Ban-the-
Box” policy is implemented, employment for formerly incarcerated individuals sharply increased
in counties and states that implement Ban-the-Box. This increase of employment for formerly
incarcerated individuals in “Ban-the-Box” implemented counties and states continues after time
T. These results suggest that (1) “Ban-the-Box” policy implementation is an exogenous shock to
employment for formerly incarcerated individuals, and that (2) “Ban-the-Box” policy helps
mitigate labor market discrimination for formerly incarcerated individuals.
[Insert Figure 1 Here]
Furthermore, we find differential effects of incarceration and “Ban-the-Box” policy by
race. We divide the sample into black and white population in order to examine how past
incarceration and “Ban-the-Box” policy affects employment opportunities differently by race.
Models 2 and 3 of Table 5 show that black formerly incarcerated individuals face greater
discrimination from employers compared to white formerly incarcerated individuals, supporting
previous findings (Pager 2003). Specifically, our results show that while the average black
individual without a criminal record works in paid employment for 6.55 months in a year, black
30
formerly incarcerated individuals are employed 0.83 months (12.7 percent) less (Table 5 Model
2). The average white individual without a criminal record works in paid employment for 7.46
months in a year, and a criminal record decreases employment by 0.49 months (6.6 percent)
(Table 5 Model 2). Interestingly, the adoption of “Ban-the-Box” mitigates such employment
discrimination for black formerly incarcerated individuals, but does not significantly increase
employment for white formerly incarcerated individuals.
9
Thus, the “Ban-the-Box” policy has
the greatest positive employment impact on the individuals who face the greatest discrimination
from employers: black formerly incarcerated individuals.
Table 6 shows the triple difference OLS result for entrepreneurship, where evidence for
our causal mechanism is indicated by lower entrepreneurship rates for formerly incarcerated
individuals after “Ban-the-Box” is adopted. Table 6 Model 1 provides consistent results as Table
4, showing that past incarceration increases one’s likelihood of engaging in entrepreneurship by
4.9 percent, compared to non-formerly incarcerated individuals. Yet, the coefficient for the
interaction of “Past Incarceration” and “Ban-the-Box indicates that the exogenous
implementation of “Ban-the-Box” does not significantly change the incarceration effect on the
likelihood of entrepreneurial entry.
In order to probe deeper, we conduct sub-sample analyses by race (Table 6 Model 2 and
Model 3). The sub-sample analyses show that formerly incarcerated individuals are less likely to
take part in entrepreneurship when “Ban-the-Box” policy is adopted, but only for the black
formerly incarcerated population. This is consistent with past scholarship, as well as the results
from Table 5 which show evidence that not only are black formerly incarcerated individuals the
9
We find that “Ban-the-Box” policy negatively impacts non-formerly incarcerated black individuals, by decreasing
their employment. This result speaks to prior research on “Ban-the-Box” employment effects such as Agan and Starr
(2017) and Doleac and Hansen (2017). We discuss further in the Discussion section of this paper.
31
most discriminated against in the labor market, but also experience the greatest increase in labor
market opportunities through Ban-the-Box. Overall, the results from Table 6 show that black
formerly incarcerated individuals, those who face the most discrimination from employers and
the greatest employment benefits from Ban-the-Box, are less likely to take part in
entrepreneurship when “Ban-the-Box” is implemented. This supports our central thesis that
formerly incarcerated individuals are more likely to pursue entrepreneurship as an alternative
route to avoid labor market discrimination.
10
[Insert Table 6 Here]
Figure 2 clearly summarizes the effect of “Ban-the-Box” policy on entrepreneurship for
formerly incarcerated individuals. The solid horizontal line at zero represents the baseline
entrepreneurship likelihood of formerly incarcerated individuals in states and counties that never
implemented “Ban-the-Box” policy. The dashed line shows the relative entrepreneurship
likelihood of formerly incarcerated individuals in states and counties that implement “Ban-the-
Box” policy at time T (labeled “Ban-the-Box”). Similar to Figure 1, Figure 2 shows that the two
lines overlap during T-2, T-1, confirming that “Ban-the-Box” policy is indeed exogenous
without any pre-trends between the counties and states that implemented and did not implement
“Ban-the-Box” policy. Also, Figure 2 shows that at time T+1, a year after “Ban-the-Box policy
is implemented, entrepreneurship for formerly incarcerated individuals sharply decreases in
counties and states that implement Ban-the-Box. This decrease of entrepreneurship for formerly
incarcerated individuals in “Ban-the-Box” implemented counties and states continues on after
10
We interpret the increase of entrepreneurship for non-formerly incarcerated black individuals after “Ban-the-Box”
policy implementation to be the impact of “Ban-the-Box” policy decreasing employment for this population. This is
in line with our overall theory that individuals are pushed into entrepreneurship due to the lack of work opportunities
in the labor market.
32
time T+1. These results show support for our hypotheses that formerly incarcerated individuals
transition into entrepreneurship because of labor market discrimination, and that the mitigation
of discrimination negatively impacts entrepreneurial entry.
[Insert Figure 2 Here]
Next, Table 7 shows the results on how past incarceration and entrepreneurship affects
annual income in dollars. Table 7 Model 1 first shows that incarceration has a significant
negative impact on yearly income. Specifically, in terms of yearly income in dollars,
entrepreneurs with a criminal record earn approximately 2,700 dollars more than employees with
a criminal record. Furthermore, while employees with a criminal record earn approximately
7,000 dollars less than employees without a criminal record each year, entrepreneurs with a
criminal record earn only 4,300 dollars less each year than entrepreneurs without a criminal
record.
11
Thus although formerly incarcerated individuals still earn significantly less than non-
formerly incarcerated individuals in entrepreneurship, the income penalty from past incarceration
decreases by 38.6 percent. Table 7 Model 2 and Model 3 show consistent results for the sub-
samples by race. These results support our theory that the income penalty that formerly
incarcerated individuals face due to labor market discrimination and stigma can be mitigated by
taking part in entrepreneurship. Particularly, our results that a portion of the income gap remains
from incarceration even in entrepreneurship seems to suggest evidence for individual-level
effects of incarceration such as the human capital or social capital erosion. Thus,
entrepreneurship helps formerly incarcerated individuals overcome the institutional level effects
11
The results are similar when using 3-year average income.
33
of labor market discrimination. For robustness, we test and find consistent results with logged
yearly income, logged hourly pay rate, and net worth.
[Insert Table 7 Here]
Lastly, Table 8 shows results for the effect of entrepreneurship on recidivism rates for
formerly incarcerated individuals. We find supportive evidence that entrepreneurship helps
prevent formerly incarcerated individuals from returning to prison, beyond the effect of
employment. Model 1 of Table 8 shows that entrepreneurship decreases the likelihood of
recidivism (measured by re-incarceration) by 5.3 percent, which is a 32.5 percent decrease from
the average recidivism rate for formerly incarcerated individuals who are employees. Model 1
supports previous research with results that show longer incarceration length increases
recidivism while the number of years since release from incarceration decreases recidivism. The
sub-sample analyses by race shows interesting results that entrepreneurship helps desist further
crime, only for black formerly incarcerated individuals (Table 8 Model 2 and Model 3). This is
consistent with our theory that entrepreneurship is most helpful as an alternative route for work
for those facing the most discrimination in the labor market: black formerly incarcerated
individuals. Table 8 Model 4 shows consistent results with Model 1, by measuring recidivism as
re-arrests.
[Insert Table 8 Here]
DISCUSSION & FUTURE DIRECTIONS
Our study shows that people who have spent time in prison are more likely to become
entrepreneurs compared to similar individuals who have not been incarcerated, signifying
34
entrepreneurship as a meaningful choice for formerly incarcerated individuals. Through our
quasi-experiment design, we verify an underlying mechanism that formerly incarcerated
individuals choose to become entrepreneurs because of the lack of employment opportunities in
the labor market. In addition, we find evidence that entrepreneurship offers formerly incarcerated
individuals the chance to overcome both economic and social barriers to successful reentry by
decreasing the income gap and recidivism rates.
These findings offer a modification to a prevailing narrative on formerly incarcerated
individuals, which has emphasized labor market discrimination and its adverse consequences on
subsequent reentry. Consistent with this narrative, we verify past scholarship that points to the
value of employment. However, we extend this research by drawing attention to
entrepreneurship as an alternative labor market choice that formerly incarcerated individuals can
pursue to mitigate the stigma associated with the mark of a criminal record. Entrepreneurship not
only helps formerly incarcerated people find work and gain competitive income, but also lowers
the likelihood of returning to prison. While employment remains an important channel to
successful reentry, we introduce entrepreneurship as an alternative way formerly incarcerated
people can achieve both economic and social reintegration.
Our study speaks to the important discussion on the intersection of race and incarceration,
by underlining the significance of race in the role of entrepreneurship for formerly incarcerated
people. Our results support prior research on the persistent racial inequality in employment
opportunities for formerly incarcerated individuals by showing that African American (black)
individuals with a criminal record are those who face the highest employment barriers. Yet, our
findings that black formerly incarcerated individuals reap the greatest advantages from
entrepreneurship, emphasizes that entrepreneurship offers the opportunity for economic and
35
social integration, particularly for those who face the greatest stigma and discrimination in the
labor market - African Americans.
Our quasi-experimental study design allows us to disentangle the underlying mechanism
and offer direct causal evidence of incarceration on entrepreneurship. While survey research can
have limitations of indirect estimates of effects, our research design utilizes an exogenous policy
shock with a triple-difference method. This allows a direct and causal measure of a criminal
record and labor market discrimination as a mechanism that drives entrepreneurial decisions.
This methodology allows us to effectively isolate the institutional effect from the individual
effect of incarceration, and identify entrepreneurship as a response to labor market
discrimination and stigma.
Addressing this important channel of reentry for formerly incarcerated individuals not
only contributes to research but also has implications for policymakers and practitioners.
Examining entrepreneurship as a valid opportunity for formerly incarcerated individuals may
draw attention to the importance of investing in programs and policies to facilitate post-
incarceration entrepreneurial activities, as well as better understand discrimination in
employment markets. As many studies have found that the lack of employment influences
formerly incarcerated individuals to return to prison, this study draws attention to the importance
of entrepreneurship as a way of decreasing recidivism. While there have been recent policy
initiatives such as the New Start Act (Marks 2019) and efforts from non-profit organizations and
educational institutions, our study is one of the first research studies to emphasize the need for
attention to entrepreneurship for formerly incarcerated people.
36
Relatedly, our paper has theoretical implications on public policies that affect people with
disadvantages, such as a criminal record. For example, our research speaks to the discussion
around the “Ban-the-Box” policy. Scholars, policy makers, and practitioners have debated the
effects of the “Ban-the-Box” policy, with serious disagreement. While some have argued that
“Ban-the-Box” policy increases discrimination against racial minorities (Agan and Starr 2018;
Doleac and Hansen 2018), others have found counter evidence suggesting that this policy
reduces discrimination (Craigie 2020, Pinard 2014, Southern Coalition for Social Justice 2013,
Clarke 2012, and Community Catalyst 2013) or negligible effects (Rose 2019). While our
research is not designed to assess the impact and effectiveness of the “Ban-the-Box” policy, we
offer some findings that relate to this conversation. In our study, we are able to observe
individuals with a criminal record before and after “Ban-the-Box” policy enactments, which has
been difficult in audit studies with fictitious job applicants (Agan and Starr 2018) and Current
Population Survey studies that don’t report incarceration variables (Doleac and Hansen 2018).
By being able to separately observe formerly and non-formerly incarcerated individuals after
“Ban-the-Box” policy, we are able to more accurately assess the policy implications for each
different group. We find that “Ban-the-Box” increases employment for formerly incarcerated
individuals (both black and white), and that “Ban-the-Box” has a significant negative effect on
the employment outcomes of African Americans who have not been incarcerated. While this
finding is a result of our specific sample, research design, and measurement of employment
(number of months employed), our findings suggest the need for more studies to investigate the
impact of “Ban-the-Box” policies on both formerly incarcerated and non-formerly incarcerated
individuals.
37
This study also provides contribution to work on incarceration as one of the first papers
to address entrepreneurship of formerly incarcerated people. While recent studies have started to
examine entrepreneurship for formerly incarcerated individuals through qualitative analyses of
entrepreneurial training programs in prisons (e.g. Cooney 2012), our study is the first study to
offer quantitative analyses on entrepreneurial transitions and outcomes for formerly incarcerated
individuals. We believe our research opens future research possibilities on formerly incarcerated
entrepreneurs. For example, scholars should examine the entrepreneurial process of formerly
incarcerated individuals and how the entrepreneurial experience of formerly incarcerated people
impacts their future employment prospects.
An increasing number of individuals are returning back to society from prisons and jails
as a consequence of mass incarceration. Thus, it becomes increasingly important to consider the
impact of incarceration on reentry and how formerly incarcerated individuals can overcome the
common pathway to unemployment and recidivism. Our paper is an initial attempt to introduce
entrepreneurship as an alternative response to the poor employment outcomes and labor market
discrimination that await formerly incarcerated individuals. Future research is needed to expand
this emphasis on entrepreneurship by exploring the antecedents, process, and diverse outcomes
of entrepreneurship for formerly incarcerated individuals. In this way, we can move toward a
more complete understanding of the labor market choices that formerly incarcerated individuals
can make in order to successfully reenter and remain in the society.
38
REFERENCES
Amit, R., Muller, E. and Cockburn, I., 1995. Opportunity costs and entrepreneurial
activity. Journal of business venturing, 10(2), pp.95-106.
Agan, A., & Starr, S. (2017). Ban the box, criminal records, and racial discrimination: A field
experiment. The Quarterly Journal of Economics, 133(1), 191-235.
Aldrich, Howard E., J. Cater, T. Jones, D. McEvoy, and P. Velleman. 1985. "Ethnic Residential
Concentration and the Protected Market Hypothesis." Social Forces 63: 996- 1009.
Alper M, Durose MR, Markman J. 2018 update on prisoner recidivism: A 9-year follow-up
period (2005-2014). Washington, DC: US Department of Justice, Office of Justice Programs,
Bureau of Justice Statistics; 2018 May.
Avery, B. and Hernandez, P. 2018. Ban the Box, U.S. Cities, Counties, and States Adopt Fair
Chance Policies to Advance Employment Opportunities for People with Past Convictions.
Washington, DC: National Employment Law Project
Bates, T. (1995). Self-employment entry across industry groups. Journal of business venturing,
10(2), 143-156.
Bates, T. (1986). Characteristics of minorities who are entering self-employment, The Review
of Black Political Economy, 15 (2), pp. 31-49
Bates, T. (2011), "Minority Entrepreneurship", Foundations and Trends® in Entrepreneurship:
Vol. 7: No. 34, pp 151-311. http://dx.doi.org/10.1561/0300000036
Becker, G. S. (1968). Crime and punishment: An economic approach. The Journal of Political
Economy, pages 169217.
Becker, G.S., 1975. Front matter, human capital: a theoretical and empirical analysis, with
special reference to education. In Human Capital: A Theoretical and Empirical Analysis, with
Special Reference to Education, Second Edition (pp. 22-0). NBER.
Blanchfower, D.G. and Oswald, A.J.1998..What makes an entrepreneur?. Journal of Labor
Economics, 16:26-60.
Bushway, S.D. and Reuter, P., 2011. Deterrence, economics, and the context of drug
markets. Criminology & Public Policy, 10, p.183.
Bushway, Shawn D. 1997. “Labor Market Effects of Permitting Employer Access to Criminal
History Records.” Working paper. University of Maryland, Department of Criminology
Bushway, S.D., 1998. The impact of an arrest on the job stability of young white American
men. Journal of research in Crime and Delinquency, 35(4), pp.454-479.
39
Caspi, A., Wright, B.R.E., Moffitt, T.E. and Silva, P.A., 1998. Early failure in the labor market:
Childhood and adolescent predictors of unemployment in the transition to adulthood. American
sociological review, pp.424-451.
Castellaneta, F., Conti, R., & Kacperczyk, O. (2018). The (Un) intended Consequences of
Lowering Entry Barriers: Evidence from an Entry Deregulation Reform in Portugal. Working
Paper
Clarke, H., 2012. Protecting the Rights of Convicted Criminals: Ban the Box Act of
2012. Washington Post.
Craigie, T.A., 2020. Ban the Box, Convictions, and Public Employment. Economic Inquiry,
58(1), pp.425-445.
Dale, Mitchell. 1976. “Barriers to the Rehabilitation of Ex-Offenders.” Crime and Delinquency
22:32237.
Doleac, J. and Hansen, B., 2017. The unintended consequences of" ban the box'': Statistical
discrimination and employment outcomes when criminal histories are hidden.
D.S. Evans, B. Jovanovic (1989) An estimated model of entrepreneurial choice under liquidity
constraints, Journal of Political Economy, 97 (4), pp. 808-827
D.S. Evans, L. Leighton (1989) Some empirical aspects of entrepreneurship, The American
Economic Review, 79, pp. 519-535
Fairlie, R. W. (1999). The absence of the African-American owned business: An analysis of the
dynamics of self-employment. Journal of Labor Economics, 17(1), 80-108.
Fairlie, R.W., 2002. Drug dealing and legitimate self-employment. Journal of Labor
Economics, 20(3), pp.538-537.
Fairlie, R., & Meyer, B. (1996). Ethnic and Racial Self-Employment Differences and Possible
Explanations. The Journal of Human Resources, 31(4), 757-793. doi:10.2307/146146
Fairlie, R. W., & Meyer, B. D. (2003). The effect of immigration on native self-employment.
Journal of Labor Economics, 21(3), 619-650.
Farrington, David P., Bernard Gallagher, Lynda Morley, Raymond J. St. Ledger, and Donald J.
West. 1986. "Unemployment, School Leaving, and Crime." British Journal of Criminology
26:335-56.
Freeman, Richard B. 1991. "Crime and the Employment of Disadvantaged Youths." National
Bureau of Economic Research Working Paper No. 3875. Cambridge, Mass.: NBER 1996.
Glazer, Nathan and Daniel P. Moynihan. 1963. Beyond the Melting Pot. Cambridge, MA: MIT
Press.
40
Gottschalk, P., 2009. Entrepreneurship and organised crime: Entrepreneurs in illegal business.
Edward Elgar Publishing.
Grogger, J., 1992. Arrests, persistent youth joblessness, and black/white employment
differentials. The Review of Economics and Statistics, pp.100-106.
Hagan, J., 1993. The social embeddedness of crime and unemployment. Criminology, 31(4),
pp.465-491.
Harding, D.J., Morenoff, J.D. and Wyse, J.J., 2019. On the outside: Prisoner reentry and
reintegration. University of Chicago Press.
Hegde, Deepak and Justin Tumilson 2018, Asymmetric Information and Entrepreneurship,
Working Paper
Hellmann, T., 2007. When do employees become entrepreneurs?. Management science, 53(6),
pp.919-933.
Holzer, H.J., Raphael, S. and Stoll, M.A., 2003. Employer demand for ex-offenders: Recent
evidence from Los Angeles. Institute for Research on Poverty, University of Wisconsin-Madison.
Imbens, G. and Wooldridge, J. 2007. What’s New in Econometrics? Difference-in-Difference
Estimation. NBER Lecture Notes.
Kacperczyk, A. J. (2012). Opportunity Structures in Established Firms: Entrepreneurship versus
Intrapreneurship in Mutual Funds. Administrative Science Quarterly, 57(3), 484521.
https://doi.org/10.1177/0001839212462675
Kacperczyk, A. and Marx, M., 2016. Revisiting the small-firm effect on entrepreneurship:
Evidence from firm dissolutions. Organization Science, 27(4), pp.893-910.
Kaeble, D., and Cowhig, M. 2018. “Correctional Populations in the United States, 2016”. US
Department of Justice, Office of Justice Programs, Bureau of Justice Statistics; 2018 April.
Kling, J.R., 1999. The effect of prison sentence length on the subsequent employment and
earnings of criminal defendants (No. 208).
Light, I.H. and Paden, J.N., 1973. Ethnic enterprise in America: Business and welfare among
Chinese, Japanese, and Blacks. Univ of California Press.
Lockwood, F., Teasley, R., Carland, J.A.C. and Carland, J.W., 2006. An examination of the
power of the dark side of entrepreneurship. International Journal of Family Business, 3(1), pp.1-
20.
Lofstrom, M. (2007). Mexican-American Self-Employment : A Dynamic Analysis of Business
Ownership.
41
Lofstrom, M., Bates, T., & Parker, S. C. (2014). Why are some people more likely to become
small-businesses owners than others: Entrepreneurship entry and industry-specific barriers.
Journal of Business Venturing, 29(2), 232-251.
Lott Jr, J.R., 1992. Do we punish high income criminals too heavily?. Economic Inquiry, 30(4),
pp.583-608.
Lyons, C.J. and Pettit, B., 2011. Compounded disadvantage: Race, incarceration, and wage
growth. Social Problems, 58(2), pp.257-280.
Min, Pyong Gap. 1988. Ethnic Small Business Enterprise. Staten Island: Center for Migration
Studies:
Müller, W., and R., Arum. 2004. “Self-Employment Dynamics in Advanced Economies.” Pp. 1-
35 in R. Arum and W. Müller (eds.), The Re-Emergence of Self-Employment: A Comparative
Study of Self-Employment Dynamics and Social Inequality. Princeton: Princeton University
Press.
Nagin, D. and Waldfogel, J., 1995. The effects of criminality and conviction on the labor market
status of young British offenders. International Review of Law and Economics, 15(1), pp.109-
126.
Pager, Devah. "The mark of a criminal record." American journal of sociology 108.5 (2003):
937-975.
Pager, D., 2007. The use of field experiments for studies of employment discrimination:
Contributions, critiques, and directions for the future. The Annals of the American Academy of
Political and Social Science, 609(1), pp.104-133.
Pager, D., Bonikowski, B. and Western, B., 2009. Discrimination in a low-wage labor market: A
field experiment. American sociological review, 74(5), pp.777-799.
Parenti, Christian. 1999. Lockdown America: Police and Prisons in the Age of Crisis. New York:
Verso
Pettit, E.M. and Lyons, C.J., 2007. Status and the stigma of incarceration: The labor-market
effects of incarceration, by race, class, and criminal involvement. In Barriers to Reentry?: The
Labor Market for Released Prisoners in Post-Industrial America (pp. 203-226). Russell Sage
Foundation.
Pettit, B. and Western, B., 2004. Mass imprisonment and the life course: Race and class
inequality in US incarceration. American sociological review, 69(2), pp.151-169.
Pezzin, L.E. 1995. “Earning Prospects, Matching Effects, and the Decision to Terminate a
Criminal Career.” Journal of Quantitative Criminology 11:29-50.
Pinard, M., 2014. Ban the Box in Baltimore. Baltimore Sun.
42
Portes, A. and Zhou, M., 1996. Self-employment and the earnings of immigrants. American
Sociological Review, pp.219-230.
Renzulli, L.A., Aldrich, H. and Moody, J., 2000. Family matters: Gender, networks, and
entrepreneurial outcomes. Social forces, 79(2), pp.523-546.
Rieple, A., 1998. Offenders and entrepreneurship. European Journal on Criminal Policy and
Research, 6(2), pp.235-256.
Rose, E., 2019. Does banning the box help ex-offenders get jobs? Evaluating the effects of a
prominent example. Journal of Labor Economics, Forthcoming
Rosti, L., & Chelli, F. (2005). Gender discrimination, entrepreneurial talent and self-
employment. Small Business Economics, 24(2), 131-142.
Sampson, R.J. and Laub, J.H., 1997. A life-course theory of cumulative disadvantage and the
stability of delinquency. Developmental theories of crime and delinquency, 7, pp.133-161.
Shoag, D. and Veuger, S. (2016). No woman no crime: Ban the box, employment, and
upskilling. AEI Working Paper.
Shover, N. (1996). Great pretenders: Pursuits and careers of persistent thieves.
Boulder, CO: Westview.
Sonfield, M., Lussier, R. and Barbato, R., 2001. The entrepreneurial aptitude of prison inmates
and the potential benefit of self-employment training programs. Academy of Entrepreneurship
Journal.
Sørensen, J. B., & Sharkey, A. J. (2014). Entrepreneurship as a Mobility Process. American
Sociological Review, 79(2), 328349. https://doi.org/10.1177/0003122414521810
Stoll, M.A. and Bushway, S.D., 2008. The effect of criminal background checks on hiring ex‐
offenders. Criminology & Public Policy, 7(3), pp.371-404.
Sugie, N.F., 2018. Work as foraging: a smartphone study of job search and employment after
prison. American Journal of Sociology, 123(5), pp.1453-1491.
Trasler, G., 1979. Delinquency, recidivism and desistance. Brit. J. Criminology, 19, p.314.
Travis, J. and Visher, C. eds., 2005. Prisoner reentry and crime in America. Cambridge
University Press.
Uggen, C., 1999. Ex-offenders and the conformist alternative: A job quality model of work and
crime. Social Problems, 46(1), pp.127-151.
Uggen, C., 2000. Work as a turning point in the life course of criminals: A duration model of
age, employment, and recidivism. American sociological review, pp.529-546.
43
Uggen, C. and Staff, J., 2001. Work as a turning point for criminal offenders. Corrections
Management Quarterly, 5, pp.1-16.
Uggen, C., Vuolo, M., Lageson, S., Ruhland, E. and Whitham, H.K., 2014. The edge of stigma:
An experimental audit of the effects of low‐level criminal records on employment.
Criminology, 52(4), pp.627-654.
Waldfogel, J. 1994. “Does Conviction Have a Persistent Effect on Income and Employment?”
International Review of Law and Economics, March.
Waldinger, R. D. (1986), Through the Eye of the Needle, New York University Press, New York
Western, Bruce. 2002. “The Impact of Incarceration on Wage Mobility and Inequality.”
American Sociological Review 67 (4): 52646
Western, Bruce, and Katherine Beckett. 1999. “How Unregulated is the U.S. Labor Market? The
Penal System as a Labor Market Institution.” American Journal of Sociology 104 (4): 103060.
Western, Bruce, and Sara McLanahan. 2000. “Fathers behind Bars: The Impact of Incarceration
on Family Formation.” Contemporary Perspective in Family Research 2:309–24.
Western, Bruce, and Becky Pettit. 1999. “Black-White Earnings Inquality, Employment Rates,
and Incarceration.” Working Paper no. 150. New York: Russell Sage Foundation
Western, B. and Pettit, B., 2005. Black-white wage inequality, employment rates, and
incarceration. American Journal of Sociology, 111(2), pp.553-578.
Western, B. and Pettit, B., 2010. Incarceration & social inequality. Daedalus, 139(3), pp.8-19.
Yang, T. and Aldrich, H.E., 2014. Who’s the boss? Explaining gender inequality in
entrepreneurial teams. American Sociological Review, 79(2), pp.303-327.
Yang, T. and Kacperczyk, A.J., 2018, July. Minority Entrepreneurship and Alternative
Opportunities Inside Established Organizations. In Academy of Management Proceedings (Vol.
2018, No. 1, p. 16903). Briarcliff Manor, NY 10510: Academy of Management.
Zeng Z. 2019. “Jail Inmates in 2017”. US Department of Justice, Office of Justice Programs,
Bureau of Justice; 2019 April.
Bureau of Justice Statistics 2016, United States Sentencing Commission 2016 (how to measure
recidivism) Administrative Office of the U.S. Courts, 2015, Bureau of Justice Statistics 2018
44
State Jurisdiction Public Start Date Private Start Date Contract Start Date
Alabama Birmingham 1 2016.02.04
Arizona State 1 2017.11.06
Coconino County 1 2017.05.10
Glendale 1 2015.09.01
Maricopa County 1 2018.01.01
Pima County 1 2015.11.10
Phoenix 1 2016.04.18
Tempe 1 2016.09.22
Tucson 1 2014.08.27
Arkansas Pulaski County 1 2016.06.28
California State 1 2010.06.25 1 2017.10.14
Alameda County 1 2007.03.01
Berkeley 1 2008.10.01
Carson 1 2012.03.06
Compton 1 2011.07.01 1 2011.07.01
East Palo Alto 1 2005.01.01
Los Angeles 1 2016.12.09 1 2016.12.09 1 2016.12.09
Oakland 1 2007.01.01
Pasadena 1 2013.07.01
Richmond 1 2011.11.22 1 2013.07.30
Sacramento 1 2017.01.01
San Francisco 1 2005.10.11 1 2014.04.04 1 2014.04.04
Santa Clara County 1 2012.05.01
Colorado State 1 2012.08.08
Denver 1 2016.07.11
Connecticut State 1 2010.10.01 1 2017.01.01
Bridgeport 1 2009.10.05
Hartford 1 2009.06.12 1 2009.06.12
New Haven 1 2009.02.01 1 2009.02.01
Norwich 1 2008.12.01
Delaware State 1 2014.05.08
New Castle County 1 2014.01.28
Wilmington 1 2012.12.10
District of Columbia District of Columbia 1 2011.01.01 1 2014.07.14
Florida Broward County 1 2016.06.14
Clearwater 1 2013.01.01
Daytona Beach 1 2015.07.01
Fort Myers 1 2015.12.07
Gainesville 1 2015.11.19
Jacksonville 1 2009.07.08
Miami Dade County 1 2015.10.06
Orlando 1 2015.05.15.
Pompano Beach 1 2014.12.01
Sarasota 1 2016.05.01
St Petersburg 1 2015.01.01
Tampa 1 2013.01.14
Tallahassee 1 2015.01.28
Table 1 Ban the Box Policies implemented by December 2018
45
Georgia State 1 2015.02.23
Albany 1 2015.03.24
Atlanta 1 2013.01.01
Augusta 1 2017.01.17
Cherokee County 1 2016.03.01
Columbus 1 2015.03.29
Fulton County 1 2014.07.16
Macon Bibb County 1 2015.02.17
Hawaii State 1 1998.01.01 1 1998.01.01 1 1998.01.01
Illinois State 1 2014.01.01 1 2014.07.19 1 2014.07.19
Chicago 1 2007.06.06 1 2014.11.05 1 2014.11.05
Indiana State 1 2017.07.01
Indianapolis 1 2014.05.25 1 2014.05.25
Kansas Johnson County 1 2016.05.19
Kansas City 1 2014.11.06
Wyandotte County 1 2014.11.06
Topeka 1 2015.07.01
Wichita 1 2017.07.09
Kentucky State 1 2017.02.01
Louisville 1 2014.03.13 1 2014.03.13
Louisiana State 1 2016.08.01
Baton Rouge 1 2015.11.10
New Orleans 1 2014.01.10
Maryland State 1 2013.10.01
Baltimore 1 2007.12.01 1 2014.04.01 1 2014.04.01
Montgomery County 1 2015.01.01 1 2015.01.01 1 2015.01.01
Prince George’s County 1 2015.04.14 1 2015.04.14
Massachusetts State 1 2010.11.04 1 2010.11.04
Boston 1 2006.07.01 1 2006.07.01
Cambridge 1 2007.05.01 1 2008.01.28
Worcester 1 2009.06.23 1 2009.06.23
Michigan Ann Arbor 1 2014.05.05
Detroit 1 2010.09.13 1 2012.06.01
East Lansing 1 2014.04.15
Genesee County 1 2014.06.01
Kalamazoo 1 2010.01.01 1 2016.05.16
Muskegon County 1 2012.01.12
Minnesota State 1 2009.01.01 1 2013.05.13 1 2009.01.01
Minneapolis 1 2006.12.01
St Paul 1 2006.12.05
Missouri State 1 2016.04.11
Columbia 1 2014.12.01 1 2014.12.01 1 2014.12.01
Jackson County 1 2016.11.06
Kansas City 1 2013.04.04 1 2018.06.09
St Louis 1 2014.10.01
Nebraska State 1 2014.04.16
Nevada State 1 2018.01.01
North Las Vegas 1 2017.02.09
46
New Jersey State 1 2015.03.01 1 2015.03.01
Atlantic City 1 2011.12.23 1 2011.12.23
Newark 1 2012.09.19 1 2012.09.19 1 2012.09.19
New Mexico State 1 2010.05.19
New York State 1 2015.09.21
Albany County 1 2017.02.13
Buffalo 1 2013.06.11 1 2013.06.11 1 2013.06.11
Dutchess County 1 2016.01.19
Ithaca 1 2015.12.23
Kingston 1 2015.09.01
Newburgh 1 2015.08.10
New York City 1 2011.10.13 1 2015.10.27 1 2011.10.13
Rochester 1 2014.05.20 1 2014.05.20 1 2014.05.20
Syracuse 1 2015.03.22 1 2015.03.22
Tompkins County 1 2016.07.05
Ulster County 1 2015.01.01
Woodstock 1 2014.11.18
Yonkers 1 2014.11.01
North Carolina Ashville 1 2016.01.19
Buncombe County 1 2016.04.19
Carrboro 1 2012.10.16
Charlotte 1 2014.02.28
Cumberland County 1 2011.09.06
Durham County 1 2012.10.01
Durham City 1 2011.02.01
Forsyth County 1 2018.04.12
Mecklenburg County 1 2016.03.16
Spring Lake 1 2012.06.25
Wake County 1 2016.05.01
Ohio State 1 2016.03.23
Akron 1 2013.10.29
Alliance 1 2014.12.01
Canton 1 2013.05.15
Cincinnati 1 2010.08.01
Cleveland 1 2011.09.26
Cuyahoga County 1 2012.09.30
Daytona Beach 1 2015.07.01
Franklin County 1 2012.06.19
Hamilton County 1 2012.03.01
Lucas County 1 2013.10.29
Massillon 1 2014.01.03
Newark 1 2015.07.20
Stark County 1 2013.05.01
Summit County 1 2012.09.01
Warren 1 2015.01.14
Youngstown 1 2014.03.19
Oklahoma State 1 2016.02.24
Oregon State 1 2016.01.01 1 2016.01.01
Multnomah County 1 2007.10.10
Portland 1 2014.07.09 1 2015.11.25
47
Pennsylvania State 1 2017.07.01
Allegheny County 1 2014.11.24
Allentown 1 2015.04.01
Beaver County 1 2018.01.25
Bethlehem 1 2016.03.14
Lancaster 1 2014.10.01
Philadelphia 1 2011.06.29 1 2011.06.29 1 2011.06.29
Pittsburgh 1 2012.12.17 1 2012.12.17
Reading 1 2015.03.09
Rhode Island State 1 2013.07.15 1 2013.07.15 1 2013.07.15
Providence 1 2009.04.01
South Carolina Spartanburg 1 2017.06.26
York County 1 2017.01.17
Tennessee State 1 2016.04.14
Chattanooga 1 2017.01.07
Hamilton County 1 2012.01.01
Memphis 1 2010.07.09
Nashville 1 2016.01.01
Texas Austin 1 2008.10.16 1 2016.03.24
Dallas County 1 2015.11.17
San Antonio 1 2016.12.07
Travis County 1 2008.04.15
Utah State 1 2017.05.08
Vermont State 1 2015.04.21 1 2017.07.01
Virginia State 1 2015.04.03
Alexandria 1 2014.03.19
Arlington County 1 2014.11.03
Blacksburg 1 2016.01.19
Charlottesville 1 2014.03.01
Danville 1 2014.06.03
Fairfax County 1 2014.09.23
Fredericksburg 1 2014.01.01
Harrisonburg 1 2014.08.26
Henry County 1 2016.07.01
Montgomery County 1 2016.01.26
Newport News 1 2012.10.01
Norfolk 1 2013.07.23
Petersburg 1 2013.09.03
Portsmouth 1 2013.04.01
Prince William County 1 2013.03.25
Richmond 1 2013.03.25
Roanoke 1 2015.01.01
Staunton 1 2016.02.25
Virginia Beach 1 2013.11.01
Washington State 1 2018.06.07 1 2018.06.07
Pierce County 1 2012.01.01
Seattle 1 2009.04.24 1 2013.01.01 1 2009.04.24
Spokane 1 2015.03.06 1 2018.06.14
Spokane County 1 2017.10.27
Tacoma 1 2015.06.30
Wisconsin State 1 2016.07.01
Dane County 1 2014.02.01
Madison 1 2014.09.05 1 2015.11.25
Milwaukee 1 2011.10.07
Milwaukee County 1 2011.10.07
*As the NLSY 1997 data is only available until year 2015, this paper focuses on the jurisdictions that adoted the Ban-the-
Box policy before December 2015. The start dates are the dates of when the policy was made effective in each
jurisdiction. Source: National Employment Law Project (2018) and local legislation.
48
Mean Standard Deviation Min Max
Full Sample (Individual-Years, N=170,696)
Entrepreneurship 0.083 0.275 0 1
Past Incarceration 0.043 0.203 0 1
Ln(Years of Education) 2.457 0.165 2.30 3.04
Ln(Yearly Income) 1.738 1.754 -6.91 5.70
Ln(Family Income) 3.644 1.518 -2.00 6.91
Number of Months Worked 6.496 4.586 012
MSA Unemployment Rate 6.155 2.633 127.80
Recidivism (Re-incarceration) 0.204 0.403 0 1
Re-arrest 0.334 0.472 0 1
Years Since Release from Incarceration 4.029 3.753 018
Number of Years Incarcerated 1.052 1.463 0.083 11.5
Table 2 Descriptive Statistics of Main Variables used in OLS Regressions, 1997 to 2015
Sub Sample of Formerly Incarcerated Individuals (Individual-Years, N=7,369)
49
Full Sample
Never Incarcerated Formerly Incarcerated
Full Sample
Past Incarceration 6.7%
Entrepreneurship 11.4% 11.0% 19.1%
Years of Education 12.74 12.85 11.49
ASVAB Ability Test 55.25 56.23 42.81
Age 27.99 27.98 28.16
Yearly Income in Dollars 25,272 26,081 16,669
Family Income in Dollars 90,224 93,466 49,613
Number of Months Worked 7.36 7.55 6.26
MSA Unemployment Rate 8.63% 8.61% 8.96%
Number of Observations 8,984 8,293 491
Sub-Sample: Black Population
Past Incarceration 8.8%
Entrepreneurship 10.9% 10.4% 18.8%
Years of Education 12.35 12.47 11.37
ASVAB Ability Test 42.91 43.51 36.24
Age 28.03 28.03 28.08
Yearly Income in Dollars 19,193 20,340 9,810
Family Income in Dollars 60,022 63,229 37,460
Number of Months Worked 6.55 6.93 4.63
MSA Unemployment Rate 8.63% 8.62% 8.71%
Number of Observations 2,335 2,077 165
Sub-Sample: White Population
Past Incarceration 5.5%
Entrepreneurship 12.1% 11.6% 20.5%
Years of Education 13.15 13.25 11.69
ASVAB Ability Test 64.16 65.02 50.51
Age 27.98 27.97 28.20
Yearly Income in Dollars 28,625 29,215 20,771
Family Income in Dollars 106,158 108,921 52,841
Number of Months Worked 7.74 7.83 7.18
MSA Unemployment Rate 8.33% 8.30% 8.69%
Number of Observations 4,665 4,387 215
*Notes: Individuals who are not included in the "Never Incarcerated" and "Formerly Incarcerated" groups are
individuals who are currently incarcerated.
Table 3 Descriptive Statistics of Main Variables, 2010
50
Incorporated
Self
Employment
Self
Employment
with Employees
(1) (2) (3) (4) (5)
Past Incarceration 0.059*** 0.056*** 0.042*** 0.005* 0.023***
(0.004) (0.005) (0.006) (0.002) (0.003)
Ln(Years of Education) -0.017** 0.017** 0.012*** -0.004
(0.006) (0.006) (0.002) (0.003)
Ln(Yearly Income) (t-1) -0.010*** -0.009*** -0.000 -0.001***
(0.001) (0.001) (0.000) (0.000)
Ln(Family Income) (t-1) 0.001 0.002** 0.001** 0.001***
(0.001) (0.001) (0.000) (0.000)
Number of Months Worked (t-1) 0.005*** 0.004*** 0.001*** 0.001***
(0.000) (0.000) (0.000) (0.000)
MSA Unemployment Rate 0.003*** 0.002*** 0.000* 0.001*
(0.000) (0.000) (0.000) (0.000)
Female -0.017***
(0.001)
Black -0.006**
(0.002)
Hispanic -0.016***
(0.002)
Asian -0.022***
(0.005)
Mixed Race -0.010
(0.007)
ASVAB Ability Test 0.000***
(0.000)
Individual Fixed Effects N N Y Y Y
Year Fixed Effects Y Y Y Y Y
MSA Fixed Effects Y Y Y Y Y
N 167,812 158,827 158,826 158,826 158,826
adj. R-sq 0.055 0.055 0.361 0.136 0.272
Entrepreneurship (Self Employment)
Table 4 Unstandardized Coefficients from OLS Regression of Entrepreneurship on Incarceration
Dependent Variable
* Note: All models exclude observations of individuals who are currently incarcerated at time t. Robust standard
errors are used in these models. + p< 0.1 * p< 0.05 ** p<0.01 *** p<0.001
51
(1) (2) (3)
Black White
Past Incarceration -0.523*** -0.838*** -0.488**
(0.097) (0.166) (0.162)
Ban-the-Box 0.055 -0.218* 0.070
(0.049) (0.091) (0.068)
Past Incarceration * Ban-the-Box 0.398* 0.847** 0.030
(0.201) (0.317) (0.332)
Ln(Years of Education) 2.296*** 2.014*** 2.704***
(0.112) (0.230) (0.154)
Ln(Yearly Income) (t-1) 0.731*** 0.606*** 0.811***
(0.012) (0.022) (0.017)
Ln(Family Income) (t-1) 0.054*** 0.120*** -0.004
(0.009) (0.015) (0.014)
MSA Unemployment Rate -0.055*** -0.104*** -0.037**
(0.008) (0.018) (0.012)
Individual Fixed Effects Y Y Y
Year Fixed Effects Y Y Y
MSA Fixed Effects Y Y Y
N 135583 35798 67710
adj. R-sq 0.452 0.479 0.438
Table 5 Unstandardized Coefficients from OLS Regression of Number of Months in Paid-
Employment on Incarceration and Ban-the-Box Policy Implementation
Dependent Variable
Number of Months in Paid-Employment
Sub-sample by Race
* Note: All models exclude observations of individuals who are currently incarcerated at time t. Robust
standard errors are used in these models. + p< 0.1 * p< 0.05 ** p<0.01 *** p<0.001
Full Sample
52
(1) (2) (3)
Black White
Past Incarceration 0.049*** 0.043*** 0.052***
(0.007) (0.012) (0.012)
Ban-the-Box 0.004 0.017** 0.003
(0.004) (0.007) (0.005)
Past Incarceration * Ban-the-Box 0.017 -0.048* 0.023
(0.016) (0.024) (0.028)
Ln(Years of Education) 0.001 0.013 -0.008
(0.007) (0.014) (0.010)
Ln(Yearly Income) (t-1) -0.010*** -0.006*** -0.011***
(0.001) (0.001) (0.001)
Ln(Family Income) (t-1) 0.002*** -0.000 0.004***
(0.001) (0.001) (0.001)
Number of Months Worked (t-1) 0.004*** 0.005*** 0.004***
(0.000) (0.000) (0.000)
MSA Unemployment Rate 0.000 0.004** -0.000
(0.001) (0.001) (0.001)
Individual Fixed Effects Y Y Y
Year Fixed Effects Y Y Y
MSA Fixed Effects Y Y Y
N 135583 35798 67710
adj. R-sq 0.355 0.374 0.360
Table 6 Unstandardized Coefficients from OLS Regression of Entrepreneurship on Incarceration
and Ban-the-Box Policy Implementation
Dependent Variable
Entrepreneurship (Self-employment)
Sub-sample by Race
* Note: All models exclude observations of individuals who are currently incarcerated at time t. Robust
standard errors are used in these models. + p< 0.1 * p< 0.05 ** p<0.01 *** p<0.001
Full Sample
53
(1) (2) (3)
Black White
Past Incarceration -7016.739*** -6767.689*** -6512.958***
(274.224) (383.735) (468.748)
Entrepreneurship -890.962*** -929.227** -1127.093***
(198.709) (353.921) (284.607)
Past Incarceration * Entrepreneurship 3582.647*** 1849.050+ 3688.572**
(869.949) (1009.618) (1328.286)
Ln(Years of Education) 33878.350*** 29568.929*** 33901.349***
(518.398) (984.791) (722.403)
Number of Months Worked 705.144*** 578.955*** 803.352***
(10.874) (17.640) (17.094)
MSA Unemployment Rate -96.118** -249.590*** -54.865
(29.285) (55.586) (48.551)
Individual Fixed Effects Y Y Y
Year Fixed Effects Y Y Y
MSA Fixed Effects Y Y Y
N 147778 38745 75876
adj. R-sq 0.618 0.602 0.630
Table 7 Unstandardized Coefficients from OLS Regression of Yearly Income on Incarceration
and Entrepreneurship
Dependent Variable
* Note: All models exclude observations of individuals who are currently incarcerated at time t. Robust
standard errors are used in these models. + p< 0.1 * p< 0.05 ** p<0.01 *** p<0.001
Yearly Income in Dollars
Sub-sample by Race
Full Sample
54
Re-arrest
(1) (2) (3) (4)
Black White
Entrepreneurship -0.053*** -0.061* 0.003 -0.064***
(0.012) (0.024) (0.016) (0.016)
Years Since Release from Incarceration -0.038*** -0.047*** -0.035*** 0.005*
(0.002) (0.003) (0.002) (0.002)
Number of Years Incarcerated 0.046*** 0.014* 0.075*** 0.033***
(0.005) (0.007) (0.009) (0.005)
Ln(Yearly Income) (t-1) 0.001 0.011 -0.005 -0.011*
(0.004) (0.007) (0.006) (0.005)
Ln(Years of Education) -0.028 0.121 0.052 -0.055
(0.043) (0.088) (0.059) (0.055)
Number of Months Worked (t-1) -0.002+ -0.000 -0.004* -0.002+
(0.001) (0.002) (0.002) (0.001)
Ln(Family Income) (t-1) 0.007* 0.003 0.002 0.001
(0.003) (0.004) (0.005) (0.003)
MSA Unemployment Rate 0.006+ -0.004 0.012** -0.001
(0.003) (0.006) (0.005) (0.004)
Individual Fixed Effects Y Y Y Y
Year Fixed Effects Y Y Y Y
MSA Fixed Effects Y Y Y Y
N 7243 2436 3114 7243
adj. R-sq 0.326 0.392 0.457 0.217
Sub-sample by Race
Full Sample
Full Sample
* Note: This sample includes only formerly incarcerated individuals. All models exclude observations of
individuals who are currently incarcerated or unemployed at time t. Robust standard errors are used in these
models. + p< 0.1 * p< 0.05 ** p<0.01 *** p<0.001
Table 8 Unstandardized Coefficients from OLS Regression of Recidivsm on Entrepreneurship
Dependent Variable
Recidivism (Re-incarceration)
55
Figure 1 Employment Trend of Formerly Incarcerated Individuals Before and After “Ban-the-Box” Policy
Implementation
* The dashed line shows the relative number of months employed for formerly incarcerated individuals in states and
counties where Ban-the-Box was implemented in year T0, compared to the baseline (solid horizontal line normalized
to zero) which represents the employment of formerly incarcerated individuals in states and counties where Ban-the-
Box was not implemented. The figure shows the employment trend of formerly incarcerated individuals in Ban-the-
Box states and counties relative to formerly incarcerated individuals in non Ban-the-Box states and counties.
-.5 0.5 11.5 2
Number of Months Employed
T-2 T-1 Ban-the-Box T+1 T+2 T+3 T+4
56
Figure 2 Entrepreneurship Trend of Formerly Incarcerated Individuals Before and After “Ban-the-Box” Policy
Implementation
* The dashed line shows the relative probability of entrepreneurship for formerly incarcerated individuals in states
and counties where Ban-the-Box was implemented in year T0, compared to the baseline (solid horizontal line
normalized to zero) which represents the probability of entrepreneurship of formerly incarcerated individuals in
states and counties where Ban-the-Box was not implemented. The figure shows the entrepreneurship trend of formerly
incarcerated individuals in Ban-the-Box states and counties relative to formerly incarcerated individuals in non Ban-
the-Box states and counties.
-.15 -.1 -.05
0.05
Probability of Entrepreneurship
T-2 T-1 Ban-the-Box T+1 T+2 T+3 T+4
57
(1) (2) (3)
PSM Matching CEM Matching
Individuals with
Arrest Records
Past Incarceration 0.035*** 0.038*** 0.031***
(0.007) (0.006) (0.006)
Ln(Years of Education) 0.048* 0.034*** 0.022+
(0.023) (0.009) (0.014)
Ln(Yearly Income) (t-1) -0.002 -0.009*** -0.008***
(0.002) (0.001) (0.001)
Ln(Family Income) (t-1) 0.002 0.001 0.002*
(0.001) (0.001) (0.001)
Number of Months Worked (t-1) 0.005*** 0.004*** 0.005***
(0.001) (0.000) (0.000)
MSA Unemployment Rate 0.003** 0.003*** 0.002**
(0.001) (0.001) (0.001)
Individual Fixed Effects Y Y Y
Year Fixed Effects Y Y Y
MSA Fixed Effects Y Y Y
N 25108 109528 52955
adj. R-sq 0.407 0.363 0.363
Appendix A Unstandardized Coefficients from Matched Sample OLS Regression of Entrepreneurship on
Incarceration
Dependent Variable
Entrepreneurship (Self Employment)
* Note: All models exclude observations of individuals who are currently incarcerated at time t. Robust standard errors
are used in these models. + p< 0.1 * p< 0.05 ** p<0.01 *** p<0.001
ResearchGate has not been able to resolve any citations for this publication.
Article
Full-text available
Ban the Box (BTB) policies mandate deferred access to criminal history until later in the hiring process. However, these policies chiefly target public employers. The study is the first to focus on the primary goal of BTB reform, by measuring the impact of BTB policies on the probability of public employment for those with convictions. To execute the analyses, the study uses data from the National Longitudinal Survey of Youth 1997 Cohort (2005–2015) and difference‐in‐difference (DD) estimation. The study finds that BTB policies raise the probability of public employment for those with convictions by about 30% on average. Some scholars argue that BTB policies encourage statistical discrimination against young low‐skilled minority males. The study employs triple‐difference (DDD) estimation to test for statistical discrimination, but uncovers no evidence to support the hypothesis. (JEL J15, J71, J78, K4).
Book
http://ontheoutsidebook.us America’s high incarceration rates are a well-known facet of contemporary political conversations. Mentioned far less often is what happens to the nearly 700,000 former prisoners who rejoin society each year. On the Outside examines the lives of 22 people—varied in race and gender but united by their time in the criminal justice system—as they pass out of the prison gates and back into society. The book takes a clear-eyed look at the challenges faced by former prisoners as they try to find work, housing, and stable communities. Standing alongside these individual portraits is a substantial quantitative study conducted by the authors that followed every state prisoner in Michigan who was released on parole in 2003 (roughly 11,000 individuals) for the next seven years, providing a comprehensive view of their post-prison education, neighborhoods, families, employment, and contact with the parole system. On the Outside delivers a powerful combination of hard data and personal narrative that shows why our country continues to struggle with the social and economic reintegration of the formerly incarcerated.
Book
Persistent thieves, criminals who resume committing crimes of burglary, robbery, vehicle theft, and ordinary theft despite previous attempts to stop, are a main focal point of American criminology and criminal justice. Cast as career criminals,“they are also one of the principal targets of the war on crime” that American governments have waged for more than two decades. Building on a theoretical interpretation of crime as choice, crime-control policies and programs justified by notions of deterrence and incapacitation have proliferated. America's urban police departments now have repeat offender units,” and many of the new state sentencing codes mandate lengthy sentences for defendants with previous convictions. Great Pretenders is based on the author's original studies and previously published research and on more than fifty autobiographies of persistent thieves. Shover uses a crime-as-choice framework and a life-course perspective to make sense of important decisions and changes in the lives of persistent thieves. He shows how the working-class origins of most persistent thieves produce both low legitimate and low criminal aspirations, even as those origins leave them ill equipped to exploit comparatively safe, lucrative, and newer forms of criminal opportunity.In this book Shover describes how many persistent thieves and hustlers identify with crime and pursue a lifestyle of life as party in which their choices alternately are made in contexts of drug-using hedonism or desperation. Their estimates of the likely payoffs from crime are severely distorted, and most give little thought to possible arrest. As they get older, however, persistent thieves make qualitative changes in the crimes they commit, and many eventually stop committing crimes altogether. The author highlights some unintended consequences of harsh crime control measures and raises critical questions about the one-size-fits-all approach to crime of recent decades.
Article
"Ban the Box" (BTB) policies restrict employers from asking about applicants' criminal histories on job applications and are often presented as a means of reducing unemployment among black men, who disproportionately have criminal records. However, withholding information about criminal records could risk encouraging racial discrimination: employers may make assumptions about criminality based on the applicant's race. To investigate BTB's effects, we sent approximately 15,000 online job applications on behalf of fictitious young, male applicants to employers in New Jersey and New York City before and after the adoption of BTB policies. These applications varied whether the applicant had a distinctly black or distinctly white name and the felony conviction status of the applicant. We confirm that criminal records are a major barrier to employment: employers that asked about criminal records were 63% more likely to call applicants with no record. However, our results support the concern that BTB policies encourage racial discrimination: the black-white gap in callbacks grew dramatically at companies that removed the box after the policy went into effect. Before BTB, white applicants to employers with the box received 7% more callbacks than similar black applicants, but BTB increased this gap to 43%. We believe that the best interpretation of these results is that employers are relying on exaggerated impressions of real-world racial differences in felony conviction rates.
Article
The past several decades have seen a decline in employment rates and labor force participation, particularly among low-skilled, minority men living in poor areas. As low-skill jobs disappear from poor places, how do marginalized job seekers navigate this landscape? Using over 8,000 daily measures of search and work collected from smartphones distributed to 133 men recently released from prison, this article presents the concept of work as foraging, where people work a variety of extremely precarious opportunities that span across job types. Sequence analysis methods describe distinct patterns of search and work that unfold over time, where most people cease their search efforts after the first month and maintain a state of very irregular and varied work. Although there is substantial heterogeneity in patterns, foraging is a common strategy of survival work.