HOW AND WHY DOES WORK MATTER? EMPLOYMENT CONDITIONS, ROUTINE ACTIVITIES, AND CRIME AMONG ADULT MALE OFFENDERS*

• Published date: 01 May 2017
• Date: 01 May 2017
• DOI: http://doi.org/10.1111/1745-9125.12134

HOW AND WHY DOES WORK MATTER?

EMPLOYMENT CONDITIONS, ROUTINE ACTIVITIES,

AND CRIME AMONG ADULT MALE OFFENDERS∗

ROBERT APEL1and JULIE HORNEY2

1School of Criminal Justice, Rutgers University

2Department of Sociology and Criminology, The Pennsylvania State University

KEYWORDS: employment, work quality, routine activities, life event calendar, media-

tion analysis

An inverse relationship between employment and crime is well established, although

the mechanisms that account for the correlation remain poorly understood. In the cur-

rent study, we investigate the role of work quality, measured objectively (hours, in-

come) as well as subjectively (commitment). A routine activities perspective is pro-

posed for the work–crime relationship, and it inspires hypotheses about the way that

work reduces crime indirectly, in part, through unstructured leisure and substance-

using behaviors that tend to carry situational inducements to offend. The results derive

from within-person analyses of monthly data provided by adult male offenders re-

cently admitted to state prison in the Second Nebraska Inmate Study (N=717; NT =

21,965). The findings indicate that employment significantly reduces self-report crime

but only when employed men report strong commitment to their jobs, whereas other

work characteristics are unrelated to crime. This indicates that, among serious crimi-

nally involved men, the subjective experience of work takes priority over its objective

characteristics. The results also indicate that routine activities only partly mediate the

relationship among work, job commitment, and crime, whereas the majority of the

work–crime relationship remains unmediated.

It is often taken for granted that employment is strongly correlated with criminal behav-

ior, and the results of a variety of studies have fairly consistently found it to be associated

with individual-level reductions in crime. For example, this has been found in nation-

ally representative samples such as the National Longitudinal Survey of Youth (NLSY;

∗Additional supporting information can be found in the listing for this article in the Wiley Online

Library at http://onlinelibrary.wiley.com/doi/10.1111/crim.2017.55.issue-2/issuetoc.

The authors wish to thank the editor of this article, Eric Baumer, along with the anonymous re-

viewers. The article has been greatly expanded and improved based on their many insights.

A note of dedication from Robert Apel: At the time this article was in production, we were stunned

and deeply saddened to learn of the loss of Julie Horney, a beloved friend and colleague to so many

in the discipline. I know that over the years many scholars have benefited as much as I have from

her career counsel and the feeling that she always had our best professional interests at heart.

Julie’s absence will be sorely felt, but this article will bring fond memories of what a pleasure it was

to be her colleague, collaborator, and friend. I hope that it honors the deep imprint she has left on

criminology, in terms of both her leadership as well as her scholarship.

Direct correspondence to Robert Apel, School of Criminal Justice, 123 Washington Street,

Newark, NJ 07102 (e-mail: robert.apel@rutgers.edu).

C2017 American Society of Criminology doi: 10.1111/1745-9125.12134

CRIMINOLOGY Volume 55 Number 2 307–343 2017 307

308 APEL & HORNEY

Apel et al., 2008; Crutchfield and Pitchford, 1997; Grogger, 1998; Wadsworth, 2006) and

Monitoring the Future (Staff, Osgood, et al., 2010; Staff, Schulenberg, et al., 2010); with

population birth cohorts (Thornberry and Christenson, 1984; Williams and Sickles, 2002;

Witte and Tauchen, 1994); in follow-up studies of incarcerated individuals (Laub and

Sampson, 2003; Sampson and Laub, 1993; Witte, 1980); among homeless samples (Hagan

and McCarthy, 1997); and in randomized evaluations comprising hard-to-employ individ-

uals with arrest histories or prison experience (Cook et al., 2015; Uggen, 2000; Uggen and

Thompson, 2003). This finding has also been replicated in a variety of non-U.S. settings

(Blokland and Nieuwbeerta, 2005; Farrington et al., 1986; Hagan, 1993; Savolainen, 2009;

Van der Geest, Bijleveld, and Blokland, 2011; Verbruggen et al., 2015).1

Despite a good deal of evidence showing that work can play an important role in

curtailing criminal behavior, surprisingly little is known about the mechanisms through

which work lowers individual involvement in crime, especially among serious offenders.

In the current study, we aim to advance our understanding of the relationship between

work and crime through an analysis of employment and offending histories in a sample

of males sentenced to a state correctional institution. The data are 36-month life event

calendars from the Second Nebraska Inmate Study, modeled after an earlier study

by Horney and colleagues (Horney and Marshall, 1991, 1992; Horney, Osgood, and

Marshall, 1995). By employing an approach that allows us to focus on within-person

change, we ask whether criminal involvement varies concomitantly with the objective

characteristics of particular jobs held (hours and income), and whether the extent of

criminal involvement depends on the degree of personal commitment to those jobs. We

also provide an empirical assessment of a routine activities perspective by determining

whether observed employment effects are mediated to any significant degree by changes

in the amount of time spent in activities that results from prior studies suggest provide

situational inducements to crime. We propose broadening the routine activities per-

spective to include not only standard leisure-time activities but also substance use. This

theoretical expansion might help understand how work may or may not be correlated

with crime among criminally involved men.

SPECIFICATION OF THE WORK–CRIME RELATIONSHIP

According to most theoretical accounts, employment provides a means of strengthen-

ing prosocial and law-abiding behavior. Rational choice theory suggests that work expe-

rience is an important source of human capital that increases the expected returns from

law-abiding behavior (Becker, 1968; Lochner, 2004). Social control theory posits that em-

ployment, in general, but high-quality jobs, in particular, are a potent source of social

capital that can override natural temptations to violate the law (Hirschi, 1969; Sampson

and Laub, 1993). Strain theory proposes that lack of success in the labor market can mo-

tivate individuals to “innovate” through criminal behavior as a way to acquire valued

1. The evidence does not universally favor the crime-reducing benefits of work. The outcomes of

other studies, many of them randomized evaluations of work programs, have found inconsistent or

inconclusive effects of work on recidivism, at least for certain populations (Farabee, Zhang, and

Wright, 2014; Redcross et al., 2010; Skardhamer and Savolainen, 2014; Visher, Winterfield, and

Coggeshall, 2005; Wilson, Gallagher, and MacKenzie, 2000; Zweig, Yahner, and Redcross, 2011).

Bushway and Apel (2012) discussed some of the policy implications of these findings.

HOW AND WHY DOES WORK MATTER? 309

goals or to alleviate negative feelings (Agnew, 1992; Merton, 1938). Interactionist theory

proposes that employment, but especially career-oriented work, coincides with transfor-

mations in reference groups as well as in the meaning of work that reduce the attractions

of crime (Matsueda, 1992; Mead, 1934). Learning theory suggests that the workplace pro-

vides a forum for exposure to a variety of models, definitions, and reinforcements that are

conducive to law-abiding behavior (Akers, 1985; Sutherland, 1947).

In the sections that follow, we elaborate on evidence consistent with the rational choice

and informal social control perspectives, which have been the most clearly articulated

perspectives on the work–crime relationship and which have several commonalities. We

then propose integrating these with the routine activities perspective, with sensitivity to

the way that employment alters time use outside of the workplace in ways that channel

individuals into better supervised settings with law-abiding companions (see Cohen and

Felson, 1979; Osgood et al., 1996). By building on the insights of Crutchfield (1989: 494),

work could influence crime through a “situation of company” whereby the unemployed

or those employed in low-quality jobs “spend significant amounts of time with nothing to

do...[i]tislikelythattheyfrequently pass time in local taverns, pool halls, and on street

corners.”

INFORMAL SOCIAL CONTROL EXPLANATIONS

According to rational choice accounts, individuals who are unemployed or employed

in low-wage jobs experience a lower “price” of criminal activity, meaning that the ex-

pected returns from illegal behavior, discounted by punishment risk (and other things

being equal), are more likely to outweigh the expected returns from legal income earn-

ing.2From a rational choice perspective, then, we would predict that individuals working

longer hours or earning higher income, because these increase the expected monetary

returns from work, are less likely to be involved in criminal activity.

Hirschi’s (1969) informal social control theory identifies the source of crime in a weak-

ening of social ties to the conventional order, emphasizing constraints that are affective

(“attachment”), rational (“commitment”), temporal (“involvement”), and moral (“be-

lief”). Sampson and Laub’s (1993) age-graded theory gives informal social control a life-

course rendering, emphasizing the accumulation of social capital as individuals undergo

conventional life transitions.3Just as they have suggested that a “good marriage,” and not

just being married, is a precondition for desistance from crime, they similarly assert that

employment per se does not provide the social capital that would lead to a reduction in of-

fending. Rather, employment should only produce effective social control “when coupled

with job stability, job commitment, and mutual ties to work” (p. 140).

2. To be clear, rational choice as imported from economics is not a theory of cognition, and although

at various points in our discussion we refer to it as a theory, the truth is that we prefer to think of it

as a model rather than as a theory of behavior (for a good discussion of models, see Solow, 1997).

At any rate, true to its utilitarian roots, it is a normative account of how people should behave,

given stable and well-ordered preferences and constraints on opportunities and resources.

3. Hirschi’s (1969) theory is static and largely limited to the explanation of adolescent misconduct.

Sampson and Laub’s (1993) theory, on the other hand, incorporates a dynamic component that

explains both stability and change in law-violating behavior across the life course. In a rough man-

ner of speaking, Hirschi’s theory of adolescent social bonds speaks to why people start committing

crime, whereas Sampson and Laub’s theory of adult social bonds explains why they stop.