Academic journal article The American Journal of Economics and Sociology

The Economics of the Criminal Behavior of Young Adults: Estimation of an Economic Model of Crime with a Correction for Aggregate Market and Public Policy Variables

Academic journal article The American Journal of Economics and Sociology

The Economics of the Criminal Behavior of Young Adults: Estimation of an Economic Model of Crime with a Correction for Aggregate Market and Public Policy Variables

Article excerpt

MARK GIUS [*]

ABSTRACT. This study uses a combination of individual-level and county-level data to estimate an economic model of crime for young adults similar to that used by Becker (1968) and Trumbull (1989). In order to estimate a model of crime in which both individual-level and county-level data are used, it is necessary to take account of the bias introduced by using aggregate-level data in conjunction with individual-level data. In order to eliminate this bias, a technique derived by Moulton (1990) is employed. Results from a logit regression model indicate that race, sex, and peer pressure have statistically significant effects on the probability that a young adult will commit a crime. Results also suggest that police presence, as measured by county-level per capita police expenditures, does not deter young adults from committing crimes.

I

Introduction

IN 1968, BECKER PUBLISHED ONE of the first articles that dealt with the economics of crime. Since then, there has been a plethora of work done on the economics of crime (Ehrlich 1973, 1975; Ehrlich and Brower 1987; Block and Heineke 1975; Block and Lind 1975a, 1975b; Carr-Hill and Stern 1973; Cover and Thistle 1988; Layson 1985; McCormick and Tollison 1984; Myers 1980, 1983; Orsagh 1973; Passell and Taylor 1977; Phillips and Votey 1975; Pirog-Good 1986; Sandelin and Skogh 1986; Schmidt and Witte 1984; Sjoquist 1973; Viscusi 1986; Witte 1980, 1983; Trumbull 1989; Hsing 1995; Britt 1994; Meera and Jayakumar 1995; Young 1993; Hull and Bold 1995; Deutsch, Hakim, and Spiegel 1990; Leung 1992; Howsen and Jarrell 1987; Benson, Kim, Rasmussen, and Zuehlke 1992).

Most of these studies attempt to ascertain the determinants of criminal behavior and/or crime rates. In addition, most of these studies use as their data aggregate measures of crime, such as county, state, or even national-level crime statistics, and aggregate measures of socioeconomic characteristics.

While many of these studies have shed light on the effects of socioeconomic characteristics and institutional factors on criminal activities, there has been some criticism of the use of aggregate data to model criminal behavior, a behavior that is an individual choice about lifestyle and income-generating opportunities. Two prior studies have used individual-level data to estimate economic models of crime (Witte 1980; Trumbull 1989). Unfortunately, both of these studies examine the criminal behavior of former prison inmates; hence their analyses and results cannot be used to determine if various socioeconomic and institutional variables have any effect on the likelihood of non-convicts committing criminal acts.

Another problem with these prior studies is that their crime statistic data do not capture all criminal activity; their data only capture reported criminal activity. Hence, aggregate studies that use reported crime statistics not only ignore the individualistic nature of crime, but also seriously underreport the level of criminal activity within any given jurisdiction.

In order to overcome the above problems, the current study estimates an economic model of crime for young adults using a data set that consists of both individual-level and aggregate variables. The individual-level data is obtained from a data set that has not been previously used for such an analysis: the National Longitudinal Survey Youth-Geocode (NLSY-Geocode) data set. This data set is uniquely suited to the estimation of an economic model of crime for young adults for two important reasons. First, the NLSY-Geocode contains a vast amount of personal information about the survey respondents, including information about any criminal activities they may have committed. Second, the Geocode data is important since this data set identifies the state and county where respondents live. Using this information in conjunction with criminal justice system data provided by the US Department of Justice, it may be possible to proxy any potential deterrent effect that police presence may have on the criminal behavior of young adults. …

Search by... Author
Show... All Results Primary Sources Peer-reviewed

Oops!

An unknown error has occurred. Please click the button below to reload the page. If the problem persists, please try again in a little while.