Welfare Programs and Property Crime
Crime is a major concern in many Latin American countries. As Fajnzylber et al (1998) state, homicide rates in the region are amongst the highest in the world. High crime rates have pernicious effects on the economy and on the quality of life. In this sense, research on policies which are effective in reducing crime are of utmost relevance.
Seminal studies postulating an economic model of criminal behavior (Becker, 1968 and Ehrlich, 1973) claim that incentives are important for an individual to engage in criminal activity. Becker’s original model states that the number of offenses and individual would commit during a particular period is a function of the probability of conviction per offense, her punishment per offense and other variables such as income generated in legal activities, law-abidingness, among others.
Amongst the main predictions of classical static models of crime, we have that crime will decrease when the probability of conviction increases and when the punishment is harsher. While the deterrence effect of tougher sentences or an increase in the probability of conviction has been assessed in developed countries, evidence for Latin America is scarce. Evidence for developed countries yield a wide range of crime elasticities with respect to the probability of conviction and sentences , ranging from -0.06 to -0.74.
As far as other factors which may affect crime, such as opportunities in legal activities, income inequality, social programs, etc, evidence is even scarcer and not very conclusive either.
Our work measures the causal effect of the introduction of a workfare program on crime. In general, it is very difficult to establish such causality, due to the existence of unobservable variables jointly determining crime rates and participation in social programs. It is often the case that areas stricken by poverty and unemployment have also a higher part of the population whose income depends on welfare. The evidence of social experiments is very scarce and in general, the scope for conducting such experiments has so far been limited.
We can expect a workfare program to have an effect on crime through several channels. First, if crime is motivated to supplement income, then workfare should have a negative effect, since workfare programs are generally associated with a monetary subsidy. Secondly, if work requirements of the program are enforced, then individuals would have less time to commit crimes, what is known in the literature as incapacitation effect. Finally, if recipients are not longer eligible for workfare once they are apprehended or convicted, workfare may have a dynamic deterrent effect on crime. If workfare programs do have an effect on crime, then total welfare effects of workfare programs may be different to the ones commonly analyzed in the literature. Several authors in the criminalistic field have studied the relationship between crime and payments associated with social programs using cross sectional data. DeFranzo (1996, 1997) and Hannon and DeFranzo's (1998a, 1998b) results indicate social programs reduce most severe crimes. Machin and Marie (2004) study the effect of a reduction in unemployment benefits in England and find a positive effect on crime. Finally, Foley (2008) shows that timing in the payment of benefits of welfare programs matters for criminal activity, especially for minor crimes. More staggered and frequent welfare payments are associated with a reduction in property crime.
As mentioned above, the main empirical problem which arises when studying the causal effect of welfare programs on crime is the existence of unobservables affecting the allocation of such programs and crime rates within a community. It is highly probable that a community where there is more population depending on social programs has also other unobserved characteristics associated with crime rates. In this sense, the number of beneficiaries of social programs for a given region is endogenous and naive regression coefficients are biased.
In this paper, we study the effect of the introduction of an important workfare program, Plan Jefes y Jefas de Hogar Desocupados, henceforth PJH, on property crime for the case of Argentina. In order to circumvent the endogeneity problem arising from unobserved factors affecting both crime and the demand for workfare, we make use of instrumental variables estimation. We show that the workfare program was assigned in response to political demands, which are unrelated to crime. Our identification strategy is based on a widely supported fact by the political economy literature: incumbent governments target spending to enhance their electoral prospects. We show that there is a strong correlation between the support for government and the granting of subsidies, and that the percentage of government support in previous elections is a valid instrument for the number of per capita workfare programs across districts. There is evidence on the use of social programs to influence voters in different Latin American countries For the case of Mexico, Green (2006) shows that there were small effects of political biases affecting the distribution of PROGRESA. , even when this program had clear eligibility rules. Manacorda et al. (2009) show that Uruguay's PANES, similar to PROGRESA, which was implemented in response to the economic crisis of 2003 strongly influenced votes in the following election.
Argentina is a middle income country located in South America, which enjoyed relatively high standards of living until the late seventies. Since 1995, property crimes have shown an increasing trend in Argentina. While there is no conclusive agreement about what has caused such increase, Garzette (2003) shows that property crime worked as a redistributive tool in the context of a worsening in socioeconomic conditions among some sectors of the population, specially the lack of opportunities in the labor market. If this hypothesis is true, then any social program working as an income maintenance program would reduce crime. Nevertheless, we cannot expect this only effect will prevail, since we can also expect that freeing time from income generating activities will also increase the amount of time available to engage in criminal activities. We make use of a special feature of the way PJH was granted across counties in order to overcome the endogeneity problem.
The paper shows that the number of per capita PJH granted in each jurisdiction was related to votes in the previous elections and not according to objective criteria for program inclusion, e.g.: poverty, unemployment, etc. We can therefore instrument the number of PJH in each region with the number of votes for the government.
Our instrumental variables estimates find strong effects of PJH on crime. Elasticity estimates are in the range of -0.7, implying an increase of 1% in the number of workfare programs per 100,000 inhabitants reduces property crime by 0.7%. There is no statistically significant effect on other type of crimes.
We conduct some robustness tests in order to rule out other potential effects behind our results. One concern might be that provinces with high government support may have systematic differences in law enforcement institutions. If this were the case, then we would find an effect for all crimes. We do not find such effect for non-property crimes. Moreover, we control our estimations for a series of observable variables which differ across provinces. Finally, we run our estimations on pre-program data and find no statistically or economically significant effect.
Cost effective policy efforts to reduce crime result in better quality of life and positive effects on economic activity, especially in countries where crime rates are high, as in Latin America. Crime literature provides some directions where efforts should be focused, though much more research in the region is needed. We showed that a welfare program introduced in a time of crisis in Argentina has negative effects on property crime. Such results are obtained using instrumental variables and exploring the fact that program assignment was not related to objective eligibility criteria but was used as a way to reward political loyalty. Our identification strategy is based on a widely supported fact by the political economy literature: incumbent governments target spending to enhance their electoral prospects. We show that there is a strong correlation between the support for government and the granting of subsidies, and that the percentage of government support in previous elections is a valid instrument for the number of per capita workfare programs across districts.
Our results support the evidence on the importance of economic incentives on crime, hinting at different mechanisms through which welfare programs can operate. Moreover, it adds to the literature of unintended consequences of social programs, which potential implications for welfare calculations. This is particularly relevant for Latin America and the design of its social policy.
Alzua, M.L, (2010), “Welfare Programs and Property Crime”, CEDLAS, Unpublished manuscript.
Becker, Gary S.,(1968) “Crime and Punishment: An Economic Approach,” Journal of Political Economy, 76 (2), 169–217.
Biderman, C, A. Schneider and J. Pinho de Mello, (2010), “Dry Law and Homicides: Evidence from the São Paulo Metropolitan Area,” Economic Journal, Vol. 120, pp. 157-182.
DeFranzo, J, (1996), “Welfare and Burglary”, Crime and Delinquency 42, pp 223-229.
DeFranzo, James, (1997), “Welfare and Homicide”, Journal of Research in Crime and Delinquency 34, pp. 395-406.
Di Tella, R, S. Edwards and E. Schargrodsky, (2010), “ The Economics of Crime:
Lessons For and From Latin America”, Chicago University Press.
Ehrlich, I.(1973) “Participation in Illegitimate Activities: A Theoretical and Empirical Investigation,” Journal of Political Economy, 81 (3), 521–565.
Fajnzylber , P, D. Lederman and N. Loayza, (2000), “Crime and Victimization: An Economic Perspective“, Economia , 219-278
Foley, C. F. (2010), Welfare Payments and Crime”, forthcoming in Review of Economics and Statistics
Garcette, N. (2004) “Property Crime as a Redistributive Tool: the Case of Argentina.”
Green, Tina (2006) “Essays on the Political Economy of Fiscal Policy”, PhD Dissertation, University of California at Berkeley
Hannon, L., and J..DeFranzo, (1998a), “Welfare and Property Crime, Justice Quarterly 15, pp. 273-287.
Lee, D and J. McCrary, (2009), The Deterrence Effect of Prison: Dynamic Theory and Evidence <http://www.econ.berkeley.edu/~jmccrary/lee_and_mccrary2009.pdf>
Miguel, E. M. Manacorda, and A. Vigorito (2010), “Government Transfers and Political Support” <http://elsa.berkeley.edu/~emiguel/pdfs/miguel_uruguay.pdf>