-
Crime and Victimization:An Economic Perspective
-
Comments
Footnotes
1. Shawn Bushway provided helpful comments for the discussion below.
2. For example, see Levitt (forthcoming).
3. Gottfredson and Hirschi (1990).
4. Wilson (1995).
5. Wilson (1996: xiii).
6. Uggen (1994).
7. Lynch (1995).
8. UN (1999, p.43).
9. James Lynch (personal communication).
10. The 1995 figure was 1.8 (Registered Murders in the Netherlands, Press Release, CBS Voorburg-Statistics Netherlands, 14 July 1998, cited by www.csdp.org/factbook/thenethe.htm).
11. UN (1999).
12. Statistics Sweden (2000).
13. Bennett and Lynch (1990, p.155).
14. Field (1999, p.11).
15. Schneider and Enste (2000); Tanzi (1999); Becker (1965).
16. UN (1999).
17. Thoumi (1995). Since the mid-1990s it appears that the major drug trafficking enterprises, associated with Cali and Medellín, have been replaced by a larger number of smaller enterprises, numbering perhaps in the hundreds. However, this comes after the period covered by Fajnzylber, Lederman, and Loayza.
18. MacCoun and Reuter (forthcoming).
19. Office of National Drug Control Policy (various issues).
1. This method was first proposed by Arellano and Bover (1995).
2. See, for example, the Monte Carlo evidence presented by Judson and Owen (1999).
3. See the Monte Carlo evidence presented by Kiviet (1995) and Judson and Owen (1999). Kiviet states, "We find that OLS has an impressingly small standard deviation, and therefore, when bias is moderate (which it is when the coefficient on the dependent variable is high), it has an attractive mean squared error" (1995, p. 70).
4. The same point is made by Bourguignon, who states that the coefficient on the Gini "becomes insignificant when a dummy variable is introduced for Latin America in the homicide regression" (1999, p. 22).
5. See Bourguignon (1999) for a formal model of this idea.
6. See, for example, Wilson (1987, 1996) on inner city violence in the United States; Levitt and Venkatesh (1998) on gang violence.
7. The countries included in the graph are Argentina, Bolivia, Brazil, Chile, Colombia, Costa Rica, Ecuador, El Salvador, Mexico, Nicaragua, Panama, Peru, Uruguay, and Venezuela.
8. This index, which was developed in a recent paper by Dahan and Gaviria (2000), is based on the correlation of schooling among teenage siblings: the higher this correlation, the lower the prospects of mobility.
9. Glaeser and others (1999) present experimental evidence showing that the available measures of social trust are indeed measures of trustworthiness.
References
Footnotes
1. International Centre for the Prevention of Crime (1998, chap. 3, p. 5).
2. Fajnzylber, Lederman, and Loayza (1998, pp. 11-15).
3. New York Times and CBS poll, quoted in Blumstein (1995, p. 10).
4. International Centre for the Prevention of Crime (1998, chap. 3, p. 2).
5. Polls conducted by Latino Barómetro, quoted in Londoño and Guerrero (1999, p. 6).
6. See Buvinic and Morrison (1999, technical note 4).
7. Glaeser (1999, p. 19).
8. International Centre for the Prevention of Crime (1998, chap. 2, p. 3).
9. World Bank (1993).
10. Buvinic and Morrison (1999, technical note 4).
11. Mandel and others. (1993).
12. International Centre for the Prevention of Crime (1998, chap. 2, p. 3).
13. Londoño and Guerrero (1999, p. 27).
14. Alternatively, in explaining the intangible costs of crime, the cumulative effect of relatively high levels of crime over a long period of time may be more important than the levels of crime at a given point in time. Thus the costs of crime could grow even in the context of stable or declining crime rates.
15. An approach that has not been applied to date in Latin America is that of using so-called hedonic estimates of housing prices to measure the economic costs of crime. In the United States, results from studies of this type indicate that a doubling of crime rates could lead to a reduction of 8 to 12.5 percent in real estate costs (Buvinic and Morrison, 1999). One advantage of these studies is that they generate estimates of the value of marginal reductions in the level of crime, as opposed to accounting estimates of the total costs of crime (Glaeser, 1999, p. 20). Indeed, the former may be most useful from a practical point of view, since most policy measures will not lead to a complete eradication but rather to marginal reductions of the level of crime.
16. Londoño and Guerrero (1999, p. 22).
17. Londoño and Guerrero (1999, p. 26).
18. Becker (1968).
19. Becker (1993, pp. 385-86, emphasis in original).
20. Becker (1993, p. 390).
21. Ehrlich (1973); Mathieson and Passell (1976).
22. Ehrlich (1975a).
23. Ehrlich (1981, p. 311).
24. Ehrlich (1975a, 1981); Levitt (1998a).
25. Becker (1968, p. 178); Ehrlich (1973, p. 528). The standard assumption in theoretical models is to consider individuals who are risk averse, but who exhibit decreasing risk aversion with increasing income (Schmidt and Witte, 1984, p. 161).
26. Grogger (1991). This result is also supportive of the prevalence of the deterrent vis-à-vis the incapacitation effects of imprisonment.
27. Levitt (1996, 1997, 1998a).
28. Levitt (1996).
29. Prison overcrowding lawsuits have been filed in the United States since 1965 on the grounds of unconstitutional conditions in prisons. Levitt shows that the filing of prison overcrowding litigation leads to the lowering of prison population growth rates, even before the courts reach any decision. Moreover, the status of prison overcrowding litigation is shown to be unrelated to previous crime rates.
30. Levitt (1997).
31. Levitt (1998a).
32. Posner (1995).
33. Glaeser and Sacerdote (1999b).
34. Glaeser, Kessler, and Piehl (1998).
35. Fleisher (1966); Ehrlich (1973).
36. Fleisher (1966) finds that a city's average family income has a negative effect on the arrest rates of young males, while Ehrlich (1973) finds that states with higher median family incomes have higher rates of violent and property crimes.
37. Fleisher (1966) measures income inequality as the difference between the average income of the second lowest quartile and the highest quartile of households, whereas Ehrlich (1973) uses the percentage of families below one-half of the median income.
38. Ehrlich (1973) assumes that the median income for the state is a good proxy for the payoffs from crime—the "opportunities provided by potential victims of crime"—while legitimate opportunities available to potential offenders may be approximated by the mean income level of those below the state's median income.
39. Fleisher (1966); Ehrlich (1973). In the words of Fleisher, "in attempting to estimate the effect of income on delinquency, it is important to consider the effects of both normal family incomes and deviations from normal due to unemployment" (1966, p. 121).
40. See the literature review in Freeman (1994). Two notable exceptions are Witte (1980) and Trumbull (1989). Trumbull's analysis is based on county-level data from North Carolina, while Witte follows a sample of North Carolina men released from prison.
41. Tauchen, Witte, and Griesinger (1994, p. 410).
42. Grogger (1997).
43. Grogger (1997). The author's econometric results on the youth wage-crime relation also help explain racial differences in rates of crime participation and the age distribution of crime.
44. Freeman (1991, p. 6).
45. Ehrlich (1975b).
46. Ehrlich (1975b, pp. 319-35).
47. Witte and Tauchen (1994). The same finding is reported in Tauchen, Witte, and Griesinger (1994, p. 410), who find a negative relation between crime and the variables for the amount of time spent at work and at school, but no significant effect from educational attainment on arrest rates. Moreover, the coefficients for the time spent at work and at school are not significantly different from one another. This finding is also present in Farrington and others (1986); Gottfredson (1985); Viscusi (1986).
48. Dilulio (1996, pp. 20-21).
49. Putnam (1993, p. 173).
50. Freeman (1986).
51. Glaeser and Sacerdote (1999a, p. S253).
52. Case and Katz (1991).
53. Glaeser, Sacerdote, and Scheinkman (1996, p. 512).
54. Sah (1991, p. 1282).
55. Blumstein (1995).
56. Grogger and Willis (1998).
57. Blumstein and Rosenfeld, (1998); Grogger (1999). Grogger (1999) argues that the costs of entry in the market for illegal drugs increased with the escalation of violence that accompanied the introduction of crack cocaine. This would have shifted the supply curve for illegal drugs upward, which would explain the reversal in drug-related violence.
58. Levitt (1998b) shows that juvenile crime is at least as responsive to criminal sanctions as is adult crime; Grogger (1997) shows that when deciding to participate in crime, youth do take into account the level of legitimate wages; and Mocan and Rees (1999) find that juvenile crime responds to arrest rates and to local economic conditions.
59. Gottfredson (1986, p. 257).
60. Gottfredson (1986, p. 256).
61. Cruz, Trigueros Argüello, and González (2000, p. 14).
62. Fox and Zawitz (2000, p. 1).
63. Glaeser (1999, p. 26).
64. Newman (1999, p. 25).
65. Glueck and Glueck (1950, 1968).
66. Gottfredson and Hirschi (1990, p. 221). Cruz, Trigueros Argüello, and González (2000) use prison survey data to study the factors that make some criminals more violent than others.
67. Farrington (1986, p. 212). Panel data also provide the researcher with a means of controlling for reverse causality and other sources of endogeneity in the explanatory variables.
68. Freeman (1994, p. 10). These studies have often found that crime rates are negatively related to the contemporaneous unemployment rate but positively related to the first lag of this variable, which has been interpreted as reflecting, respectively, the effects of reduced criminal opportunities and reduced opportunity costs of crime.
69. This section draws heavily on Fajnzylber, Lederman, and Loayza (1999, 2000), and Lederman, Loayza, and Menéndez (2000).
70. For details on definitions and sources of crime data and other variables, see table A1 in the appendix.
71. Soares (1999).
72. Donohue (1998, p. 1425).
73. Newman (1999), statistical appendix.
74. To control for quality we excluded countries that had tenfold or greater increases in the reported number of crimes from one year to another. The presumption underlying this criterion is that such large jumps in the series could only be due to changes in definitions or reporting standards. For more detailed information on how the data was cleaned up, see the appendix in Fajnzylber, Lederman, and Loayza (1998).
75. The basic WHO regression sample comprises twenty OECD countries, ten Latin American countries, five Caribbean countries, four East Asian countries, three eastern European and central Asian countries, and three African and Middle Eastern countries.
76. The GMM estimator was developed by Chamberlain (1984); Holtz-Eakin, Newey, and Rosen (1988); Arellano and Bond (1991); Arellano and Bover (1995).
77. For a more complete exposition of the GMM dynamic panel methodology, see Fajnzylber, Lederman, and Loayza (2000).
78. Arellano and Bond (1991).
79. Ehrlich (1975b).
80. See Glaeser, Sacerdote, and Scheinkman (1996).
81. See Leung (1995).
82. See Sah (1991); Posada (1994).
83. An alternative explanation is that economic conditions may have a cognitive impact on individuals by affecting their moral values or tolerance for crime.
84. Ehrlich (1973, pp. 538-40).
85. Bourguignon (1998, p. 2).
86. Neapolitan (1997); LaFree (1999).
87. Soares (1999).
88. See Fajnzylber, Loayza, and Lederman (2000). Although the sign and significance of the estimated coefficients for the key crime determinants are robust, their magnitude is not very stable in different regressions. This is hardly surprising given that, first, the samples across regressions are not the same and, second, we estimate the coefficients using an instrumental-variable approach.
89. Some countries changed their stance toward the death penalty between 1970 and 1994; therefore, the death-penalty indicator used in the regressions ranges between 0 and 1.
90. Lack of data prevents us from controlling directly for the joint endogeneity of the drug-related variables, as we do in the case of our core economic variables. We use them as country averages (that is, without time variation) to minimize their within-country endogeneity with crime rates. In the case of the dummy for drug-producing countries, the production of illegal drugs responds mostly to climatic characteristics (such as abundant rain in the forests of Colombia and Bolivia) and geographic location (such as the proximity of Mexico to the United States, with its high demand for drugs). Thus this variable is not driven by prevalent crime rates in the country. At any rate, we recognize that we do not control for potential between-country endogeneity of the drug-producers dummy or the drug-possession crimes rate.
91. See Glaeser, Sacerdote, and Scheinkman (1996); Glaeser and Sacerdote (1999a).
92. See, for example, Blumstein and Rosenfeld (1998).
93. De Gregorio and Lee (1998).
94. Esteban and Ray (1994); Collier and Hoeffler (1998).
95. Esteban and Ray (1994).
96. See Contreras (1997).
97. See Fajnzylber, Lederman, and Loayza (1999) for details about the construction of the polarization index.
98. Mauro (1995); Easterly and Levine (1997); Collier and Hoeffler (1998).
99. Although the inequality result is maintained even after controlling for income polarization and ethnic division, we acknowledge that social mobility is another potentially important variable, one that is omitted here. We thank Alejandro Gaviria for pointing this out. Unfortunately, as far as we know, there is no internationally comparable data set with indicators of social mobility.
100. See Glaeser, Sacerdote, and Scheinkman (1996).
101. Dilulio (1996).
102. Glaeser and Sacerdote (1999a).
103. Rubio (1997).
104. The World Values Survey is coordinated by the Institute for Social Research, University of Michigan.
105. Muller and Seligson (1994).
106. See Collier (1998).
107. As with crime rates, we express the social capital indicators in natural logarithms. Since these indicators are given in different units, it is necessary to express them in logs to be able to compare their coefficients and interpret them as the effect on crime rates of (approximately) a percentage change in each indicator.
108. Glaeser and others (1999, p. 5) point out that their results, which are based on an experiment conducted on a sample of Harvard undergraduates, show that "while trust survey questions [such as the one from the WVS] are bad at predicting the levels of trust, they may be good at predicting the overall level of trustworthiness in a society." If these results were applicable to our sample of countries, then our results would need to consider the conceptual difference between trust (defined by Glaeser and others as "the commitment of resources to an activity where the outcome depends on cooperative behavior") and trustworthiness (defined as "behavior that increases the returns to people who trust you"). At the national level, however, it is virtually impossible to distinguish between these two concepts, because having a large number of people with trust must be highly correlated with the level of trustworthiness.
109. Soares (1999).
110. Kaufmann, Kray, and Zoido-Lobaton (1999).
111. See the case studies financed by the World Bank: Cruz, Trigueros Argüello, and González (2000); Funsalud (2000); Instituto Apoyo (2000); Piquet (2000); Vélez and others (2000). Some of the homicide rates cited here are disputed by alternative sources of information that are available in each city, and the studies cited contain detailed discussions about the alternative sources.
112. See Glaeser and Sacerdote (1999a); Gaviria and Pagés (1999).
113. For simple but formal theoretical models of the incentives to commit crimes, see Fajnzylber, Lederman, and Loayza (1998) and the appendix in Lederman, Loayza, and Menéndez (2000).
114. The case studies of Cali, San Salvador, Rio de Janeiro, and São Paulo also control for the individual's ethnic origin.
115. For Mexico City, see Funsalud (2000); for Rio de Janeiro and São Paulo, see Piquet (2000); for San Salvador, see Cruz, Trigueros Argüello, and González (2000).
116. An alternative sociological explanation of this result is that employed individuals spend more time in public areas during their commute to and from the workplace than do unemployed individuals (see Piquet, 2000).
117. For Cali, see Vélez and others (2000); for Lima, see Instituto Apoyo (2000); for Mexico City, see Funsalud (2000); for Rio de Janeiro and São Paulo, see Piquet (2000); for San Salvador, see Cruz, Trigueros Argüello, and González (2000).
118. For a discussion on how the public disclosure of information on crime and victimization can be used as a tool for fighting crime, see Lederman (1999).