Crime, Corruption and Institutions

Abstract: This paper explores the link between crime and corruption, compares their magnitudes, determinants and their effects on growth rates. The study uses a large cross country data set containing individual responses to questions on crime and corruption along with information on the respondents’ characteristics. This data set is supplemented by country level indicators from a variety of sources on a range of macro variables and on institutions in the respondent’s country of residence. A methodological contribution of this study is the estimation of an ordered probit model based on outcomes defined as combinations of crime and bribe victimisation. The principal results include the evidence that while a male is more likely to be a corruption victim, a female is more exposed to crime, especially, serious crime. Older individuals and those living in the smaller towns and cities are less exposed to crime and corruption due presumably to their ability to form informal networks that act as protective mechanisms. With increasing levels of income and education, an individual is more likely to report both crime and bribe victimisation. A crime victim is more likely than a non victim to report receiving demands for a bribe. The results suggest that variables such as inequality, unemployment rate and population size have a strong effect on the country’s crime and corruption statistics though the sign and significance of the country effects are not always robust. However, the paper does provide robust evidence that a stronger legal system and a happier society result in a reduction in both crime and corruption. While the study finds that both crime and corruption rates decline as a country becomes more affluent, as measured by its per capita GNP at PPP, there is no evidence of a strong and uniformly negative impact of either crime or corruption on a country’s growth rate. There is limited OLS based evidence of a non linear relationship between growth and corruption rates, though the significance of the corruption effect on growth disappears on the use of IV estimation. The paper also provides evidence that there has been a decline in both crime and corruption during the latter half of the 1990s. This is true even after controlling for the individual and country characteristics. JEL Codes: C13, D63, H80, I31, K42.