Essays On Income Inequality And Crime In Chinese Transitional Period

During the last thirty years of economic reform and opening up to the outside world, Chinese economy has enjoyed an unprecedented rapid growth, blessing Chinese people with a continuously bettering economic welfare. The fast economic growth created a quick social economic transition far more complicated than economic growth itself.In this paper,we confine our focus on the surging crime rates in Chinese transitional period, and explore it under the analytical framework of the economics of crime.Based on related literatures and stylized facts,we firstly deduce comparative statics for income inequality, urbanization and law enforcement to crime,and then estimate the theoretical crime supply function using Chinese provincial panel and country-level time series data.The empirical fingings are as follows.Firstly, the widening relative income inequality and absolute income inequality both give birth to a surge in crime rates. Based on a provincial panel spanning from 1988 to 2007, fixed effect estimate indicates that the elasticity of relative income inequality to crime ranges from 0.34 to 0.45,and 1% rise of absolute income inequality will result in an increase in crime rates of the amount at least 0.37% to 0.52%. Since the vast unobserved social economic difference in provincial level may correlate with crime,and there are also other shocks from time dimension such as the strike hard movements and crime policy adjustments,omitting provincial fixed effect or year fixed effect will lead to an underestimate for the impact of income inequality on crime.Secondly, fast urbanization is another cause of surging crime in Chinese transitional period. Unlike the findings from samples of developed countries, we find that urbanization not only brings about a concentration of criminal opportunities making crime participation more attractive,the increase in income of the poor and improvement of labor market condition during urbanzation may also reduce crime.Based on the same provincial panel from 1988 to 2007,our empirical analysis shows that the crime increase effect of urbanization will take dominance when the urbanization rate is less than 50%,and crime-reduction effect will will be more influential after that. These results are robust with different urbanization index setups,which imply that bias may occurs in estimating the urbanization effect using samples from developing countries without including the square term of urbanization variable.Thirdly, income inequality and law enforcement have different long-term effect on different types of crimes. Based on the co-integration analysis using the country level time series data from 1981 to 2007,we find that there are co-integrations among theft,property crime and grand total crimes with income inequality and law enforcement variables, and no long run relationship exists when coming to robbery,assault and other violent crimes.The long-term elasticity of income inequality and clearance rate to theft,property crime and grand total crime ranges respectively from 1.88 to 2.13 and from -0.93 to-1.14.The short run ECM analysis shows that the temporarily low crime rates moving away from widening income inequality will be adjusted to long run equilibrium path, which indicates that the reduction of crime in the long run will depend on the continuous improvement of the living conditions for the poor.Last but not the least, crime punishment is another important factor in explaining the dynamic evolution of crime rate in Chinese transitional period. however,due to the typical simultaneity problem between crime rates and law enforcement variables, the estimate for crime deterrence effect using cross-section data will generally lead to underestimate bias. In this dissertation, we take advantage of the exogenous change in law enforcement variables by the two strike hard movement taking place in 1983 and 1996 to tackle this simultaneity problem using a GMM approach for dynamic panel. Our estimate indicates that the elasticity of crime deterrence ranging from -0.45 to -0.70. Since we've controlled the probability of punishment the significantly negative sign of dummies for the strike hard years just shows up the deterrence effect of the increased punishment in that period.Compared with existing domestic literatures focus on country level time series data, our empirical work based on provincial panels not only takes advantage of bigger sample size, but the possibility for control for provincial fixed effect and year dummies also make the estimates suffer less from unobserved heterogeneity and measurement error, which then guarantee a more robust estimate. And unlike studies using country level panels, our empirical work enjoys the benefit of vast varieties in social economic conditions in provincial level but sharing a same statutory system. Thus,in this sense,the estimates in this dissertation supplement literatures with new evidence from China, and also provide baseline for crime policy in transitional period.