The Effect of Unemployment on Crimes

This research is about the effect of unemployment on crimes in Puerto Rico. Data from the United State Department of Labor and the Police Department of Puerto Rico was used. A regression analysis will be perform to prove a statistical relationship between crimes and unemployment. The dependent variables are the total crimes, the property crimes and the violent crimes per years. All that variables were tested against the independent variable that is the average unemployment rate in the year i since 1980 to 2008.  As a result, the unemployment rate has a correlation with crimes; and it is significant at 5% level on the total crimes (p-value 0.001), property crimes (p-value 0.000) and violent crimes (p-value 0.023). This correlation does not mean that there is a causal relationship (the unemployment is the cause of the crimes) but it is a good signal to start a deep analysis.  The relationship between all the property crimes are positives, as higher unemployment more crimes. But with the violent crimes is not the same. Maybe it occurs by omitted variables bias that affects the analysis and should be included to increase the accuracy.  Also an instrumental variable is needed to the unemployment rate to reduce the reverse causality; perhaps there is unemployment because companies do not want to install in areas where there are crimes. The main findings are that reducing 1% of unemployment, the total crimes per year will be reduced in 3% that this mean 2,979 less deaths per each percentage reduced.