Abstract
This thesis aims to explore three key determinants of crime in the Becker-Ehrlich model: labour market conditions, the criminal justice system and family background. We firstly apply dynamic panel data analysis to test the effect of the labour market conditions and criminal justice system on crime for the UK and US. We test the effect of deterrence and unemployment on burglary, theft and robbery. We adopt the methods used by Reilly and Witt (1996) and replicate them on new data for England and Wales. We improve on the previous analyses by introducing dynamics and estimating the model in the GMM framework. The results reveal that higher clear-up rate predicts lower levels of burglary, theft and robbery. Past crime rates positively predict current crime rates and unemployment rate effects are positive and significant in the system GMM specification. The second empirical chapter tests the impact of the labour market opportunities of those most likely to commit crime (unskilled males) on area arrest rates in the US utilising the GMM estimation. We have put together crime data from the Uniform Crime Reports and data on the labour market conditions from the Current Population Survey from 1964 to 2008. We find a positive and significant relationship between unemployment and property crime arrest rates. However, for weekly earnings, the only clear and significant effect is on the burglary arrest rates. In the third empirical chapter we set up a logistic model to test for the association of parents’ criminal background with children’s contact with the police. We use self-reported data from the Offending Crime and Justice Survey (2003-2006). We find that children whose parents/guardians have committed at least one crime have a 2. 5 times higher odds of committing an offence than children of non-criminal parents. The effect of a custodial sentence of the parents increases the odds of the children offending by 3. 5 times.