This paper uses panel data techniques to examine the relationship between unemployment and a range of categories of crime in New Zealand. The data cover sixteen regions over the period 1984 to 1996. Random and fixed effects models are used to investigate the possibility of a causal relationship between unemployment and crime. Hypothesis tests show that the inclusion of random region effects and fixed time effects produces the most efficient model. The main finding of the paper is that unemployment cannot explain changes in the overall crime rate, although significant effects on crime are found for some subcategories of crime.