This paper focuses on modelling the macroeconomic linkages between property crime and economic activity indicators using unobserved components time series models. Earlier studies of cyclical behaviour in crime data relate analysed crime series to consumption, GDP, unemployment, and/or inflation rates and conclude that criminal offences are sen- sitive to the cycle in economic activity. To study this behaviour in recorded burglary and theft data for England and Wales, in the period 1955 to 2001, we adopt both univariate and multivariate time series approaches within the unobserved components models. Univariate time series analysis suggests that recorded burglary (theft) data is subject to stochastic cycle processes with typical business cycle frequencies of approximately 5 and 10 years. In the multivariate framework, recorded burglary (theft) series is simultaneously modelled in a trivariate model, where unemployment and real GDP time series are modelled as endogenous. We also estimate a five-variate model, where we simultaneously model burglary, theft, unemployment, real GDP, and police variables. Some interesting findings in these analyses are: (i) the cycle in recorded burglary is almost fully determined by the economic business cycles; (ii) explanatory variables such as conviction rates and length of imprisonment affect the shorter term dynamics more than the longer term dynamics; (iii) motivational and opportunity effects between economic time series and crime can be distinguished in our modelling framework.
|Number of pages||16|
|Journal||Journal of Economics and Statistics / Jahrbücher für Nationalökonomie und Statistik|
|Publication status||Published - 2012|