We consider the association between victimization and offending behaviour by using data from the Youth Lifestyles Survey. We consider the effect of violent and non-violent offending on the probability of being a victim of violent and non-violent crime and find a positive association between these by using univariate probit estimates. However, taking into account the endog- enous nature of offending and victimization via a bivariate probit model, we find that univariate estimates understate the association. We suggest that policy recommendations should only be informed by the bivariate analysis of the association between offending and victimization. Copyright 2004 Royal Statistical Society.
We investigate the estimation of dynamic models of criminal activity, when there is signi®cant under-recording of crime. We give a theoretical analysis and use simulation techniques to investigate the resulting biases in conventional regression estimates. We ®nd the biases to be of little practical signi®cance. We develop and apply a new simulated maximum likelihood procedure that estimates simultaneously the measurement error and crime processes, using extraneous survey data. This also con®rms that measurement error biases are small. Our estimation results for data from England and Wales imply a signi®cant response of crime to both the economic and the enforcement environment.
Abst ractFol l owi ng t he r ecent wor k of Dhi r i et al (1999) at t he Hom e Offi ce predi ct i ng recorded burgl ary and t heft for Engl and and W al es t o t he year 2001, econom et ri c and t i m e seri es m odel s have been const ruct ed for predi ct i ng recorded resi dent i al burgl ary t o t he sam e dat e. A com pari son bet w een t he Hom e Of fi ce econom et ri c predi ct i ons and t he l ess al arm i ng econom et r i c predi ct i ons m ade i n t hi s paper i dent i f i es t he di f f erences as st em m i ng f r om t he part i cul ar set of vari abl es used i n t he m odel s. However , t hese econom et ri c m odel s adopt an error-correct i on form whi ch appears i n bot h cases t o be t he m ai n r eason why t he m odel s predi ct i ncreases i n burgl ary. To i dent i f y t he r ol e of err orcorrect i on i n t hese m odel s, t i m e seri es m odel s have been bui l t for t he purpose of com pari son, al l of whi ch predi ct subst ant i al l y l ower num bers of r esi dent i al burgl ari es. The next t hree years w oul d appear t o offer an opport uni t y t o t est t he ut i l i t y of error-correct i on m odel s i n t he anal ysi s of cri m i nal behavi our.Keywords: Resi dent i al Bur gl ary: err or-correct i on; t i m e seri es forecast i ng * I am grat eful t o part i ci pant s at t he Hom e Of f i ce ' Tr ends i n Cr i m e' sem i nar, hel d i n D ecem ber 1999, for t hei r hel pful and const ruct i ve com m ent s.
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.