Summary
Abstract
The objective of the present study is to determine whether or not pre-recorded evidence raises the probability of a conviction in cases of Domestic Violence (DV) assault. In order to answer this question, we use administrative police and court data from the Australian state of New South Wales (NSW). The data contains information for cases involving at least one DV assault charge that was finalised between 1 June 2015 and 31 August 2018. Using these data we exploit exogenous variation in the availability of pre-recorded evidence in an Instrumental Variables framework. We find that pre-recorded evidence increases the probability of a conviction by six percentage points. In relative terms, when compared to cases without pre-recorded evidence, this equates to an increase of 7.9 per cent. When we restrict our sample to the one in four cases that proceed to a defended hearing, we find that pre-recorded evidence raises the probability of a conviction by 17.1 percentage points (a relative increase of 24.5%).