A Statistical Profiling Model of Long-Term Unemployment Risk in Ireland

This paper develops a statistical profiling model of long-term unemployment risk in Ireland using a combination of administrative data and information gathered from a unique questionnaire that was issued to all jobseekers making a social welfare claim between September and December 2006 who were then tracked for eighteen months. We find that factors such as a recent history of long-term unemployment, advanced age, number of children, relatively low levels of education, literacy/numeracy problems, location in urban areas, lack of personal transport, low rates of recent labour market engagement, spousal earnings and geographic location all significantly impact the likelihood of remaining unemployed for 12 months or more. While the predicted probability distribution for males was found to be relatively normal, the female distribution was bimodal, indicating that larger proportions of females were at risk of falling into long-term unemployment. We find evidence that community based employment schemes for combating long-term unemployment have little effect as participants re-entering the register typically experience extended durations. Finally, we argue that the adoption of an unemployment profiling system will result in both equity and efficiency gains to Public Employment Services.