An Econometric Analysis of Disability and Labour Force Participation in Ireland The Impact of Unobserved Heterogeneity and Measurement Error

This thesis provides benchmark estimates of the impact of disability on labour force participation in Ireland. Given the current focus on disability policy, this research is both necessary and timely for policymakers. In estimating the effect of disability on labour force participation, we are faced with many secondary questions and methodological problems, many of which this thesis aims to address. To this end, it contributes significantly to both national and international literature. The core questions addressed are; • What is the impact of disability on participation? • How significant are unobserved effects and state dependence? • Is there differential measurement error in self-reported disability? • What are the separate impacts of unobserved effects and measurement error in a labour force participation model? Previous research (Bound, 1991 and Lindeboom and Kerkhofs, 2002) has set out the main methodological issues involved, namely classical and differential measurement error, endogeneity from participation to disability and endogeneity via unobserved heterogeneity. This thesis uses novel methods to address all of these issues and concludes with a new estimate of the impact of disability on participation. Key results include; • the base effect of disability is to reduce labour force participation by approximately 30 percentage points, (participation rates for the non-disabled is 70 per cent); • unobserved heterogeneity (comprising mainly of state dependence) accounts for 50 per cent of this base estimate; • the disabled/ill labour force group are twice as likely to mis-report a severely limiting disability compared to what they would report if assessed as employed; • the true impact of disability is to reduce participation by about 15 percentage points. Compared to the original estimate of 26 percentage points, about half of the bias is due to unobservables, and the remaining half is a combination of differential and classical measurement error.

Name dc:type.qualificationname

phd

Level dc:type.qualificationlevel

doctoral

Grantor dc:publisher.institution

National University of Ireland Maynooth

Year dc:date.issued

2006

Author dc:creator

• Gannon, Brenda

• Economics, Finance & Accounting

Language dc:language

en