Panel estimates of the determinants of British regional male incapacity benefits rolls 1998–2006

Abstract

This article explores the determinants of the proportion of the working age male population claiming Incapacity Benefits (IB), across the 11 British Government Office Regions, for the period 1998 to 2006. Three different approaches are adopted to modelling register dynamics: first, treating IB stocks as if they were trend-stationary, albeit with persistence, and estimating reduced form models for their logs; second, treating IB stocks as if they were nonstationary and examining their long-run determinants plus short-run equilibrium reversion properties; third, focusing on the determinants of gross inflows and outflows that together drive IB stocks. Given the nature of the data, no approach is ideal yet the models provide reasonably robust evidence that labour market changes–specifically falling unemployment rates and rising real earnings–have contributed to falling male IB stocks over the period.

Acknowledgements

Thanks to two anonymous referees and seminar participants at WPEG2007 for helpful comments on an earlier draft. Also, thanks to the Northern Ireland Department of Enterprise, Trade and Investment for supplying ready-prepared quarterly, regional LFS data.

Notes

¹ Here we follow the usual convention in the British IB literature of using the term ‘incapacity benefits’ to denote all income replacement IBs, i.e. incapacity benefit itself, but also Severe Disablement Allowance and Credits Only cases. We use the abbreviation ‘IB’ as a shorthand for all these IBs, not just for incapacity benefit, which itself is commonly abbreviated to ‘IB’.

² www.erini.ac.uk/sandproject

³ IVB was the predecessor to IB.

⁴ Cross-section approaches at a more disaggregated spatial level (e.g. Nolan and Fitzroy, 2003; McVicar, 2007) do not face the same constraint, but can tell us little about dynamics and face a greater risk of omitted variable bias because they cannot control for unobserved heterogeneity across space.

⁵ The ECM1 and ECM3 models impose the following restrictions based on the cross-section results from McVicar (2007): γ₁ = 1, γ₂ = 0.5, γ₃ = −1, γ₄ = 0 and γ₅ = 0.

⁶ Even with the flow data, at least for inflows, we need to account for possible dynamics because many claimants go through a 6-month period on Statutory Sick Pay (SSP) before showing up on the IB register.

⁷ Claimant data by age group show a sharp fall in the proportion of male IB claimants that are aged 60–64 since 1999Q3.

⁸ Note the contrast with the US where evidence suggests that DI stocks have risen as a result of falling real earnings and rising replacement rates.