Introduction

The trend towards earlier exit from the labour force, in particular among men, has been one of the most important changes in the composition of the labour force in industrialised countries over the past forty years. Because this trend has been going on at the same time as a trend towards increased longevity, the financial stability of most social security systems around the world has been threatened.

Since the decline in labour force participation has coincided with the build-up of income security systems in most countries, the economic incentives to exit the labour force inherent in these schemes has been suggested as an explanation for the trend towards earlier retirement. One way to examine this hypothesis is to model the retirement decision econometrically, and a growing literature of studies that try to do this exists (see Lumsdaine and Mitchell, 2000, for an overview). Palme and Svensson (2003) and Karlström, Palme and Svensson (2002) follow this approach using Swedish data.

This study follows a somewhat different approach: instead of modelling the retirement decision we describe in detail the retirement behaviour in different groups of the labour force. We then describe the economic incentives to exit from the labour market in these groups. The idea behind this approach is that, rather than doing a causal analysis of retirement, we study to what extent the observed retirement behaviour matches up with the observed retirement incentives.