Research That Matters (January 17 - 20, 2008)

Purpose: Most of the early works on early retirement emphasized economic incentives and health status and focused on the retirement behavior of male workers at the individual level, rather than women or both genders, socio-demographic environment, and both within and across countries. Thus, the main purpose of this study is to distinguish socio-demographic factors from the generosity of welfare programs in analyzing the determinants of early retirement at the country level. The second purpose of this study is to identify gender differences in the pattern of and trends in the labor force participation rate (LFPR) of persons aged 55 and 64 across countries. This study presents which and how socio-demographic environment influences male and female older persons' labor force participation differently. Thirdly, this study shows how the typology of welfare states as a means of the country-variations in social-economic environment is related to the pattern of early retirement across gender.

Method: The study employs Gøsta Esping-Andersen' typology (1990) of welfare states in categorizing nine developed countries (U.S., Australia, Canada, Germany, France, Italy, Sweden, Demark, and Finland). The datasets are from the Organization for Economic Cooperation and Development (OECD) Labor Statistics, OECD Social expenditure Database, OECD revenue Statistics, National accounts of OECD countries, and OECD Health Database. I examine the interactive effect of social welfare, economic, and demographic characteristics on the LFPR of persons aged 55 to 64 from 1984 to 2001. This study uses the Breusch-Pagan test, the Hausman test, and time series cross-sectional regression analysis (TSCSREG).

Results: the LFPRs of female older persons are increasing while those of males are declining during this period (1984-2001). The finding explains that a decline in the LFPRs of male older persons is offset by an increase in the LFPRs of female older person. The descriptive analysis shows that there is a positive relationship (r=.6, p<.0001) between old age dependency ratio and old age expenditure across the three types of welfare state. Controlling for economic incentive variables, the results of TSCSREG demonstrate that the old age dependency ratio affects males' and females' early retirement differently: the LFPR of male older person decreases (beta -1.2, p<.0001) and that of female older persons increases (beta=1.5, p<.0001) when the old-age dependency ratio increases by one. Considering the types of welfare state, the social-democratic welfare state has the highest level of incapacity benefits and unemployment benefits among three types of welfare states. The conservative welfare state has the highest increasing rate (13.3%) and the highest level (25.7) of old-age dependency ratio among the three types.

Implications: This study demonstrates that economic incentives per se do not work well to reverse the pattern of early retirement. This study shows that the labor participation decision of older workers is different by gender because men and women respond differently to institutional constrains and environment at the country level. In addition, the three types of welfare state are able to stress both similarities and dissimilarities of social welfare policy as well as social-demographic environment related to early retirement across countries.