retirement pensions and with the existence of social pensions for individuals who have not
contributed enough to the social security system during their lives.
Our paper belongs to a second generation of studies that address the issue of poverty in old
age using longitudinal data1. The use of longitudinal data enables us to trace changes in the
economic status of actual individuals or households that experience an event like
retirement.
A few papers have dealt with the matter of income dynamics linked to retirement, although
most of them base their results on the British2 and on the North American realities3. Ours is
the first study based on data for Portugal.
Most studies find that the exit from the labour force after reaching retirement age increases
the chances of downward income mobility [Zaidi (2001), Holden, Burkhauser and Feaster
(1988), Bardasi, Jenkins, and Rigg (2002)]. Disney, Grundy, and Johnson (1998) refer that
in the UK the average income after retirement represents between 70% and 80% of the pre-
retirement income. OECD (2001) confirms these values (around 80%) when considering
nine developed countries, and mentions that the UK shows the worst values of them. Grad
(1990) finds that in the USA full retirees earn on average 46% of pre-retirement income if
they receive one type of retirement benefit and 60% of pre-retirement income if they
receive two types of retirement benefits, which are even lower figures.
In the opposite direction, and using a cohort analysis, Williamson and Smeeding (2004)
find that in Canada and Sweden poverty rates fall as the cohort moves past 65 years of age.
Although they do not study directly the effect of retirement poverty rates, their result gives
us some relevant information, since the legally-fixed retirement age is 65 years both in
Sweden, and in Canada. Osberg (2001) focuses especially on the bottom income deciles of
population to compare the pre-retirement and post-retirement situation. He finds that in
some countries, like Canada and the UK, the presence of a floor to old age security benefits
1 This is more recent in Europe and in the rest of the world than in the US.
2 See Bardasi, Jenkins and Rigg (2002), Disney, Grundy, and Johnson (1998), Johnson, Stears and Webb
(1998), Kingson and Arsenault, (2000), Osberg (2001), Williamson and Smeeding (2004), OECD (2001), and
Zaidi (2001).
3 Grad (1990), Haveman, Holden, Wilson and Wolfe (2003), Holden, Burkhauser and Feaster (1988), Hurd
(1990), Osberg (2001), Williamson and Smeeding (2004), and OECD (2001).