Essays on Subjective Survival Probabilities, Consumption, and Retirement Decisions

Abstract

Recent pension reforms in industrialized countries are, in part, motivated by the increased life expectancy. As individuals are expected to take more responsibility in their retirement planning and savings decisions, it is important to understand whether they are aware of improvements in life expectancy. The second chapter of this thesis investigates the extent to which individuals’ beliefs about their own life expectancy coincide with the realizations. In household surveys, individual life expectancy is usually measured with questions about the probability of survival to a target age. To compare individuals’ survival probabilities with their actual mortality risk, I use data from the DNB Household Survey (DHS) supplemented with (administrative) data on actual mortality from the causes of death registry. The main finding of the second chapter is that the Dutch underestimate their life expectancy, on average. This finding is important from a policy perspective since the underestimation of life expectancy may lead to insufficient retirement savings and early retirement. In the third chapter I investigate if the underestimation of life expectancy by the Dutch can perhaps be partly explained by their insufficient knowledge on population life expectancy. I find that the Dutch’s underestimation of population life expectancy significantly explains the underestimation of their remaining lifetime. This finding suggests that if individuals were better informed about population life expectancies, the accuracy of their survival beliefs could improve considerably. As a result, a policy of providing respondents information about population life expectancies may help them making more optimal economic decisions such as saving and retirement decisions. The fourth chapter investigates whether individuals use their subjective probabilities of survival when they make economic decisions. In particular, I focus on the role of subjective survival probabilities in explaining the consumption decisions of the elderly. The theory predicts that the mortality risk increases with age; therefore, individuals increase current consumption at the expense of future consumption. In line with the predictions of the theory I find that there is some evidence in favor of wealth decumulation by elderly Americans after retirement and consumption growth decreases with higher subjective mortality rates for elderly singles. On the other hand, the finding that subjective survival probabilities do not explain couples’ consumption decisions deserves further analysis. The last chapter of this thesis studies the drop in consumption at retirement, i.e. the so-called retirement-consumption puzzle. In contrast with earlier studies in the literature, I investigate consumption behavior of retirees also after they re-enter the labor force. This chapter concludes that moving back to labor force after retirement is very common among older Americans and most of the unretirement transitions are anticipated before retirement. Moreover, in line with the predictions of the theory consumption does not respond to unretirement if it is fully anticipated. One of the reasons for this finding could be that post-retirement jobs pay much less than pre-retirement jobs and therefore individuals’ income does not increase significantly when they unretire. Overall, the findings suggest that individuals have saved enough to smooth their consumption around retirement, on average.