Might the financial security of working Americans during retirement be jeopardized by their ability to cash out their pension plans when they leave a job? Federal tax rules discourage such actions, but the limited evidence available suggests the practice is common. This paper takes advantage of long-term longitudinal data in the Health and Retirement Study to update prior findings, investigate cohort differences and study the long-term consequences of pension cash-out at job separation. Armour, Hurd, and Rohwedder find that pension cash out is more concentrated among workers who experience economic or health shocks around the time of job separation. The most recent cohort of older workers more often cashed out pension balances and more frequently used the balances for spending or to pay off debt. This is likely due to most of the job separations for this cohort occurring during or in the aftermath of the Great Recession, which brought about economic shocks at higher frequency. Long-term outcomes for those who cashed out pension balances are worse than for those who did not cash out, but so were their baseline characteristics. Taking this together with the fact that outcomes are the same across populations of workers with or without access to pension cash-out, the researchers conclude that the worse outcomes among workers who cashed out are due to the experience of shocks rather than due to access to the cash-out option.
Understanding how healthy life spans are changing over time is central to public policy. A good deal of medical spending is predicated on the idea that more intensive treatments improve quality-adjusted life expectancy. Measuring healthy life expectancy is thus a first start in understanding the value of medical advance. Further, policies such as increasing the age of eligibility for Social Security or Medicare only make sense if healthy life expectancy is increasing for the vast bulk of the population. Accurate measurement of healthy life expectancy is thus essential in the welfare evaluation of such policies. Data on life expectancy is easy to obtain, but data on healthy life expectancy is more difficult. To a great extent, this is because there is no single measure of good or bad health commonly accepted in the literature. The authors' past work, along with much of the literature, focuses on disabled and non-disabled life expectancy. In this paper, Chernew, Cutler, Ghosh, and Landrum define disability as an indicator for whether an individual has impairment with any Activity of Daily Living or Instrumental Activity of Daily Living. They calculate the number a years a person turning 65 in different years can expect to live with and without a disability. Their past research (Cutler et al., 2014) shows that disability free life expectancy has increased significantly at older ages in the United States. Between 1992 and 2005, for example, life expectancy increased by 0.7 years. Disability-free life expectancy increased by 1.6 years; disabled life expectancy fell by 0.9 years. Other results have reached similar conclusions about increases in disability-free life expectancy over time (Crimmins et al. 1997, 2001, 2009), though the data used in past studies were less complete. However, little research has examined why disability-free life expectancy has increased so greatly, and in particular what role medical advance may have
played in this.
This paper was distributed as Working Paper 22306, where an updated version may be available.
Reliable measures of disease prevalence are crucial for answering many empirical research questions in health economics, including the causal structures underlying the correlation between health and wealth. Much of the existing literature on the health-wealth nexus relies on survey data, for example those from the Health and Retirement Study (HRS). Such survey data typically contain self-reported measures of disease prevalence, which are known to suffer from reporting error. Two more recent developments — the collection of biomarkers and the linkage with data from administrative sources such as insurance claims — promise more reliable measures of disease prevalence. In this paper, Heiss, McFadden, Winter, Wuppermann, and Zhu systematically compare these three measures of disease prevalence.
This paper investigates the effects of spatial housing price risk on housing choices over the life-cycle. Housing price risk can be substantial but, unlike other risky assets which people can avoid, most people want to eventually own their home thereby creating an insurance demand for housing ownership early in life. This paper's contribution instead is to focus on the importance of ownership as a hedge against future house price risk as individuals move up the ladder. Banks, Blundell, Oldfield, and Smith show that people living in places with higher housing price risk should own their first home at a younger age, should live in larger homes, and should be less likely to refinance. These predictions are shown to hold using comparable panel data from the United States and United Kingdom.
This paper was distributed as Working Paper 21255, where an updated version may be available.
Suicide rates, life evaluation, and measures of affect are all plausible measures of the mental health and wellbeing of populations. Yet in the settings Case and Deaton examine in this paper, correlations between suicide and measured well-being are at best inconsistent. Differences in suicides between men and women, between Hispanics, blacks, and whites, between age groups for men, between countries or U.S. states, between calendar years, and between days of the week, do not match differences in life evaluation. By contrast, reports of physical pain are strongly predictive of suicide in many contexts. The prevalence of pain is increasing among middle-aged Americans, and is accompanied by a substantial increase in suicides and deaths from drug and alcohol poisoning. The researchers' measure of pain is now highest in middle age. Pain comes with low positive affect in middle age. In normal times, in the absence of the pain epidemic, suicide and life evaluation are likely unrelated, leaving unresolved whether either one is a useful overall measure of population well-being.
Poterba, Venti, and Wise consider assets when households were last observed prior to death in the Health and Retirement Study (HRS) and then trace assets backwards to the age when the household was first observed in the HRS. The researchers find that for most households assets in the last year observed (LYO) were very similar to assets of households in the first year observed (FYO). The researchers then estimate the relationship between individual attributes — in particular education, changes in health and changes in family composition — and the change in assets between the FYO and the LYO. The authors obtain estimates for HRS respondents who were 51 to 61 in 1992 and for AHEAD respondents who were age 70 and over in 1993. In addition, the authors obtain estimates by family status pathway — two-person in FYO and LYO, one-person in both years, and one-person in the LYO and two-person in the FYO.
What is the socially optimal level of liquidity in a retirement savings system? Liquid retirement savings are desirable because liquidity enables agents to flexibly respond to pre-retirement events that raise the marginal utility of consumption. On the other hand, pre-retirement liquidity is undesirable when it leads to under-saving arising from, for example, planning mistakes or self-control problems. This paper compares the liquidity that six developed
economies have built into their employer-based defined contribution (DC) retirement savings systems. Beshears, Choi, Hurwitz, Laibson, and Madrian find that all of them, with the sole exception of the United States, have made their DC systems overwhelmingly illiquid before age 55.
This paper was distributed as Working Paper 21168, where an updated version may be available.
Complex patients with many comorbid conditions are among the highest-cost users of Medicare, and they constitute an important source of growth in Medicare expenditures. In this paper, Bhattacharya and MaCurdy analyze the universe of 2009 Medicare claims to characterize the complexity of patients with multiple comorbid conditions. The analysis finds that such patients cannot be placed into a small number of clinical bins; instead, the number of different combinations of comorbid conditions is staggeringly large and there are often very few patients with any particular combination of conditions. Furthermore, Medicare expenditures on patients grow non-linearly with the number of comorbid conditions afflicting patients. The results have important implications for existing risk adjustment methods used by Medicare, which do not sufficiently account for the way interactions among comorbid conditions tend to increase costs. Finally, the results suggest that disease management and care coordination programs will face a difficult challenge in coping with the heterogeneity of patient health conditions.
What Determines End-of-Life Assets? A Retrospective View
Movies, Margins and Marketing: Encouraging the Adoption of Iron-Fortified Salt
Suicide, Age, and Wellbeing: an Empirical Investigation
House Price Volatility and the Housing Ladder