The Effect of Subjective Survival Probabilities on Retirement and Wealth in the United States

Abstract

We explore the proposition that expected longevity affects retirement decisions and accumulated wealth using micro data drawn from the Health and Retirement Study for the United States. We use data on a person's subjective probability of survival to age 75 as a proxy for their prospective lifespan. In order to control for the presence of measurement error and focal points in responses, as well as reverse causality, we instrument subjective survival probabilities using information on current age, or age at death, of the respondent's parents. Our estimates indicate that increased subjective probabilities of survival result in increased household wealth among couples, with no effect on the length of the working life. These findings are consistent with the view that retirement decisions are driven by institutional constraints and incentives and that a longer expected lifespan leads to increased wealth accumulation.

Subjective mortality expectations and consumption and saving behaviours among the elderly

“…In the 2000 HRS survey, 9.5% of respondents stated that their subjective longevity probability was 0% and 10.7% stated it was 100%. Several studies such as Delavande, Perry, and Willis (2006), Bloom et al (2006), and Gan, Hurd, and McFadden (2003) suggest procedures for adjusting stated probabilities. For the main specification, I decided against correcting stated probabilities, because even somewhat unrealistic expectations might still be what agents base their decisions on.…”

“…For the main specification, I decided against correcting stated probabilities, because even somewhat unrealistic expectations might still be what agents base their decisions on. However, I also include a specification, which follows the approach by Delavande, Perry, and Willis (2006), Bloom et al (2006), andFang et al (2007), who use information on parental longevity as instrumental variables for subjective mortality expectations in order to adjust for focal answers. The exact definition of variables for parental longevity in my study follows Delavande, Perry, and Willis (2006).…”

Asset Allocation of Two‐Person Households under Different Longevity Expectations

“…The sample is limited to families where both spouses are between 46 and 65 years old. While the HRS is not representative of the population below age 51, we follow the previous research on subjective survival probability (SSP) that includes respondents in late 40s to increase the sample size (Bloom et al 2006;Hurd and McGarry 2002;Payne et al 2013). In addition, we exclude households with missing responses to the "final say" question and those where the spouses disagree on who has the "final say."…”

“…The even-numbered columns present IV estimates (̂( 2) IV and̂( 4) IV ), and the odd-numbered columns present coefficient estimates from uninstrumented counterparts (̂( 1) and̂ ( 3) ). Following Bloom et al (2006), binary indicators for parents being alive and continuous variables for parents' age (current or at death) are used as instruments. This strategy builds upon the evidence that subjective survival estimates are anchored on experience with parents' mortality outcomes.…”

“…This study uses the self-assessed probability of survival in the Health and Retirement Study (HRS) to proxy for each partner's perceived investment horizon. Past literature suggests that individuals use private health information to form expectations about their chance of survival to a certain age (Hurd 2009) or remaining life years (Rappange, Brouwer, and van Exel 2016), and that such subjective forecast has a major influence on household consumption (Salm 2010), saving (Bloom et al 2006), and portfolio choice (Spaenjers and Spira 2015). Using this measure and rich information on asset holdings, we examine how households incorporate different investment horizons of the husband and wife into the asset allocation decision.…”