"Subjective life expectancies, time preference heterogeneity and wealth inequality” (with Jonna Olsson)
There is a substantial heterogeneity in life expectancy in the population. However, an individual's consumption-savings decision is not necessarily guided by the objective statistical life expectancy, but rather by the individual's beliefs about survival. In this paper, we document a systematic bias in survival beliefs: individuals with a low survival probability relative to their peers underestimate their life expectancies, while individuals with a high survival probability overestimate theirs. To gauge the effect of heterogeneity in life expectancy (objective and subjective) on savings rates and ultimately wealth inequality, we introduce shocks to survival beliefs into an otherwise standard overlapping-generations model. We show that with a bequest motive calibrated to match asset decumulation in old age, such a model exhibits a counter-factual savings behavior as individuals increase their savings when their life expectancy drops. Nevertheless, the overall wealth inequality in the economy is virtually unaffected by heterogeneity in survival beliefs, contrary to previous literature which finds stronger effects of heterogeneous discount factors.
"Health dynamics and heterogeneous life expectancies" (with Jonna Olsson)
In this paper, we provide improved estimates for age-dependent health transitions and survival probabilities for different subsamples of the US population. The estimated yearly transition matrices can be used in any life-cycle model where health and survival dynamics is of interest. The results show substantial heterogeneity in life expectancy in the population. For a 70-year-old man in excellent health, the probability of reaching his 80th birthday is around 75%, while the corresponding probability for a man in poor health is just below 40%. There is also substantial inequality in life expectancy between different educational groups. In the group with less than a high school degree, the life expectancy at the age of 50 is 75 years, while the average for those with some college education or more is 80 years. This difference is due to two factors. First, at the age of 50, overall health is worse in the group with lower education. Second, even conditional on health status, the health dynamics and survival probabilities for this group are worse also from the age of 50 and onwards. We estimate that the difference in life expectancy across education groups mainly stems from the worse health and survival dynamics after the age of 50.
Work in progress
"Wealth Inequality and Portfolio Choice"
In this project I explore how portfolio choice can be used to explain wealth inequality, a mechanism that so far has been mostly absent in the literature. Recent empirical research based on administrative data has shown that the portfolio composition of wealthy households vastly differs from those of poorer ones, and that there is substantial return heterogeneity across the wealth distribution. Endogenous heterogeneous returns (for example due to differences in portfolio composition) could thus be used an additional mechanism to generate wealth inequality in heterogeneous-agent macro models. Since households choose very similar portfolios with a counterfactually high share of risky assets in standard models, behavioral elements such as biased subjective expectations are added to the framework to generate a higher dispersion in portfolio composition and returns.
"On the Redistributive Effects of Government Bailouts in the Mortgage Market" (with Dirk Krueger and Kurt Mitman)
In this paper we investigate positive and normative implications of bailout guarantees for mortgage lenders. The implicit guarantee for debt issued by Fannie Mae and Freddie Mac in the US prior to 2008 led to lower borrowing costs for financial intermediaries in the mortgage market, which were passed on to households taking out mortgages. This in turn might have affected the distribution of real estate ownership and leverage among US households. On the other hand, the bailouts in 2008 required injections of $187bn of taxpayer money into these companies. The paper attempts to quantify this trade-off of lower borrowing rates and higher taxes to fund bailouts in a model featuring heterogeneous households and aggregate uncertainty in the form of (severe) recessions, identifying beneficiaries and losers of such a policy as well as overall welfare effects.
"Self-selection into Retirement and Social Security Reform" (with Jonna Olsson)
We investigate optimal retirement behavior in a life-cycle model of agents who are heterogeneous with respect to age, assets, productivity, health and social security entitlements. Our model replicates stylized facts observed in the US, such as the delayed retirement of high earners. Furthermore, unlike in earlier work, we carefully model health transitions estimated from a panel of the elderly which allows us to quantify the effects of (expected) longevity on retirement decisions. We use our model to evaluate reforms to the US Social Security system, such as changes to the full retirement age or the earnings test for early retirees, and the heterogeneous effects on high vs low earners. Preliminary results indicate that the earnings test as currently implemented in the US is a major factor in delaying retirement for high earners.