This paper studies why, after the onset of recessions, low-wealth workers experience larger falls and slower recoveries in earnings than high-wealth workers. I show that differences in job-switching and job-losing rates play an important role in explaining these dynamics. To do so, I build a quantitative search and matching model with incomplete markets and on-the-job search in which wages are determined by an alternating offer bargaining protocol that accommodates risk-averse workers and wealth accumulation. The wages of job-switchers result either from Bertrand-competition between firms or, if the poaching firm is sufficiently more productive than the incumbent, from one-on-one negotiation between poacher and worker. This model includes an ingredient I document empirically: over the first fifteen months following a job switch workers experience a 6.4 percentage point increase in their job-loss probability. Through this model I conclude that cyclical differences in job-switching and job-losing by wealth, which the model can endogenously reproduce, explain 40 percent of the gap in earnings recovery between low- and high-wealth workers following the Great Recession. I then apply the model to study the post-Pandemic behavior of job-switching and show that fiscal stimulus alleviated its fall and sustained its recovery.

How does health affect one’s preferences and ability to self-insure? How does it impact one’s valuation of government insurance? I build a life-cycle model in which health affects not only survival, earnings, and medical expenses but also, and importantly, the marginal utility of consumption. While there is previous evidence showing that the effect of health on preferences is important, the literature has not reached a consensus on either its size or its direction. I calibrate my model using data on health, consumption, and income from the Panel Study of Income Dynamics. I find that bad health reduces the marginal utility of (non-medical) consumption and that this effect lowers savings over the life cycle and decreases consumption in old age. I also show that a model without health-dependent preferences does not replicate the degree of self-insurance against health shocks observed in the data. Finally, I find that health-dependent preferences reduce the household valuation of means-tested government insurance programs in the United States.

While “’Tis impossible to be sure of any thing but Death and Taxes” Bullock (1716), the structure of taxes and their burden have undergone large and frequent changes over time. We provide a brief history of U.S. federal income tax reforms since the 1960s, calculate effective federal income tax rates for each wave of the Panel Study of Income Dynamics, and discuss how effective taxation changed from 1969 to 2016. We show that most tax regimes are short-lived and that the variation in taxes over time and across groups is large. We also use an estimated dynamic model of couples and singles to show that the various tax regimes that we estimate imply very different labor market and saving behavior. These findings stress the importance of studying and modeling tax changes over time and across groups.