Research Database

Many U.S. households have significant wealth late in life, contrary to the predictions of a simple life-cycle model. In this paper, we document stark differences between U.S. and Sweden regarding out-of-pocket medical and long-term-care expenses late in life, and use them to investigate their role in discouraging the elderly from dissaving. Using a consumption-saving model in retirement with significant uninsurable expense risk, we find that medical expense risk accounts for a quarter of the U.S.-Sweden difference in retirees' dissaving patterns. Furthermore, medical expense risk affects primarily financial assets, while its impact on housing is limited.

While the savings of retired singles tend to fall with age, those of retired couples tend to rise. We estimate a rich model of retired singles and couples with bequest motives and uncertain longevity and medical expenses. Our estimates imply that while medical expenses are an important driver of the savings of middle-income singles, bequest motives matter for couples and high-income singles, and generate transfers to non-spousal heirs whenever a household member dies. The interaction of medical expenses and bequest motives is a crucial determinant of savings for all retirees. Hence, to understand savings, it is important to model household structure, medical expenses, and bequest motives.

We study optimal savings policies when there is a dual concern about undersaving for retirement and income inequality. Agents differ in present bias and earnings ability, both unobservable to a planner with paternalistic and redistributive motives. We characterize the solution to this two-dimensional screening problem and provide a decentralization using realistic policy instruments: mandatory savings at low incomes but a choice between subsidized savings vehicles at high incomes—resembling Social Security, 401(k), and IRA accounts in the US. Offering more savings choice at higher incomes facilitates redistribution. To solve large-scale versions of this problem numerically, we propose a general, computationally stable, and efficient active-set algorithm. Relative to the current US retirement system, we find significant welfare gains from increasing mandatory savings and limiting savings choice at low incomes.

We develop a job ladder model with labor reallocation across firms and regions, and estimate it on matched employer-employee data to study the large and persistent real wage gap between East and West Germany. We find that the wage gap is mostly due to firms paying higher wages per efficiency unit in West Germany and quantify a rich set of frictions preventing worker reallocation across space and across firms. We find that three spatial barriers impede East Germans’ ability to migrate West: migration costs, a preference to live in the East, and fewer job opportunities received from the West. The estimated model highlights that the spatial barriers needed to generate the large wage gap between East and West are small relative to the frictions preventing the reallocation of labor across firms. Therefore, policies that directly promote regional integration lead to smaller aggregate benefits than equally costly hiring subsidies within region.

Why do young firms pay less? Using confidential microdata from the US Census Bureau, we find lower earnings among workers at young firms. However, we argue that such measurement is likely subject to worker and firm selection. Exploiting the two-sided panel nature of the data to control for relevant dimensions of worker and firm heterogeneity, we uncover a positive and significant young-firm pay premium. Furthermore, we show that worker selection at firm birth is related to future firm dynamics, including survival and growth. We tie our empirical findings to a simple model of pay, employment, and dynamics of young firms.

Tracking human activity in real time and at fine spatial scale is particularly valuable during episodes such as the COVID-19 pandemic. In this paper, we discuss the suitability of smartphone data for quantifying movement and social contact. We show that these data cover broad sections of the US population and exhibit movement patterns similar to conventional survey data. We develop and make publicly available a location exposure index that summarizes county-to-county movements and a device exposure index that quantifies social contact within venues. We use these indices to document how pandemic-induced reductions in activity vary across people and places.

Over the last two decades, the literature on comparative development has moved from country-level to within-country analyses. The questions asked have expanded, as economists have used satellite images of light density at night and other big spatial data to proxy for development at the desired level. The focus has also shifted from uncovering correlations to identifying causal relations, using elaborate econometric techniques including spatial regression discontinuity designs. In this survey we show how the combination of geographic information systems with insights from disciplines ranging from the earth sciences to linguistics and history has transformed the research landscape on the roots of the spatial patterns of development. We discuss the limitations of the luminosity data and associated econometric techniques and conclude by offering some thoughts on future research.

Defaults have been shown to have a powerful effect on retirement saving behavior yet there is limited research on who is most affected by defaults and whether this varies based on features of the choice environment. Using administrative data on employer-sponsored retirement accounts linked to survey data, we estimate the relationship between retirement saving choices and individual characteristics – long-term discounting, present bias, financial literacy, and exponential-growth bias – under two distinct choice environments: an opt-in regime and an auto-enrollment regime. Consistent with our conceptual model, we find that the determinants of following the default and contribution behavior are regime-specific. Under the opt-in regime, financial literacy plays an important role in predicting total contributions, active saving choices, and maxing out contributions in the tax-preferred account. In contrast, under the auto-enrollment regime, present bias is the most significant behavioral predictor of contribution behavior. A causal interpretation of the estimates suggests that auto-enrollment increases saving primarily among those with low financial literacy.

We harmonize the results of 42 different data sets and studies dating back to the early 20th century to construct a time series of college attendance patterns for the United States. We find an important reversal around the time of World War II: before that time, family characteristics such as income were the better predictor of college attendance; afterwards, academic ability was the better predictor. We construct a model of college choice that can explain this reversal. The model's central mechanism is an exogenous rise in the demand for college that leads better colleges to become oversubscribed. These colleges institute selective admissions and raise their quality relative to the remaining colleges, as in Hoxby (2009). Rising quality at better colleges attracts high-ability students, while falling quality at the remaining colleges dissuades low-ability students, generating the reversal.

The location of individuals determines their job opportunities, living amenities, and housing costs. We argue that it is useful to conceptualize the location choice of individuals as a decision to invest in a ‘location asset’. This asset has a cost equal to the location’s rent, and a payoff through better job opportunities and, potentially, more human capital for the individual and her children. As with any asset, savers in the location asset transfer resources into the future by going to expensive locations with good future opportunities. In contrast, borrowers transfer resources to the present by going to cheap locations that offer few other advantages. As in a standard portfolio problem, holdings of this asset depend on the comparison of its rate of return with that of other assets. Differently from other assets, the location asset is not subject to borrowing constraints, so it is used by individuals with little or no wealth that want to borrow. We provide an analytical model to make this idea precise and to derive a number of related implications, including an agent’s mobility choices after experiencing negative income shocks. The model can rationalize why low wealth individuals locate in low income regions with low opportunities even in the absence of mobility costs. We document the investment dimension of location, and confirm the core predictions of our theory with French individual panel data from tax returns.

For whom has earnings risk changed, and why? To answer these questions, we develop a filtering method that estimates parameters of an income process and recovers persistent and temporary earnings for every individual at every point in time. Our estimation flexibly allows for first and second moments of shocks to depend upon observables as well as spells of zero earnings (i.e., unemployment) and easily integrates into theoretical models. We apply our filter to a unique linkage of 23.5m SSA-CPS records. We first demonstrate that our earnings-based filter successfully captures observable shocks in the SSA-CPS data, such as job switching and layoffs. We then show that despite a decline in overall earnings risk since the 1980s, persistent earnings risk has risen for both employed and unemployed workers, while temporary earnings risk declined. Furthermore, the size of persistent earnings losses associated with full year unemployment has increased by 50%. Using geography, education, and occupation information in the SSA-CPS records, we refute hypotheses related to declining employment prospects among routine and low-skill workers as well as spatial theories related to the decline of the Rust-Belt. We show that rising persistent earnings risk is concentrated among high-skill workers and related to technology adoption. Lastly, we find that rising persistent earnings risk while employed (unemployed) leads to welfare losses equivalent to 1.8% (0.7%) of lifetime consumption, and larger persistent earnings losses while unemployed lead to a 3.3% welfare loss.

This paper introduces a new long-run dataset based on archival data from historical waves of the Survey of Consumer Finances. The household-level data allow us to study the joint distributions of household income and wealth since 1949. We expose the central importance of portfolio composition and asset prices for wealth dynamics in postwar America. Asset prices shift the wealth distribution because the composition and leverage of household portfolios differ systematically along the wealth distribution. Middle-class portfolios are dominated by housing, while rich households predominantly own equity. An important consequence is that the top and the middle of the distribution are affected differentially by changes in equity and house prices. Housing booms lead to substantial wealth gains for leveraged middle-class households and tend to decrease wealth inequality, all else equal. Stock market booms primarily boost the wealth of households at the top of the distribution. This race between the equity market and the housing market shaped wealth dynamics in postwar America and decoupled the income and wealth distribution over extended periods. The historical data also reveal that no progress has been made in reducing income and wealth inequalities between black and white households over the past 70 years, and that close to half of all American households have less wealth today in real terms than the median household had in 1970.

This paper analyzes the optimal response of the social insurance system to a rise in labor market risk. To this end, we develop a tractable macroeconomic model with risk-free physical capital, risky human capital (labor market risk) and unobservable effort choice affecting the distribution of human capital shocks (moral hazard). We show that constrained optimal allocations are simple in the sense that they can be found by solving a static social planner problem. We further show that constrained optimal allocations are the equilibrium allocations of a market economy in which the government uses taxes and transfers that are linear in household wealth/income. We use the tractability result to show that an increase in labor market (human capital) risk increases social welfare if the government adjusts the tax-and-transfer system optimally. Finally, we provide a quantitative analysis of the secular rise in job displacement risk in the US and find that the welfare cost of not adjusting the social insurance system optimally can be substantial.

We investigate cyclicality of variance and skewness of household labor income risk using PSID data. There are five main findings. First, we find that head's labor income exhibits countercyclical variance and procyclical skewness. Second, the cyclicality of hourly wages is muted, suggesting that head's labor income risk is mainly coming from the volatility of hours. Third, younger households face stronger cyclicality of income volatility than older ones, although the level of volatility is lower for the younger ones. Fourth, while a second earner helps lower the level of skewness, it does not mitigate the volatility of household labor income risk. Meanwhile, government taxes and transfers are found to mitigate the level and cyclicality of labor income risk volatility. Finally, among heads with strong labor market attachment, the cyclicality of labor income volatility becomes weaker, while the cyclicality of skewness remains.

We estimate the effect of industrial robots on employment, wages, and the composition of jobs in German labor markets between 1994 and 2014. We find that the adoption of industrial robots had no effect on total employment in local labor markets specializing in industries with high robot usage. Robot adoption led to job losses in manufacturing that were offset by gains in the business service sector. We analyze the impact on individual workers and find that robot adoption has not increased the risk of displacement for incumbent manufacturing workers. They stay with their original employer, and many workers adjust by switching occupations at their original workplace. The loss of manufacturing jobs is solely driven by fewer new jobs for young labor market entrants. Moreover, we find that, in regions with higher exposure to automation, labor productivity increases while the labor share in total income declines.

We offer causal evidence of higher returns to experience in big cities. Exploiting a natural experiment that settled refugees across labor markets in Denmark between 1986 and 1998, we find that refugees initially earn similar wages across locations. However, those placed in Copenhagen exhibit 35% faster wage growth with each additional year of experience. Faster sorting of workers towards the type of establishments, occupations, and industries typically found in cities accounts for the vast majority of this urban wage growth premium.

This paper uses a unique large-scale survey administered in April 2020 to assess disparities on several dimensions of wellbeing under rising COVID-19 infections and mitigation restrictions in the US. The survey includes three modules designed to assess different dimensions of well-being in parallel: physical health, mental and social health, and economic and financial security. The survey is unique among early COVID-19 data efforts in that provides insight on diverse dimensions of wellbeing and for subnational geographies. I find dramatic declines in wellbeing from pre-COVID baseline measures across both people and places. Place-level variation is not well explained by local characteristics that either precede or coincide with the pandemic. Analysis by demographic groups also shows large and unequal declines in wellbeing in the COVID era. Hispanic, younger, and lower-earning individuals all faced disproportionately worsening economic conditions, as did those with school-aged children. I conclude that place-based relief policies are unlikely to be efficient relative to support targeted to the neediest individuals. I also find that individual COVID-19 exposure and risk show concerning relationships with employment, protective behavior, and mental health. Those with direct COVID-19 exposure through their households continue working similar hours to others, and those with recent fever symptoms or elevated risk for COVID complications are not reducing their work hours or taking additional precautions, despite negative mental health status changes indicating concern. These findings suggest that some support policies might be directly targeted to households with confirmed infections or heighted risk.

We document new facts about risk in male wages and earnings, household earnings, and pre- and post-tax income in the Netherlands and the United States. We find that, in both countries, earnings display important deviations from the typical assumptions of linearity and normality. Individual-level male wage and earnings risk is relatively high at the beginning and end of the working life, and for those in the lower and upper parts of the income distribution. Hours are the main driver of the negative skewness and, to a lesser extent, the high kurtosis of earnings changes. Even though we find no evidence of added-worker effects, the presence of spousal earnings reduces the variability of household income compared to that of male earnings. In the Netherlands, government transfers are a major source of insurance, substantially reducing the standard deviation, negative skewness, and kurtosis of income changes. In the U.S. the role of family insurance is much larger than in the Netherlands. Family and government insurance reduce, but do not eliminate nonlinearities in household disposable income by age and previous earnings in either country.

We study the impact of increased pandemic-related childcare responsibilities on custodial mothers by telework compatibility of their job. We estimate changes in employment outcomes of these mothers in a difference-in-difference framework relative to prime-age women without children and a triple-difference framework relative to prime-age custodial fathers. Mothers' labor force participation decreased between 0.1 to 1.5 percentage points (ppts) relative to women without dependent children and 0.3 to 2.0 ppts compared to custodial fathers. Conditional on being in the labor force, the probability of being unemployed fell by 0.7 ppts relative to childless women. Conditional on being employed, leave take-up increased by 0.7 ppts. These patterns were especially prominent among custodial mothers with a college degree or higher in telework-compatible jobs. Compared to women without children, mothers working as teachers and white-collar workers disproportionately left the labor market at the end of the 2020-2021 virtual school year. These mothers likely struggled balancing remote work while simultaneously supporting their children's virtual schooling needs. The disparity between mothers and fathers widened over time, indicating the prevalence of inequality in sharing household duties even today. By the start of the 2021-2022 school year, eighteen months after the pandemic began, mothers' employment was still adversely impacted by childcare disruptions. Our findings emphasize that while flexible work has been shown to increase women's labor supply, it is not sufficient to ensure continued and increasing levels of women's labor force participation if accessible and affordable childcare is unavailable while they work for pay.

I examine the impact of the COVID-19 shock on parents' labor supply during the initial stages of the pandemic. Using difference-in-difference estimation and monthly panel data from the Current Population Survey (CPS), I compare labor market attachment, non-work activity, hours worked, and earnings and wages of those in areas with early school closures and stay-in-place orders with those in areas with delayed or no pandemic closures. While there was no immediate impact on detachment or unemployment, mothers with jobs in early closure states were 68.8 percent more likely than mothers in late closure states to have a job but not be working as a result of early shutdowns. There was no effect on working fathers or working women without school age children. Mothers who continued working increased their work hours relative to comparable fathers; this effect, however, appears entirely driven by a reduction in fathers’ hours worked. Overall, the pandemic appears to have induced a unique immediate juggling act for working parents of school age children. Mothers took a week of leave from formal work; fathers working fulltime, for example, reduced their hours worked by 0.53 hours over the week. While experiences were different for mothers and fathers, each are vulnerable to scarring and stunted opportunities for career growth and advancement due to the pandemic.