Research Database

Differences in household saving rates are a key driver of wealth inequality. But what determines these differences in saving rates and wealth accumulation? We provide a new answer to this longstanding question based on new empirical evidence and a new modeling framework. In the data, we decompose U.S. household wealth into its main portfolio components to document two new empirical facts. First, the variation in wealth by income is mainly driven by differences in participation in asset markets rather than by the amounts invested. Wealth differences are a matter of to have or not to have. Second, the large heterogeneity in asset market participation closely follows observed differences in access to asset markets. Combining these two facts, we develop a new model of life-cycle wealth accumulation in which income-dependent market access is the key driver of differences in asset market participation and saving rates by income. The calibrated model accurately captures the joint distribution of income and wealth. Eliminating heterogeneity in access to asset markets increases wealth accumulation in the bottom half of the income distribution by 32%. Facilitating access to employer-sponsored retirement accounts improves broad-based wealth accumulation in the U.S. economy. Historical data support the model’s prediction.

Like most countries, the Chilean government closed schools as part of its pandemic public health mandates in order to restrict social-contact activities and reduce disease spread. We study the impact of central planner variation in school re-openings on parental labor supply, focusing on the initial three months after schools partially re-opened. We find that mothers’ labor force participation (LFP) decreased by 5.1 percentage points (ppts) one month after re-opening relative to mothers near closed schools. It decreased 9.5 ppts among householder mothers. These magnitudes weakened over time. Two or three months out, mothers who stayed in the labor force saw minimal increase in their ability to actively work and, more specifically, to work in informal jobs. In contrast, fathers’ LFP immediately increased anywhere from 2.0 to 2.9 ppts. Unplanned care disruptions during the re-opening of schools, an artifact of quarantine policies related to sickness and exposure, had differential effects on parental labor supply. Our findings support a theory that parental labor supply is uniquely sensitivity to the care transitions of children both in terms of gender and the householder status of the parent. Policies that encourage consistency in care transitions would largely benefit mothers’ ability to stay engaged in the labor force and advance in paid jobs and careers, especially when they are the householder.

Neighborhoods have a profound and lasting impact on children’s economic outcomes later in life, challenging the equality of opportunity promised by the American Dream. We develop a dynamic spatial equilibrium model in which children’s education choices are shaped by the costs and returns to education in their childhood location. Local returns depend on the moving-cost-adjusted education wage premia in all locations and local costs on the per-student school funding raised from local taxes. In the calibrated model, equalizing school funding across all students decreases differences in education outcomes across US counties and increases intergenerational mobility. However, the reform reduces the supply of educated workers in locations where the demand for them is highest, lowering aggregate output. Policies that instead broaden access to counties with good education outcomes increase intergenerational mobility without reducing output.

We present two valuation models which we use to account for the annual data on price per share and dividends per share for the CRSP Value-Weighted Index from 1929 to 2023. We show that it is a simple matter to account for these data based purely on a model of variation over time in the expected ratio of dividends per share to aggregate consumption under two conditions. First, investors must receive news shocks regarding the expected ratio of dividends per share to aggregate consumption in the long run. Second, the discount rate used to evaluate the impact of this news on the current price per share must be low. We use the approach of Campbell and Shiller (1987) and Campbell and Shiller (1988) to argue that the cash flow news in our model is not a stand-in for changes in expected returns: with our model parameters, returns are not predictable and price dividend spreads and ratios predict dividend growth at model-implied magnitudes. We illustrate which parameter choices account for differences between our results and prior findings in the literature. We conclude that the answer to Shiller’s (1981) question “Do stock prices move too much to be justified by subsequent movements in dividends?” is “not necessarily.”

While health affects many economic outcomes, its dynamics are still poorly understood. We use k-means clustering, a machine learning technique, and data from the Health and Retirement Study to identify health types during middle and old age. We identify five health types: the vigorous resilient, the fair-health resilient, the fair-health vulnerable, the frail resilient, and the frail vulnerable. They are characterized by different starting health and health and mortality trajectories. Our five health types account for 84% of the variation in health trajectories and are not explained by observable characteristics, such as age, marital status, education, gender, race, health-related behaviors, and health insurance status, but rather, by one’s past health dynamics. We also show that health types are important drivers of health and mortality heterogeneity and dynamics. Our results underscore the importance of better understanding health type formation and of modeling it appropriately to properly evaluate the effects of health on people’s decisions and the implications of policy reforms.

We study how negative sentiment around an industry impacts beliefs and behaviors, focusing on demands for racial justice after the murder of George Floyd and the salience of the “defund the police” movement. We assess stakeholder beliefs on the impact of protests on the stock prices of police-affiliated firms. In our survey experiment, laypeople and finance professionals predicted more negative stock price outcomes when they lacked details on the products supplied by such firms. Exposure to narratives about the context of the protests further reduced the prediction accuracy of these groups. In contrast, product information improved the prediction accuracy of respondents. Turning to real-life behavior, we find that mutual funds exposed to protests were 20% less likely to hold police stocks, after the protests, than funds in areas without protests. Political support for maintaining police funding, though in the majority, declined by 4.3 percentage points in protest areas. The salience of the “defund the police” narrative led to significant overreactions in both financial predictions and real-life behavior.

Using rural household survey data from West Bengal, we find that voters respond positively to excludable government welfare benefits but not to local public good programs, while reporting having benefited from both. Consistent with these voting patterns, shocks to electoral competition induced by exogenous redistricting of villages resulted in upper-tier governments manipulating allocations across local governments only for excludable benefit programs. Using a hierarchical budgeting model, we argue these results provide credible evidence of the presence of clientelism rather than programmatic politics.

After 1980, larger US cities experienced substantially faster wage growth than smaller ones. We show that this urban bias mainly reflected wage growth at large Business Services firms. These firms stand out through their high per-worker expenditure on information technology and disproportionate presence in big cities. We introduce a spatial model of investment-specific technical change that can rationalize these patterns. Using the model as an accounting framework, we find that the observed decline in the investment price of information technology capital explains most urban-biased growth by raising the profits of large Business Services firms in big cities.

We explore the evolution of income inequality and mobility in the U.S. for a large number of subnational groups defined by race and ethnicity, using granular statistics describing income distributions, income mobility, and conditional income growth derived from the universe of tax filers and W-2 recipients that we observe over a two-decade period (1998–2019). We find that income inequality and income growth patterns identified from administrative tax records differ in important ways from those that one might identify in public survey sources. The full set of statistics that we construct is available publicly alongside this paper as the Income Distributions and Dynamics in America, or IDDA, dataset. Using two applications, we illustrate IDDA’s relevance for understanding income inequality trends. First, we extend Bayer and Charles (2018) beyond earnings gaps between Black and White men and document that, unlike those for other groups, earnings for both Black men and Black women fell behind earnings for White men following the Great Recession. This trend lasted through 2019, the end of the data period. Second, we document a significant reversal in the convergence of earnings for Native earners in Native areas.

Property taxes impact the housing distribution across generations. Low property taxes lead to concentrated ownership among elderly empty-nesters, limiting housing for financially constrained young families. Conversely, high property taxes act as a “forced mortgage,” reducing upfront downpayments and enabling greater homeownership among younger households. We show in an overlapping generations model that raising property taxes in low-tax California to match those in higher-tax Texas increases homeownership in California by 4.6% and among younger households by 7.4% in steady state. Asset taxes can reallocate housing to higher-valuation households in the presence of financial constraints, providing an independent rationale for property taxes.