Research

We study the optimal design of means-tested transfers and progressive income taxes. In a simple analytical model, we show that adding a transfer to a log-linear tax induces welfare gains almost as large as in the second-best allocation. Transfers allow for more progressive average than marginal tax-and-transfer rates, achieving redistribution while preserving efficiency. In a rich dynamic model, we quantify the optimal fiscal plan. We use new flexible functions featuring targeted transfers and progressive income taxes, delivering a good empirical fit across the income distribution. Transfers should be larger than currently in the U.S. and financed with moderate income-tax progressivity.

This paper provides novel evidence that the added worker effect – labor force entry upon spousal job loss – is substantially stronger for young than old households. Using a life cycle model of two-member households in a frictional labor market, we study whether this age-dependency is driven by heterogeneous needs for or availability of spousal insurance. Our framework endogenizes asset and human capital accumulation, as well as arrival rates of job offers, and is diciplined against US micro data. By means of counterfactuals, we find a strong complementarity across both margins: A large added worker effect requires both high spousal earnings potential (human capital) relative to the primary earner and limited access to other means of self insurance (assets). Either one individually does not generate a sizable response of spousal labor supply to the job loss of a primary earner, but their interaction can account for the observed age differential in the added worker effect.

We study the sources of individual earnings dynamics in frictional labor markets, particularly focusing on the role of firms. First, using administrative matched employer-employee data from Denmark, we establish a strong connection between the likelihood of large earnings changes for workers and the revenue growth of their employers. This relation is driven by a higher incidence of both job losses and earnings changes of stayers in shrinking firms. Second, we build a random search model of heterogeneous firms, choosing how many jobs to create, and heterogeneous workers, choosing search effort both on- off-the-job, calibrated to be consistent with key statistics of earnings inequality, earnings dynamics, and labor market flows. Wage setting with two-sided limited commitment and productivity shocks for both workers and firms are crucial to generate earnings dynamics with many small but some very large earnings changes as in the data, as well as the right relation between firm and worker outcomes. Third, we explore policies – targeting either workers directly or indirectly through firms – designed to mitigate earnings fluctuations in this model where earnings dynamics are endogenous to policies.

In the last decades the US has experienced a large increase in both income inequality and the standard of living. In the workhorse models of optimal income taxation, larger inequality calls for a more redistributive welfare state over time. Yet, the level of income - what standard of living the poor can afford - plays no role for optimal taxes in these homothetic models. This paper incorporates non-homothetic preferences into the optimal income tax problem to capture rising living standards. We derive a Mirrleesian nonlinear income tax formula to make transparent how rising living standards affect distribution and efficiency concerns. The dominant force is a fall in the dispersion of marginal utilities as an economy becomes richer. We quantify the importance of this channel relative to the rise in inequality in a dynamic incomplete markets setup, calibrated to the US in 1950 and 2010. Even though inequality was rising substantially in this period, we find that the "level" effects significantly dampen the desired increase in redistribution.