Working Papers


Region-based policies in China have been the source of controversy for many years. Existing policies concerning education and migration produce inequities in quality of and access to opportunity, favoring those in rich provinces. We consider three relevant policies of note. First, decentralized public K-12 education spending implies that richer regions have better-funded public schools. Second, the Ministry of Education allocates more seats at public colleges to richer provinces relative to the number of provincial applicants. Finally, the Hukou, or residential permit, system imposes substantial costs to migrate to rich areas by restricting access to local Hukou, which is required for individuals to access important public resources such as education, subsidized housing markets, and healthcare. We build a structural spatial overlapping generations model that encapsulates China's institutional setting and calibrate it to recent Chinese data. We then use the calibrated model to estimate effects of changing these policies on intergenerational mobility, welfare, education, and the income distribution, with particular interest in outcomes for children born at the bottom of the parental income distribution. We find that lifting Hukou restrictions and equating public spending levels increases intergenerational mobility and improves outcomes for the poorest born in poor provinces, but at the expense of the poorest born in rich provinces. Modifying college allocations to a merit-based or equity-based admissions system changes college enrollment outcomes without substantial aggregate effects, and positive effects are largely limited to children born to richer parents.


I use individual-level longitudinal survey data from the Indonesian Family Life Survey to characterize intergenerational mobility in Indonesia. I produce estimates of mobility in Indonesia for the full economy as well as urban and rural subsamples of children and find that relative mobility is high relative to other developed countries. I also explore potential mechanisms of mobility such as education and location and sector of work. Contrary to previous evidence in the United States, I find no evidence of neighborhood effects between urban and rural areas.


U.S. Employment and Opioids: Is There a Connection?

(with Janet Currie and Molly Schnell), Health and Labor Markets (2019)


This chapter uses quarterly county-level data from 2006 to 2014 to examine the direction of causality in the relationship between per capita opioid prescription rates and employment-to-population ratios. We first estimate models of the effect of per capita opioid prescription rates on employment-to-population ratios, instrumenting opioid prescriptions for younger ages using opioid prescriptions to the elderly. We find that the estimated effect of opioids on employment-to-population ratios is positive but small for women, while there is no relationship for men. We then estimate models of the effect of employment-to-population ratios on opioid prescription rates using a shift-share instrument and find ambiguous results. Overall, our findings suggest that there is no simple causal relationship between economic conditions and the abuse of opioids. Therefore, while improving economic conditions in depressed areas is desirable for many reasons, it is unlikely on its own to curb the opioid epidemic.

Work in Progress

Decomposing School Starting Age Effects

(with Dejan Kovač)