Job Market Paper
The Long-Run Effects of Parental Unemployment in Childhood
Parental job loss is a large, negative shock to the household that can affect children in both the short- and long-run. Little is known, however, about how the long-run impacts of job loss on children vary with the child's age at the time of displacement. This paper provides the first empirical evidence of the long-run effects of parental unemployment on children exposed before age 10 (and as young as 2), a period thought to be critical for child development. Using administrative tax data covering the universe of children born in Canada between 1972 and 1985 and random forest proximity matching, I estimate the causal effects of parental job loss experienced at different points in childhood on a child's income attainment. I find that children exposed to parental unemployment at ages 2 to 10 experience losses of 3 to 4 rank points in average earnings attainment in adulthood (approximately $2,500 per year). These children are also 36% more likely to receive welfare as adults and 4% less likely to pursue post-secondary education. Consistent with critical periods of child development, children who experience parental job loss before age 10 experience larger reductions in income attainment than children exposed at older ages. Decomposing these estimates, I show that the majority of my treatment effects are attributable to the timing of income losses experienced during childhood, as well as unemployment-induced moves to neighourhoods with less opportunity.
Working Papers
The Places They Left Behind: The Great Migration and Southern Development (Under review)
During the Great Migration, millions of black families left the South in pursuit of economic opportunities in the North. Its effects on Northern cities have been widely studied, and it is well established that the Great Migration in part caused white flight, high levels of urban segregation, and the poor economic prospects of black inner city residents. Despite the extensive work on the North, relatively little attention has been paid to how the counties where exodus began were affected. Using a reverse shift-share instrument identified off of Northern pull factors, I bring modern causal methods to study how black emigration affected Southern development. Given the South's history of slavery and de jure segregation, I pay special attention to how the gains from development were shared among the black and white residents of a given county. I find that the Great Migration increased educational attainment and reduced many measures of socio-economic poverty: geographic segregation, crime rates, poverty rates, and incidence of single parenthood. Despite all of these improvements, I find little evidence that the Great Migration improved the lot of white or black residents of the South. The county-level improvements are driven by changes in the demographic composition caused by the Great Migration rather than improvements for residents themselves.
Work in Progress
The Emergence of the Modern Welfare System: Evidence from the Mothers’ Pension Program 
(with Shari Eli, Price Fishback, and Adriana Lleras-Muney)
The Mothers' Pension program was the first welfare program in the United States. It provided the blueprint for the modern day welfare system, legislating transfers to poor mothers with dependants in 46 states. We investigate the origins of the Mothers' Pension program. We use newly collected data on county-level pension payments to shed light on some old questions. First, why was this legislation adopted? Second, what factors affected the implementation of the legislation and are those the same factors that determined the original adoption of the program? Despite robust legislative coverage, the actual funding and implementation of these programs was left to the county governments and very few pensions were paid as a result. We train a random forest to predict the payment of pensions at the county-level, and find that more affluent and educated counties with larger immigrant populations tended to be more likely to pay pensions.
The Great Equalizer? The Intergenerational Transmission of Human Capital and Optimal Public School Provision 
(with Torsten Jaccard)
Parents differ in their ability to endow children with human capital and therefore earnings potential. The literature modelling parental decision-making in this vein, however, often presents the trade-off for parents as a matter of consumption: a parent can consume for themselves or purchase human capital for their child. This paper introduces a model of intergenerational inequality in which parents endogenously invest time in the development of their child's human capital. The productivity of a parent's time spent teaching her child is increasing in her human capital as is her wage rate of market labour, thus introducing ambiguity in the relationship between parental human capital and labour supply. We document the flexibility of this model through a number of exercises. We consider how the availability of teachers, in both private and public schools, allows parents to outsource their human capital time investments and reduces intergenerational income inequality. We solve the planner's problem, choosing the optimal level of skills provided by public schools funded with either (i) a labour income tax or (ii) a tax on monetary bequests. In both cases, the optimal level of public schooling selected by the planner reflects the strength of intergenerational correlation of innate ability. The less correlated innate ability is across generations, the greater the social gains from redistributive public schooling as this counteracts the concentration of high ability children in borrowing constrained households and therefore increases societal productivity.
Optimal Unemployment Insurance Benefits with Dependents
Unemployed parents may be forced to cut back on human capital investments made in their children, knowing that skills which are not developed today may affect future learning opportunities. These dynamic complementarities in human capital formation are not accounted for in the Baily-Chetty formula for optimal unemployment insurance. I add dependants to the canonical model of unemployment insurance, and the derive optimal unemployment insurance schedule with an arbitrary human capital production function. If there are dynamic complementarities in human capital formation, my model implies that unemployed parents with younger children will search for work with effort which is more inelastic to unemployment insurance benefits than parents with older children. Consequently, the optimal unemployment benefit schedule should pay higher benefits to unemployed parents with young children. Using administrative data from the Labour Force Survey in Canada, I exploit a natural experiment in 1990 where Canadian unemployment insurance benefits were increased across the country as a result of a legislative logjam (Baker and Rea, 1994) to test for heterogeneity in search effort by age of dependants.