Research

Working Papers

Why do wars occur? We exploit a natural experiment to test the longstanding hypothesis that leaders declare war because they fail to internalize the associated costs. We test this moral hazard theory of conflict by compiling data on the 9,210 children of 3,693 US legislators who served in the U.S. Congress during the four conscription-era wars of the 20th century: World War I, World War II, the Korean War, and the Vietnam War. We test for agency problems by comparing the voting behavior of congressmen with draft-age sons versus draft-age daughters. We estimate that: (i) having a draft-age son reduces legislator support for pro-conscription bills by 10-17%; (ii) support for conscription increases by a quarter as a legislator’s son crosses the upper age threshold; and (iii) legislators with draft-age sons are more likely to win reelection. These results are consistent with a political agency model in which voters punish politicians for choosing unpopular policies. Our findings provide new evidence that agency problems contribute to political violence, and that elected officials can be influenced by changing private incentives.

I propose a new policy that endows firms with limited-duration, virtual shares in their own workers' future realized earnings growth. The policy motivates firms to assist workers with a key challenge of the modern world: increasing skills well into adulthood. I label the policy “generalized experience rating” (GER) because it builds on the more narrow experience rating embodied in state unemployment insurance systems. GER can be interpreted as an efficient Pigouvian tax, and as a mandate alleviating an adverse selection problem. I discuss many design issues, and argue the policy's potential benefits may well outweigh its unintended consequences.

Asian Americans are the only non-white US racial group to experience long-term, institutional discrimination and subsequently exhibit high income. I re-examine this puzzle in California, where most Asians settled historically. Asians achieved extraordinary upward mobility relative to blacks and whites for every cohort born in California since 1920. This mobility stemmed primarily from gains in earnings conditional on education, rather than unusual educational mobility. Historical test score and prejudice data suggest low initial earnings for Asians, unlike blacks, reflected prejudice rather than skills. Post-war declines in discrimination interacting with previously uncompensated skills can account for Asians' extraordinary upward mobility.

I combine several datasets spanning 1970-2010 to document a rise and decline in educational intergenerational mobility for cohorts reaching their late 20s from 1930-2010. Because the early rise in mobility may reflect bias from selective mortality, I develop a method to estimate mobility for young adults in census data back to 1940. The method relies on children age 26-29 who co-reside with parents and adjusts for missing independent children. Census results confirm mobility gains 1940-70. Mobility gains were economically large, similar across genders, larger for blacks, and driven by rising equality in high school attainment, while post-1980 mobility losses were driven by rising inequality in college attainment. Panel data on mobility by state, year, and race suggest robust positive correlations of mobility with earnings growth, earnings equality, relative teacher pay, and white population shares for both whites and blacks.

Publications

How do parental layoffs and their large attendant income losses affect children's long-term outcomes? This question has proven difficult to answer due to the endogeneity of parental layoffs. I overcome this problem by exploiting the timing of 7 million fathers' layoffs when children are age 12-29 in administrative data for the United States. Layoffs dramatically reduce family income but only slightly reduce college enrollment, college quality and early career earnings. These effects are consistent with a weak estimated propensity to spend on college out of marginal parental income. I find that larger effects based on firm closures stem from selection.

Credence goods such as health care, legal services and auto repair create a conflict of interest by requiring experts to diagnose and treat uninformed consumers. Mistreatment of consumers appears widespread empirically, but a simple explanation for mistreatment under realistic assumptions has proven elusive. I generalize the credence good model of Dulleck and Kerschbamer (2006) to incorporate the highly realistic assumption that consumers do not observe expert cost functions. The model guarantees equilibrium mistreatment in a wide range of price-setting and market environments. The model also yields testable implications regarding the nature of mistreating firms and the direction of mistreatment.

In Project STAR, 11,571 students in Tennessee and their teachers were randomly assigned to classrooms within their schools from kindergarten to third grade. This paper evaluates the long-term impacts of STAR by linking the experimental data to administrative records. We first demonstrate that kindergarten test scores are highly correlated with outcomes such as earnings at age 27, college attendance, home ownership, and retirement savings. We then document four sets of experimental impacts. First, students in small classes are significantly more likely to attend college and exhibit improvements on other outcomes. Class size does not have a significant effect on earnings at age 27, but this effect is imprecisely estimated. Second, students who had a more experienced teacher in kindergarten have higher earnings. Third, an analysis of variance reveals significant classroom effects on earnings. Students who were randomly assigned to higher quality classrooms in grades K-3 -- as measured by classmates' end-of-class test scores -- have higher earnings, college attendance rates, and other outcomes. Finally, the effects of class quality fade out on test scores in later grades but gains in non-cognitive measures persist.

Data Sets