Working Papers

Do Black Politicians Matter?

Abstract: This paper exploits the history of Reconstruction after the American Civil War to estimate the causal effect of politician race on public finance, using the number of free blacks in the antebellum era (1860) as an instrumental variable (IV) for black political leaders during Reconstruction. Free blacks were particularly overrepresented as officeholders, but their geographic distribution before the Civil War was unrelated to local preferences for redistribution, electoral outcomes, the tenure of black elected officials, political competition, or voter education campaigns. IV estimates show that an additional black official increased per capita county tax revenue by $0.20, more than an hour's wage at the time. Consistent with the stated policy goals of black officials, I find positive effects of black politicians on land tenancy and show that exposure to black politicians increased black literacy by 6% and decreased the black-white literacy gap by more than 7%. The effects were not persistent, however, disappearing entirely once black politicians were removed from office at Reconstruction's end. These results suggest that politician race has large effects on public finance and individual outcomes over and above electoral preferences.


Non-randomly Sampled Networks: Biases and Corrections
(with Chih-Sheng Hsieh, Stanley Ko, and Jaromir Kovarik)

Abstract: This paper analyzes statistical issues arising from examining networks from non-representative samples of the population. We first characterize the biases in both network statistics and the estimates of network effects under non-random sampling theoretically and numerically. Sampled network data systematically bias the properties of observed networks and suffer from non-classical measurement-error problems if applied as regressors, even in representative samples. Apart from the sampling rate and the elicitation procedure, these biases depend in a non-trivial way on which subpopulations are missing with higher probability. Hence, the diagnosis of the statistical issues in a particular data set should be performed on a case-by-case basis. We then propose a methodology, adapting post-stratification weighting approaches to networked contexts, which enables researchers to recover several network-level statistics and reduce the biases in the estimated network effects. The main advantage of the proposed methodology is that it can be applied to network data collected via both designed and non-designed sampling procedures and is straightforward to implement. We apply our approach to two commonly used network data sets and show that abstracting from non-representativeness of the sample dramatically changes the results of regression analysis.


Is the Best Interest of the Child Best for Children? Educational Attainment and Child Custody Assignment
(with Yang Chen)

Abstract: Between the 1970s and 1990s, state custody laws moved from maternal preference to the “best interests of the child” doctrine, which gives fathers and mothers equal treatment in child custody assignment. We exploit exogenous variation across states in the timing of this custody law change to estimate the long-term implications of exposure to a gender-neutral custody law regime. We find that childhood exposure to gender neutral custody laws has a negative effect on educational attainment. A child exposed to gender neutral custody law is less likely to graduate from high school by 1.5 to 2.0 percentage points.


Do Gender Neutral Custody Laws Increase Divorce Rates?
(with Yang Chen)

Abstract: Between the 1970s and the 1990s, state custody laws moved from maternal preference to the “best interest of the child” doctrine, giving fathers and mothers equal treatment in child custody decisions in the case of marital dissolution. We exploit exogenous variation across states in the timing of the legal changes to identify the effect of custody law reform on divorce. We find that changes in custody laws raised divorce rates in the long term. The divorce rate began to increase approximately seven years after a state’s adoption of the new custody law and persisted thereafter. The magnitude of the increase was between 0.1 and 0.2 divorces per 1,000 people per year, increasing the divorce rate by more than 5%. Changes in custody laws also increased the likelihood of being separated by roughly 0.5 percentage points for women and 0.3 percentage points for men. We also show that states’ movement from maternal preference to gender-neutral custody laws was independent of the adoption of unilateral divorce laws. The results suggest that child custody law reform plays an important and overlooked role in marital dissolution in the United States.


Moveable Feasts: A New Approach to Endogenizing Tastes
(with Paul Rhode)

Abstract: We provide a new empirical approach to endogenizing tastes in consumer demand. We argue that tastes can be understood as the result of utility maximizing behavior in the past, whose properties can be used to partially endogenize tastes. As the old maximization problem depends critically on relative prices, we use old relative prices to endogenize tastes, overcoming many of the empirical criticisms of the taste formation literature while at the same time being consistent with a broad class of existing theoretical approaches to taste and preference formation. To test the empirical implications of our approach, we estimate the demand for food using unique household consumption and price data from the nineteenth century. We use contemporaneous relative prices and old relative prices from the home countries of immigrants measured fifteen years prior to our consumption survey. We first establish that the old relative prices are uncorrelated with the contemporaneous relative prices. We then find that older relative prices have a large and significant effect on the demand for food. On average, a one standard deviation in the old relative price changes the current food budgetshare by .2 standard deviations. We also provide suggestive evidence of persistence—the effect of old relative prices on demand persists more than 40 years later. We conclude by noting how our empirical strategy can be used to parameterize changes in tastes in both microeconomic and macroeconomic contexts.


On Family Allocation Strategy in the Late Nineteenth Century

Abstract: I analyze the intrahousehold allocation of resources among nineteenth century industrial families. The narrative record and economic theory suggest that we should find allocation differences by gender. Using a large survey of industrial households in the late nineteenth century, I find no evidence of gender bias in household allocations to children, nor can I reject the hypothesis that allocations were efficient. These findings cannot be explained by parental egalitarianism. I find that parents were strategic out of necessity—the future cooperation of children was unknown and highly uncertain, tempering any desire for gender bias in household allocations. Narrative and quantitative evidence supports this conclusion.