arXiv subjects
Publications and source records attributed to Shuyang Sheng.
We study the estimation of treatment effects using samples stratified by treatment status. Standard estimators of the average treatment effect and the local average treatment effect are inconsistent in this setting. We propose consistent estimators and characterize their asymptotic distributions.
We investigate the returns to adolescent friendships on earnings in adulthood using data from the National Longitudinal Study of Adolescent to Adult Health. Because both education and friendships are jointly determined in adolescence, OLS estimates of their returns are likely biased. We implement a novel procedure to obtain bounds on the causal returns to friendships: we assume that the returns to schooling range from 5 to 15% (based on prior literature), and instrument for friendships using similarity in age among peers. Having one more friend in adolescence increases earnings between 7 and 14%, substantially more than OLS estimates would suggest.
We propose logit-based IV and augmented logit-based IV estimators that serve as alternatives to the traditionally used 2SLS estimator in the model where both the endogenous treatment variable and the corresponding instrument are binary. Our novel estimators are as easy to compute as the 2SLS estimator but have an advantage over the 2SLS estimator in terms of causal interpretability. In particular, in certain cases where the probability limits of both our estimators and the 2SLS estimator take the form of weighted-average treatment effects, our estimators are guaranteed to yield non-negative weights whereas the 2SLS estimator is not.
This paper shows that the endogeneity test using the control function approach in linear instrumental variable models is a variant of the Hausman test. Moreover, we find that the test statistics used in these tests can be numerically ordered, indicating their relative power properties in finite samples.
This paper investigates social interactions in endogenous groups. We specify a two-sided many-to-one matching model, where individuals select groups based on preferences, while groups admit individuals based on qualifications until reaching capacities. Endogenous formation of groups leads to selection bias in peer effect estimation, which is complicated by equilibrium effects and alternative groups. We propose novel methods to simplify selection bias and develop a sieve OLS estimator for peer effects that is \sqrt{n}-consistent and asymptotically normal. Using Chilean data, we find that ignoring selection into high schools leads to overestimated peer influence and distorts the estimation of school effectiveness.
This paper explores strategic network formation under incomplete information using data from a single large network. We allow the utility function to be nonseparable in an individual's link choices to capture the spillover effects from friends in common. In a network with n individuals, an individual with a nonseparable utility function chooses between 2^{n-1} overlapping portfolios of links. We develop a novel approach that applies the Legendre transform to the utility function so that the optimal link choices can be represented as a sequence of correlated binary choices. The link dependence that results from the preference for friends in common is captured by an auxiliary variable introduced by the Legendre transform. We propose a two-step estimator that is consistent and asymptotically normal. We also derive a limiting approximation of the game as n grows large that simplifies the computation in large networks. We apply these methods to favor exchange networks in rural India and find that the direction of support from a mutual link matters in facilitating favor provision.