Research - Zhou

This paper develops a semi-parametric model for heterogeneous treatment effect in the presence of peer influence over a network. We propose a novel framework that non-parametrically models individual-level treatment responses via functions in a reproducing kernel Hilbert space (RKHS) to flexibly capture non-linearity induced by covariates while accommodating structural peer effects, both endogenous and contextual. We address the reflection problem using an instrumental variables (IV) strategy that leverage higher order neighbors across the network graph. An iterative algorithm estimates the linear parameters and the function jointly. The paper derives conditions for identification and asymptotic properties of the estimators. Monte Carlo simulations highlight the method's easy implementability and performance even when the dimension of covariates is large relative to the sample size. We revisit a real social network experiment, applying our procedure to recover the non-linearity in treatment effect and obtain sizable gains in important counterfactual policies.

Economic models produce moment inequalities, which can be used to form tests of the true parameters. Confidence sets (CS) of the true parameters are derived by inverting these tests. However, they often lack analytical expressions, necessitating a grid search to obtain the CS numerically by retaining the grid points that pass the test. When the statistic is not asymptotically pivotal, constructing the critical value for each grid point in the parameter space adds to the computational burden. In this paper, we convert the computational issue into a classification problem by using a support vector machine (SVM) classifier. Its decision function provides a faster and more systematic way of dividing the parameter space into two regions: inside vs. outside of the confidence set. We label those points in the CS as 1 and those outside as -1. Researchers can train the SVM classifier on a grid of manageable size and use it to determine whether points on denser grids are in the CS or not. We establish certain conditions for the grid so that there is a tuning that allows us to asymptotically reproduce the test in the CS. This means that in the limit, a point is classified as belonging to the confidence set if and only if it is labeled as 1 by the SVM.

Recent years have seen some sizable increase in consumer prices in Mexico. Statistics show that housing prices have risen to become the second largest contributor to the inflation over the past few years next to food expenditure. Recent literature mainly focuses on the effects of financialization and gentrification. In this paper, we aim to investigate the role of one other important factor in the rising prices of the local housing markets, short-term rental listings. We propose to use detailed Airbnb data along with national data on credit taken to purchase real estate properties as well as data from some other public sources to quantitatively analyze the impact, namely the direct and the spillover effects, of new Airbnb listings on the prices of the neighboring real estate properties. We employ a novel network interference model to disentangle the spillover effect from the direct effect and provide semiparametric consistent estimates of the effects.