Research

    Working Papers:

Statistical inference of heterogeneous treatment effects (HTEs) is confounded when  economic units interact because treatment effects may vary by pre-treatment variables, post-treatment variables (that measure the exposure to other units’ treatment statuses), or both. It invalidates the standard hypothesis testing techniques used to infer HTEs. In this paper, I develop statistical inference  methods  to detect HTEs and disentangle the drivers of treatment effects heterogeneity in populations where units interact. Specifically, I incorporate clustered interaction into the potential outcomes model and propose kernel-based test statistics for the null hypotheses of (a) no HTEs by treatment assignment (or post-treatment exposure variables) for all pre-treatment variables values; and (b) no HTEs by pre-treatment variables for all treatment assignment vectors. To disentangle the source of heterogeneity in treatment effects, I recommend a multiple-testing algorithm.  I prove the asymptotic properties of the proposed test statistics via a  modern poissonization technique. Furthermore, I propose bootstrap methods that better approximate the null distributions in finite-sample settings. Monte Carlo simulation evidence corroborates the theoretical findings in the paper. Finally, to gain practical insight, I illustrate the usage of the proposed tests in an empirical application using an experimental data set from a Chinese weather insurance program.  [Paper, Summary] Job market Paper





In this paper we study treatment assignment rules in the presence of social interaction. We construct an analytical framework under the anonymous interaction assumption, where the decision problem becomes choosing a treatment fraction. We propose a multinomial empirical success (MES) rule that includes the empirical success rule of Manski (2004) as a special case. We investigate the non-asymptotic bounds of the expected utility based on the MES rule. Finally, we prove that the MES rule achieves the asymptotic optimality with the minimax regret criterion. [arXiv, Replication Code] Under review, November 2022  




We develop a Stata command csa2sls that implements the complete subset averaging two-stage least squares (CSA2SLS) estimator in Lee and Shin (2021). The CSA2SLS estimator is an alternative to the two-stage least squares estimator that remedies the bias issue caused by many correlated instruments. We conduct Monte Carlo simulations and confirm that the CSA2SLS estimator reduces both the mean squared error and the estimation bias substantially when instruments are correlated. We illustrate the usage of csa2sls in Stata by two empirical applications. [arXiv] Under review, July 2022  

    Work in Progress:


1.  Randomization Inference of Heterogeneous Treatment Effects Under Interference.

I design tests of heterogeneous treatment effects (HTEs) when units interact. Specifically, I study valid randomization tests for heterogeneous   treatment effects in the presence of network interference. I model network interference into the potential outcomes framework using the concept of network exposure mapping. Then, I consider three non-sharp null hypotheses that represent different contrasts of homogeneous treatment effects in this setting. I face two main challenges: the null hypotheses are not sharp, and nuisance parameters are present. I explore potential solutions to these problems and study the factors that affect the power and size distortions of the proposed  tests in each instance. Finally, I present testable restrictions that guarantee improvements in the proposed test. Draft available upon request



2.  Measuring the Impact of Financial Bailouts: An Intervention Analysis Approach.

Over the last three decades, several empirical studies have been undertaken to assess whether financial bailouts have led to an improved balance of payments and current account balances, lower inflation, stable currency, higher growth, etc. These studies have employed a variety of methodologies and covered different samples. The results are, however, conflicting. This paper provides a different outlook by using a different methodology - intervention analysis with transfer functions - to measure the impact of financial bailouts on some of the aforementioned macroeconomic variables. I use data from Korea and Thailand before and after they enrolled in International Monetary Fund (IMF) programs during the 1997/1998 Asian currency crisis. The results highlight the ability of the intervention analysis technique to determine the time path of the programs’ impacts. The results indicate that the bailout packages had positive effects on the macroeconomic variables under consideration in the long run but no significant impacts in the short run.