“Identification and Estimation of Market Size in Discrete Choice Demand Models”.(Job Market Paper)
Presentations: BC, BU-BC Econometrics Workshop, IIOC, CEA, EARIE
Within the framework of Berry (1994) and Berry, Levinsohn, and Pakes (1995), the existing empirical industrial organization literature often assumes that market size is observed. However, the presence of an unobservable outside option is a common source of mismeasurement. Measurement errors in market size lead to inconsistent estimates of elasticities, diversion ratios, and counterfactual simulations. I explicitly model the market size, and prove point identification of the market size model along with all demand parameters in a random coefficients logit (BLP) model. No additional data beyond what is needed to estimate standard BLP models is required. Identification comes from the exogenous variation in product characteristics across markets and the nonlinearity of the demand system. I apply the method to a merger simulation in the carbonated soft drinks (CSD) market in the US, and find that assuming a market size larger than the true estimated size would underestimate merger price increases.
“Who is Most Affected by Soda Taxes? Evidence from Purchases At-Home and Away-From-Home”, with Xirong Lin.
Presentations: SEHO, BC
Using a novel dataset that includes at-home and away-from-home food purchases, we study who is affected by soda taxes. We nonparametrically estimate a random coefficient nested logit model to exploit the rich heterogeneity in preferences and price elasticities across households, including SNAP participants and non-SNAP-participant poor. By simulating its impacts, we find that soda taxes are less effective away-from-home while more effective at-home, especially by targeting the total sugar intake of the poor, those with high total dietary sugar, and households without children. Our results suggest that ignoring either segment can lead to biased policy implications.
“Instrument-free estimation of triangular equation systems with the trigmm command”, with Heejun Lee, Arthur Lewbel and Susanne M. Schennach.
Revision requested at Stata Journal.
In this article, we introduce the Stata package trigmm. The trigmm command performs an estimation for the parameters of a triangular two equation system without instruments and reports standard errors. The method is based on Lewbel, Schennach, and Zhang (Journal of Business & Economic Statistics, forthcoming), who have proposed sufficient conditions for identification and derived associated moment conditions. The estimation is conducted by casting the moment conditions into the built-in Stata command gmm. The usage of package trigmm is illustrated with simulated data and sample commands.
“Identification of a Triangular Two Equation System Without Instruments”, with Arthur Lewbel and Susanne M. Schennach.
Forthcoming at Journal of Business & Economic Statistics.
We show that a standard linear triangular two equation system can be point identified, without the use of instruments or any other side information. We find that the only case where the model is not point identified is when a latent variable that causes endogeneity is normally distributed. In this non-identified case, we derive the sharp identified set. We apply our results to Acemoglu and Johnson’s (2007) model of life expectancy and GDP, obtaining point identification and comparable estimates to theirs, without using their (or any other) instrument.
“Assessing Sensitivity to Unconfoundedness: Estimation and Inference”, with Matthew A. Masten and Alexandre Poirier (Replication files).
Forthcoming at Journal of Business & Economic Statistics.
ssc install tesensitivity from within Stata.This paper provides a set of methods for quantifying the robustness of treatment effects estimated using the unconfoundedness assumption (also known as selection on observables or conditional independence). Specifically, we estimate and do inference on bounds on various treatment effect parameters, like the average treatment effect (ATE) and the average effect of treatment on the treated (ATT), under nonparametric relaxations of the unconfoundedness assumption indexed by a scalar sensitivity parameter c. These relaxations allow for limited selection on unobservables, depending on the value of c. For large enough c, these bounds equal the no assumptions bounds. Using a non-standard bootstrap method, we show how to construct confidence bands for these bound functions which are uniform over all values of c. We illustrate these methods with an empirical application to effects of the National Supported Work Demonstration program. We implement these methods in a companion Stata module for easy use in practice.
“Buying in Small Quantities”, with Yuzhi Yao.
Dollar stores feature products sold in small sizes, which are not necessarily cheaper in terms of unit price. The behavior of purchasing in smaller quantities, especially among low-income consumers, is puzzling given the quantity discount associated with bulk-buying. To unravel this phenomenon, we develop a structural model to disentangle potential explanations: limited access due to lack of transportation, liquidity constraints, and storage costs.
“Identifying Models With Mismeasured Endogenous Regressors Without Instruments: with an Application to Monopsony in Academic Labor Markets”, with Zhanhan Yu.
We extend the model considered in Lewbel, Schennach, and Zhang (2023) to allow for measurement errors in the endogenous regressor. One limitation of Lewbel, Schennach, and Zhang (2023) is that they require the common latent variable to be a scalar, while we extend their results to allow for a vector of unobservable shocks, and it contains measurement error as a special case. The correction utilizes higher-order moments of variables. We apply this approach to study the monopsony power in academic labor markets at public research universities, addressing measurement error concerns in faculty salaries.
“Identification of Discrete-Continuous Models with Large Choice Sets and Complementarity”.