On the estimation of treatment effects with endogenous misreporting

Abstract

as well as conference participants at the Tinbergen Institute, IZA, CERGE-EI, the Southern Economic Association, and the American Society of Health Economists, for helpful comments. The views expressed herein are those of the authors and do not necessarily reflect the views of the National Bureau of Economic Research. NBER working papers are circulated for discussion and comment purposes. They have not been peer-reviewed or been subject to the review by the NBER Board of Directors that accompanies official NBER publications.

The Effect of SNAP and School Food Programs on Food Security, Diet Quality, and Food Spending: Sensitivity to Program Reporting Error

“…There have been a number of articles analyzing the impact of measurement error in a binary independent variable on estimated regression coefficients (Aigner 1973;Bollinger 1996;Hausman et al 1998;Lewbel 2007;Hu 2008;DiTraglia and Garcia-Jimeno 2017;Hu et al 2017;Nguimkeu et al 2017). Here, we derive the bias in OLS and IV both for the case where the measurement error is exogenous and where it is endogenous, for a simple model with no additional covariates other than the binary variable (for a derivation of bias with an additional covariate vector, see Nguimkeu et al 2017).…”

Estimating the Associations between SNAP and Food Insecurity, Obesity, and Food Purchases with Imperfect Administrative Measures of Participation

“…A growing literature documents the problems with relying on survey measures of program participation, which suffer from significant reporting error, when conducting impact evaluations (Meyer, Mok, and Sullivan 2015;Mittag 2016;Nguimkeu, Denteh, and Tchernis 2019). Administrative data are ordinarily assumed to be the "gold standard" to overcoming these econometric challenges, but relatively little evidence exists on the potential problems with administrative records or econometric strategies to address them.…”

“…Several studies have documented the expected negative relationship between SNAP and food insecurity ( Van Hook and Ballistreri 2006;Nord and Prell 2011;Schmidt, Shore-Shephard, and Watson 2016), but others have found statistically insignificant or even positive associations (Gundersen and Oliveira 2001;Huffman and Jensen 2003;Hofferth 2004;Wilde and Nord 2005;Hoynes and Schanzenbach 2016). SNAP is also often found to be positively correlated with obesity, but some studies find insignificant or negative effects (Meyerhoefer and Pylypchuk 2008;Gundersen 2015;Almada, McCarthy, and Tchernis 2016;Denteh, 2017;Almada and Tchernis 2018;Nguimkeu, Denteh, and Tchernis 2019).…”

“…2 For instance, the estimated false negative rates for SNAP in various surveys range from 20% to almost 50% (Mittag 2013;Meyer, Mittag, and George 2018). There is a growing literature suggesting that the estimated effect of a misclassified binary explanatory variable (such as SNAP participation) may be substantially biased and may even yield "wrong signs" (Kreider 2010;Kreider et al 2012;Nguimkeu, Denteh, and Tchernis 2019). Within a one-sided model of endogenous misreporting, Nguimkeu, Denteh, and Tchernis (2019) provide sign-switching results for the ordinary least squares (OLS) estimator even when participation is exogenous.…”

“…There is a growing literature suggesting that the estimated effect of a misclassified binary explanatory variable (such as SNAP participation) may be substantially biased and may even yield "wrong signs" (Kreider 2010;Kreider et al 2012;Nguimkeu, Denteh, and Tchernis 2019). Within a one-sided model of endogenous misreporting, Nguimkeu, Denteh, and Tchernis (2019) provide sign-switching results for the ordinary least squares (OLS) estimator even when participation is exogenous. In this case, they show that the OLS estimator yields the wrong sign if misreporting is endogenous, with the size of the signswitching region increasing with the rate of false negatives and decreasing with the true participation rate.…”