EcoSta 2022: Start Registration
View Submission - EcoSta2022
A0264
Title: Measurement errors in panel data regression: A direct estimation approach Authors:  Alexandra Soberon - Universidad de Cantabria (Spain) [presenting]
Winfried Stute - University of Giessen (Germany)
Abstract: The estimation of a multiple mismeasured regressor errors-in-variables model with panel data is considered. Using the dependence structure of this data as an additional source of information, we are able to provide a correction for measurement error. More precisely, closed-form two-step estimators are obtained as solutions to estimating equations that exploit the information contained in the second-order moments of the residuals and quasi-residuals obtained by partialling out perfectly measured regressors. Then, the resulting estimators are valid even when distributional assumptions such as the non-normality of the error variables cannot be justified. The asymptotic properties of these estimators are analyzed for both random effects and fixed effects and the finite sample properties are shown via Monte Carlo simulations. Also, the methodology is used in a corporate-finance application of regressions with mismeasured regressors.