Title: Two-sample estimation as an alternative to instrumental variable estimation in the presence of omitted variables
Authors: Masayuki Hirukawa - Setsunan University (Japan) [presenting]
Irina Murtazashvili - Drexel University (United States)
Artem Prokhorov - University of Sydney (Australia)
Abstract: When conducting regression analysis, econometricians often face the situation in which some regressors are unavailable in the data set at hand (e.g., an ability measure in wage regression). They typically treat the missing regressors as omitted, and for consistent estimation of model parameters, they attempt to find valid instruments for the regressors that are suspected to be endogenous due to their possible correlation with the omitted regressors. We argue that we can also estimate parameters by combining the original data set with another one containing the `missing' regressors. A consistent, two-sample estimator is proposed as an alternative to the matched sample indirect inference estimator, and its asymptotic properties are explored. Moreover, we refer to dimension reduction methods of matching variables and conduct Monte Carlo simulations.