international econometric journal
in Russian language
This essay presents a brief review of the theory and empirical implementation of the discontinuity design for estimation of treatment effects of programs. Using published studies as examples, we consider the main difficulties in applications of the discontinuity design in practice and interpretation of results.
This essay briefly describes the main features of some well-known multinomial discrete choice models: the standard logit, the mixed logit, and the probit.
Sample selection problems are pervasive when working with micro economic models and datasets of individuals, households or firms. During the last three decades, there have been very significant developments in this area of econometrics. Different type of models have been proposed and used in empirical applications. And new estimation and inference methods, both parametric and semiparametric, have been developed. These notes provide a brief introduction to this large literature.
This essay covers the principles and methodology of nonparametric estimation of a mean regression. The emphasis is put on kernel smoothing, but non-kernel methods are also reviewed.
Semiparametric models are characterized by a finite- and infinite-dimensional (functional) component. As such they allow for added flexibility over fully parametric models, and at the same time estimators of parametric components can be developed that exhibit standard parametric convergence rates. These two features have made semiparametric models and estimators increasingly popular in applied economics. We give a partial overview over the literature on semiparametric modelling and estimation with particular emphasis on semiparametric regression models. The main focus is on developing two-step semiparametric estimators and deriving their asymptotic properties. We do however also briefly discuss sieve-based estimators and semiparametric efficiency.