In this workshop we will review the most popular nonparametric and semiparametric estimation methods, focusing on cross section observations. Standard methods for model estimation in the social sciences rely heavily on functional form assumptions and distributions of unobserved components. Nonparametric and semiparametric methods relax these assumptions, reducing the risk of misspecification errors. These methods are applicable to a wide variety of estimation problems. We will illustrate some of these applications with examples, including density estimation, regression estimation, partially linear models, single-index models, and propensity score methods. Emphasis will not be on math but on how these methods are applied in practice. We will discuss implementation of the procedures in R.
Dr. Escanciano is Associate Professor of Economics and Adjunct Associate Professor of Statistics at Indiana University. He received the PhD. in Economics from University Carlos III in Madrid in 2004 and has been at IU since 2006. His research interests fall broadly into the area of econometric theory, with emphasis on specification testing, semiparametric estimation and identification and risk management.