Multivariate Time-series Models

Peter A. Zadrozny's research is largely devoted to solving computational and statistical problems arising in the estimation and use of multivariate economic time-series models. With regard to computational problems, he has developed and applied a general method for accurately and quickly computing solutions of linear rational expectations models. Regarding statistical problems, he has developed and applied maximum-likelihood and optimal-instrumental-variables methods for estimating multivariate time-series models when data have complications such as different variables having different observation frequencies or different publication delays.

At Ifo, Mr Zadrozny plans to continue joint work with Klaus Wohlrabe and Stefan Mittnik on a method of efficiently modelling volatilities of multiple time series with a new method based on dynamic factors. He also plans to work on "An Econometric Method for Decomposing Total-Input Productivity Into Input-Specific Productivities", which he will present in a seminar.

Peter Zadrozny received his PhD in Economics from the University of Chicago and taught economics at New York University, the University of Houston, Washington State University, the University of Warsaw and Johns Hopkins University. He has also been a researcher at the Center for Economic Studies at the US Bureau of the Census in Washington, DC. Since 1998, has been doing statistical and economic research at the US Bureau of Labor Statistics in Washington, DC, where he is now in the Division of Price Index Number Research. Mr Zadrozny is an associate editor of the journals Computational Economics and Journal of Time Series Analysis. Among other journals, he has published papers in Econometrica, Econometric Theory, Journal of Economic Dynamics and Control, Journal of Productivity Analysis, Journal of Time Series Analysis, Computational Economics and Economics of Planning.