In this short course I will overview estimation and inference in regression models of increasing parameter dimension, meaning that the parameter dimension K grows with the sample size N. In the first part, I will discuss basic and new results on least squares, quantile regression processes, generalized method-of-moments, and Bayes-type estimators under the condition that K/N tends to zero, including functional central limit theorems, strong approximations, and uniform inference. I will illustrate these results with econometric applications, such as inference on intersections bounds, inference under shape constraints, and instrumental regression. In the second part, I will discuss basic and new results on L1 penalized least squares, quantile regression processes, and m-estimators under the condition that K/N tends to infinity. I will illustrate these results with econometric applications, including instrument selection and growth regressions.
High dimensional econometric modelling
Date & Time
From: 14 April 2011
Until: 15 April 2011
Type
Masterclass
Venue
The Institute for Fiscal Studies
7 Ridgmount Street,
Fitzrovia,
London,
WC1E 7AE
Fitzrovia,
London,
WC1E 7AE
Prices
HE Delegates: £80
Other Delegates: £1300