A0161
Title: Estimation and inference in high-dimensional panel data models with interactive fixed effects
Authors: Oliver Linton - University of Cambridge (United Kingdom) [presenting]
Maximilian Ruecker - University of Ulm (Germany)
Michael Vogt - University of Ulm (Germany)
Christopher Walsh - Newcastle University (United Kingdom)
Abstract: New econometric methods are developed for estimation and inference in high-dimensional panel data models with interactive fixed effects. The approach can be regarded as a non-trivial extension of the very popular common correlated effects (CCE) approach. A projection device is first constructed to eliminate the unobserved factors from the model by applying a dimensionality reduction transform to the matrix of cross-sectionally averaged covariates. The unknown parameters are then estimated by applying lasso techniques to the projected model. For inference purposes, a desparsified version of the lasso-type estimator is derived. While the original CCE approach is restricted to the low-dimensional case where the number of regressors is small and fixed, methods can deal with both low- and high-dimensional situations where the number of regressors is large and may even exceed the overall sample size. Theory for the estimation and inference methods is derived both in the large-$T$-case, where the time series length $T$ tends to infinity, and in the small-$T$-case, where $T$ is a fixed natural number. Specifically, the convergence rate of the estimator is derived, and it is shown that its desparsified version is asymptotically normal under suitable regularity conditions. The theoretical analysis is complemented by a simulation study and an empirical application to characteristic-based asset pricing.
