Document Actions

BWPEF 1606
Andrea Nocera
EM Estimation of Dynamic Panel Data Models with Heteroskedastic Random Coefficients

Abstract
In this paper, we show how to combine the EM algorithm with the Restricted Maximum Likelihood (REML) method to estimate iteratively both the average effects and the unit-specific coefficients as well as the variance components in a wide class of dynamic heterogeneous panel data models. The estimation procedure can also be adapted to allow for cross-section dependence. Compared to existing methods, our approach allows for heteroskedastic random coefficients, and leads to an unbiased estimation of the variance components of the model without running into the problem of non-positive definite covariance matrices typically encountered in random coefficients models. Monte Carlo simulations reveal that the proposed estimator has good properties even in small samples. A novel approach to investigate heterogeneity of the sensitivity of sovereign spreads to government debt is presented.

Keywords: EM algorithm, restricted maximum likelihood, correlated random coefficient models, dynamic heterogeneous panels, debt intolerance, sovereign credit spreads.

JEL Classification: C13, C33, C63, F34, G15, H63.

View this paper.