Intersectoral Labour Reallocation and Employment Volatility: A Bayesian Analysis using a VAR-GARCH-M model

Based on the macroeconomic VAR model for total employment and sectoral employment shares developed by Campbell and Kuttner (1996) we extend the model to a multivariate ARCH in mean (ARCH- M) model. We investigate the question of whether volatile growth in sectoral employment shares has an impact on total employment. The estimation method we use is the Gibbs-Metropolis algorithm for a Bayesian vector ARCH (B-VAR) model. This model is a standard tool in Financial econometrics and was developed by Engle (1986). The Bayesian estimation gives an exact small sample solution. It is found that a model incorporating a GARCH-M structure performs better than a simple VAR. Moreover sectoral shocks can account for more than 60% of the variance of total employment growth within a VAR-GARCH-M framework.