What Drives Worker Flows?

Melbourne Institute Working Paper No. 34/07

Date: November 2007


Chew Lian Chua
Robert Dixon
G. C. Lim


This paper applies a multi-state latent factor intensity model to worker flows to obtain insights about the determinants of entry and exit rates pertaining to various labour market states. The analysis shows that one activity factor underpins the decision to move from employment and from unemployment and this result may be of special interest to policy makers concerned with understanding the rate of departures from the pool of both the employed and (especially) the unemployed. The paper also shows how to estimate a non-linear state space model using a Gibbs sampler that encompasses a Metropolis-Hastings algorithm as well as the auxiliary particle filter to estimate the latent process. The advantage of the approach is that it provides a parsimonious and efficient way to obtain key information about behaviour in labour markets.

Download Paper

  • Bayesian statistics
  • Gibbs sampler
  • Metropolis-Hastings
  • Auxiliary particle filter
  • Worker flows
  • unemployment rate.