The Maltese Beveridge curve : an analysis of its developments and shift factors

Abstract

This study provides an analysis of the developments of the Beveridge curve for Malta and tests for the statistical significance of the Beveridge relation and observed shifts in the curve, over the sample period of 2002Q1 to 2021Q4. Furthermore, the factors underlying statistically significant shifts are tested. This is done by first utilising an Autoregressive Distributed Lag (ARDL) model to test for the negative relationship between the unemployment rate and vacancy rate and to test for the statistical significance of shifts in the Beveridge curve. Furthermore, together with the ARDL, the Bounds testing model of Pesaren et al., (2001) is utilised to obtain an unrestricted ErrorCorrection Model (ECM), incorporating both the short-run and long-run relationships in one model. In the next step, the Probit model is used to test for factors that may contribute to shifts in the curve, together with a Logit model to check for the robustness of the estimations. Results from the ARDL and ECM estimations suggest that there exists a short-run and longrun negative relationship between unemployment rates and vacancy rates. Moreover, there has been a statistically significant inward shift in the Beveridge curve during 2016, implying that there is a lower unemployment rate for a given level of the vacancy rate. It resulted that the main underlying factor behind this shift is the change in the share of foreign population in Malta. This is mainly due to the influx of foreign workers, which is a phenomenon that has been on the increase in the past years as a result of EU accession.