Abstract:
This paper aims at providing an analysis of the dynamic properties of the CSA Isfol quarterly series on job vacancies, in order to assess the possibility of its use as a leading indicator for the employment and unemployment fluctuations. In particular, we study the cross-correlative structure emerging between definite sections of vacancies and the corresponding sections in the standard employment and unemployment series. The analysis employs both standard descriptive tools, and the modern econometric concepts developed for the multivariate analysis of non stationary time series. The empirical structure of the data is thus estimated employing the VAR-VECM methodology, while persistence profiles, impulse-response analyses and forecast error decompositions are employed for the assessment of the dynamical properties. The results suggest that the ISFOL-CSA vacancies series is a valid cyclic predictor for employment and unemployment both at the aggregate level and respect to some interesting decompositions of the aggregate series. As a consequence, a labour market index based on help wanted may constitute a valuable leading indicator of the labour market trends and cycles and of their skill biases implications.
