Evaluation of Dynamic Participation in Portuguese Active Employment Programmes

Over the past decades, European labour market policies, like the Portuguese, offer not only a set of simultaneous active programmes, to the universe of unemployed registered individuals, but also the possibility of successive participation in different active programmes. Such dynamic participation patterns have as a consequence different average causal effects to participants. Aware of this reality, the present paper discusses how to overcome the limitations of the traditional labour market microeconometric evaluation literature, eminently static, and empirically estimate the impact of dynamic selection. In recent years it is possible to report several developments concerning the introduction of dynamic elements in the traditional evaluation econometric models. For instance, is important to report that the causal effects of dynamic sequences make use of potential results and allow the introduction of intermediate results which determine subsequent sequences. The application presented in this research study uses an administrative dataset containing extensive individual information concerning the active public intervention on the Portuguese labour market. Based on this dataset, it will be shown that it is possible to make a quantitative analysis on the impact participation of an unemployed individual on a sequence of active measures offered by the unemployment office services. The present empirical study analyses the power of the dynamic nonparametric model as a fundamental tool for an active labour market policy evaluation, and allows, even if in an exploratory way, the evaluation of the consequences of sequential participation. Indeed, based on propensity score matching microeconometric evaluation techniques - computed through a probit model - is possible to compute the causal effect of the decision to postpone participation in an active labour programme vis-a-vis the decision of consecutive labour programme participation, on the one hand, and continuous non-participation, on the other. The evaluation results of a consecutive participation decision in comparison with the alternative of non-participation or postponed participation suggest a low degree of effectiveness by the Portuguese public intervention in the labour market. Moreover, the empirical application of a dynamic propensity score methodology seems to be a quite useful tool for the estimation of dynamic causal effects when the effectiveness of a sequence of participation decisions is tested compared to a distinct sequence of decisions.