Participación laboral y ciclo económico

Abstract

This thesis consists of four self-contained essays structured as follows. It mainly consists of two parts leaving aside this introduction. Part II includes chapters 2 to 4, three chapters devoted to provide new empirical findings about the cyclical relationship between unemployment and labor force participation. Chapter 2 looks for linear and non-linear relationships between unemployment and labor force participation rate in Spain and the US, using quarterly data _ F.snnpla He Dnrtnrario from 1976:3 to 2015:2. In this chapter we revisit how participation rates are affected by business cycle fluctuations, while accounting for different sources of asymmetries. In particular, we explore the potential existence of asymmetries into two directions by age groups and checking whether the relationship is time varying. By using Spanish quarterly data over the period 1976-2015, results point to the existence of nonlinear relationships between unemployment and labor force participation for different age groups. In particular, we provide evidence for a nonlinear discouraged worker effect for the oldest and middle-aged groups, while an added worker effect is found for youth when the unemployment rate is below a threshold. Chapter 3 examines studies the long-run relationship between unemployment and labour force participation rate, from an alternative way, by taking into account possible different sources of asymmetries: structural breaks and the possible existence of asymmetries by gender and age groups, using Spanish data. In the last regime, the current one, our findings support the discouraged for the aggregate labour force participation rate. However, the analysis by gender reports mixed evidence -discouraged for males and added worker effect for females-. Finally when we consider the asymmetries by age, the evidence is mixed again: discouraged for the younger added for the middle age group and invariance for the older. Chapter 4, re-examines the relationship between the labor force participation rate and the unemployment rate for Spain, looking for Granger causality relationships. The novelty of this study is to explore asymmetries in several ways: by gender, by age and allowing time dependence by using the new tests recently proposed by Hatemi-J (2012). This analysis completes the previous ones, carried out in this dissertation, in order to provide a comprehensive view of the exact nature of this relationship. The article finds that there is either bidirectional long-run Granger causality running from labor force participation to the unemployment rate for the aggregate and for gender and age groups, except for females and for the middle age group. However if time-asymmetries are also considered, different causal relationships prevail for different groups and for every regime. Finally, the contribution of the Part III, deals with the long term. Chapter 5 estimates an unobserved components model to explore the existence of hysteresis in labour participation. In particular, this approach should allow disentangle the effect of unemployment evolution in different business cycle phases on labour force participation -the cyclical effects summarised in the literature in discouraged and added worker effects- and the permanent part, the part that exhibits resilience and it is incorporated to the natural component of the participation rate. In addition, and by using an extended version of this model -including non-linearities- we also check the potential existence of different estimates of this relationship in different periods -in expansions and recessions-. By using this approach, the participation effect of changes in the unemployment rate is analysed. The analysis allows to look not only for cyclical effects -i.e. testing added, discouraged or invariance worker effect- but also for hysteresis. In sum we use an alternative framework for exploring the dynamic effects of downturns and upturns on labour force participation both in aggregate and by gender, that is the greater/lesser responsiveness labour force participation to downturns/upturns and whether this shocks show persistence. The empirical analysis of the labour force participation rate is a hot policy issue at the time of writing, specially in countries like Spain, where not only the deep depression of the economic activity but also the existence of a strong added worker effect -empirical findings of Congregado et al (2014) pointed to this direction among females- are behind the exponential growth experienced by the Spanish unemployment rate. Now, where Spanish economy shows symptoms of recovery and the unemployment rate is starting to fall, the bad news is that it is not the consequence of higher employment but the consequence of a fall in the participation rate - discouraged effect seems to be dominating now the net effect and no hysteresis seems to be present-. Our results provide robust evidence of hysteresis for the aggregate and for males but not for females, in the linear version. Attending to the cyclical effects our linear unobserved component model also provide evidence of a discouraged effect among males and invariance in the total and females labor force participation rates. However, when the nonlinear model is considered, then we provide evidence of hysteresis not only for males and total but also for females, although in the positive regime for males and for the total. The evidence on the discouraged and invariance hypotheses in this nonlinear model confirms the results obtained in the linear one. The study concludes with a final chapter, containing some concluding remarks and the future research agenda.