بیمه بیکاری و ترک  کار در اسپانیا

The aim of this paper is to shed some light on the potential relationships between the unemployment insurance system and labour market turnover. This study assumes the incentives embedded in the unemployment insurance system have a heterogeneous impact, depending on the type of labour market transition (quits versus layoffs and recalls versus new job entrances) and on a worker's attachment to the labour market (gender and type of contract). The layoff hazard rate increases as workers qualify for unemployment benefits, whilst the quit hazard rate remains stable. Similarly, employment inflow increases sharply after the exhaustion of unemployment benefits. The timing and importance of the exit differ between recalls and new job entry and depend on a worker's attachment to the labour market. The results show that unemployment benefits appear to favour job turnover and both firms' and workers' decisions seem to matter.

The labour market is in a constant state of flux. There is a continuous flow of workers into and out of employment, and from one job to another. Understanding job turnover is the key to understanding how the labour market operates. Turnover is necessary because it helps allocate workers to those jobs where they are most productive and allows employers to hire and fire according to economic conditions. It is not always optimal, however. Some groups of workers experience high layoff rates without ever advancing to better positions (Rebollo, 2011 and Gagliarducci, 2005). And some groups of firms face high firing rates without improving their productivity levels (Dolado and Stucchi, 2008 and Bassanini et al., 2008). One of the factors blamed for excessively high turnover in the labour market is the design of the Unemployment Insurance System (UIS). Hence the aim of this paper is to shed some light on the potential relationships between the UIS and labour market. Accordingly, we analyse the Spanish labour market for the period 2000–2007. Several features distinguish the Spanish labour market from other European labour markets. Firstly, it has a generous UIS financed by uniform payroll taxes. Uniform payroll taxes are frequently criticised for giving rise to too many layoffs, reducing the mean duration of employment and increasing unemployment (see Anderson and Meyer, 2000, Cahuc and Malherbet, 2004, Fath and Fuest, 2005 and Blanchard and Tirole, 2008). Secondly, employment turnover is notably higher than in other European countries, with recent figures showing that nearly 50% of workers have held their current job for six months or less and almost 30% for no more than a year. Thirdly, more than 80% of newly signed contracts are temporary, and Spain's temporary employment rate has remained above 30% since the beginning of the nineties. Fourthly, more than a third of the unemployed who find a job return to their former employer. The effects of UIS benefits on job turnover compound labour supply and demand forces and their relative importance continues to be an empirical issue. A number of empirical studies have already examined how certain characteristics of the UIS play out with respect to the duration and outcome of unemployment spells. Typically, these previous studies show that higher replacement ratios lead to longer unemployment spells and that the probability of escaping unemployment increases as unemployment benefit entitlements are exhausted. To understand whether demand or supply incentives are behind this effect, the researcher must take into account whether the unemployment spell finally ends in recalls or in a new job entrance and how the UIS affects the length of employment. In relation to the first point, the idea is that recall versus entry into a new job may involve several different causal mechanisms, all requiring explicit consideration in the analysis of the effect the UIS has on job turnover (Katz, 1986 and Jurajda, 2002). In relation to the second point, Juradja (2002) has shown that evaluating the UIS by studying only its effects on unemployment duration may lead to an underestimation of the total impact the UIS has on job turnover and hence on the unemployment rate. The influence of UIS eligibility parameters on employment duration, in contrast, has received scant attention and none of the empirical studies found takes into account the potential behavioural differences between layoffs and quits. These distinctions between different types of employment inflow and outflow are key to determining whether the UIS also affects a firm's hiring and firing decisions (as implicit contract theory shows, see Feldstein, 1976) and not only workers' decisions, as assumed in traditional analysis. For instance, one could easily argue that layoffs are triggered by productivity shocks whilst quits are triggered by reservation wage shocks (Blanchard and Tirole, 2008). This paper conducts a more comprehensive analysis of the potential effects of the UIS on job turnover. Firstly, we take into account potential selection effects through the estimation of a multivariate mixed proportional hazard model—multiple spell and multiple states with competing risks—allowing for jointly-distributed unobserved heterogeneity. Although sample selection effects might be important in these types of analysis,1 few empirical papers take them into account. In particular, the analysis considers three distinct initial states: employment, involuntary unemployment, and voluntary unemployment. Secondly, we define a competing risk model for employment and unemployment spells as follows: within the state of employment, the analysis differentiates between quits (leading to voluntary unemployment) and layoffs (leading to involuntary unemployment). Within the involuntary unemployment state, we consider whether the spell ends in recall or the worker's entry into a new firm. Within the voluntary unemployment state we only consider exit to employment, as job quitters probably face zero recall expectations. Thirdly, given the strong duality of the Spanish labour market, we allow for heterogeneous effects of the UIS system between workers holding permanent contracts and those holding temporary ones. Finally, the whole analysis is performed separately according to gender. Although several dimensions of the UIS affect the labour market, we shall concentrate on two of its key components: Entry Requirement (ER) and Potential Benefit Duration (PBD). ER refers to the minimum number of weeks individuals have to work over a specified period in order to qualify for UIS benefits. PBD refers to the maximum number of weeks the unemployed worker is entitled to draw UIS benefits. In Spain, both parameters (ER and PBD) depend on the number of weeks worked over the six years prior to the onset of unemployment. The empirical method is to look for spikes in the employment and unemployment hazard profiles exploiting cross-sectional and longitudinal variations in ER and PBD parameters, respectively. As mentioned above, we allow these UIS parameters to differ between temporary and permanent contracts. Note that the influence of UIS benefits on search behaviour and reservation wage policy might differ depending on the type of contract. Another key feature of this analysis is the use of an administrative dataset (Longitudinal Working Life Sample, LWLS) that allows compiling full employment histories and analysing the distribution of employment and unemployment durations as affected first by ER and then by PBD. It is very important to use an administrative dataset in this type of analysis since it avoids the existence of seam bias,2 a serious problem for estimating duration models, as it affects the timing of transitions. Our results show that omitting the role of unobserved heterogeneity and dependence between the different states will hide the fact, amongst other things, there are certain types of workers whose labour market path is characterised by high exit probability from employment and long unemployment duration. The analysis presented also points to various behavioural consequences of the UIS on job turnover. Firstly, we observe that employment inflow and outflow are influenced by the UIS, varying the intensity of the effect by gender, type of contract and type of transition. In general, these effects stand out for those segments with a loose attachment to the labour market, such as females and temporary workers. Secondly, we show that employers might play a role in the timing of the layoff, as well as in the timing of the outflow from unemployment. Thus, the layoff hazard rate increases when the worker qualifies for UIS benefits; whilst job quit decisions remain unaffected. The exit rate from unemployment for benefit recipients records sharp increases around the time that benefits run out.3 Interestingly, the recall hazard rate peaks one month prior to the exhaustion of benefits for workers previously hired on permanent contracts. Meanwhile, the new job hazard rate peaks when UIS benefits run out. In light of these findings, it can be concluded that the observed ‘moral hazard’ effects the UIS has on employment and unemployment duration cannot all be attributed to worker reactions alone. Note that the importance of these results resides in the finding that the UIS tends to reduce the time an individual spends in employment throughout their labour market career. The rest of the paper is organized as follows. Section 2 describes the main characteristics of the Spanish UIS and Section 3 outlines the theoretical framework and reviews existing empirical literature. The data and econometric model are presented in 4 and 5, respectively. The results of the empirical analysis are given in Section 6. The paper's conclusions are summarized in the final section.

The current design of the UIS might provide incentives for workers and employers to increase labour market turnover since the UIS appears to have a negative effect on employment duration whilst increasing unemployment incidence. The study reported in this paper reveals that when an employee qualifies for unemployment benefits there is a spike in the layoff hazard rate, but none in the quit hazard rate. The UIS is also found to have a major impact on unemployment duration. The recall and new job hazard rates increase notably around the time of benefit exhaustion. Interestingly, the incidence of the UIS on employment and unemployment transitions is highest for females and temporary workers; that is, for workers with loose attachment to the labour market and who suffer the largest turnover rates in Spain. We have also presented evidence to show that the economic incentives explaining unemployment duration may differ depending on whether the shift to unemployment is due to a layoff or a quit and whether the nature of the layoff is temporary (ending in recall) or permanent (ending in new job entry). Hence, the results forthcoming show that workers and firms seem to have some influence on the timing of the outflow from both employment and unemployment and use it to their advantage whenever the current characteristics of the UIS allow. However, these incentives might generate excessive labour market turnover, with shorter employment spells and longer unemployment spells. Note that the importance of these results lies in the fact that the UIS seems to reduce the time spent in employment throughout an individual's working life by both directly increasing the probability of exit from employment and indirectly increasing unemployment duration. These findings need to be considered in the Spanish economy, in which over 80% of newly signed contracts are temporary and more than 30% of unemployed workers return to their previous firm. Given these results, a potential reform of the UIS designed to reduce the average unemployment duration and the unemployment rate should consider both sides of the labour market: on the one hand, the current design of the UIS distorts a firm's hiring and firing decisions, and on the other hand, it also has behavioural consequences on a worker's decisions.