Incentives for indefinite employment in Spain Statements and Comments. 1. Main elements of the policy DISCUSSION PAPER

Transcription

1 Incentives for indefinite employment in Spain Statements and Comments Luis Toharia Universidad de Alcalá 1. Main elements of the policy One of the most persistent features of the Spanish labour market in the last fifteen years or so has been the high proportion of temporary or fixed-term workers existing among employees. The two expressions ( temporary workers and fixed-term workers ) have been used synonymously in Spain, probably for two reasons: first, since 1984 fixed-term contracts could be used to fill normal, i.e. continuous in time, jobs; secondly, temporary help agencies were only legalized in The main characteristic of a fixed-term worker is that termination of employment is almost costless for the employer: some severance payment has to be paid but that termination cannot be contested in court by the worker. In the mid-1980s, the main problem of the Spanish labour market was the extremely high unemployment rate (which was over 20 per cent for the first time in 1985). An extensive reform of the Labour Code was passed in 1984, in an effort to curb the persistent decline of employment which then lasted for almost ten years. One of the most significant elements of that reform was the generalisation of an already existing, though severely limited, fixed-term contract (so-called the Employment-Promotion Temporary Contract, hereafter EP-TC) which firms could use on a quite free basis for any job whether permanent or temporary. Partly as the result of this policy, but probably mostly due to various other general economic and structural factors, employment grew in the late 1980s at unprecedented rates, although the labour market retained two abnormal features: first, unemployment remained relatively high (16 per cent by 1991); secondly, unstable contracts went constantly on the rise: in terms of the LFS definition, the proportion of employees holding fixed-term contracts increased from 16 percent in 1987 to over 33 percent in 1992 (see Figure 1). During this period, one of the elements of labour market policy were a number of subsidies which were given to specific groups although not particularly to those hired under openended ( indefinite ) contracts. Starting in 1992, policy-makers started to be worried about the magnitude of the temporary workforce. However, as unemployment remained high, any move towards restricting fixed-term contracts was considered dangerous for employment growth. Thus, in 1992, only a minor change was introduced, which extended the minimum duration of EP-TC from 6 to 12 months. The effect of this change was the opposite from what could have been expected: the number of temporary contracts sharply increased and their actual duration decreased, as firms shifted from the employment-promotion contracts to more purely temporary contracts (which had always existed in Spain), in principle only allowed for temporary tasks. Given the large number of these contracts, the control of their actual temporary nature was impossible, and this favoured the shift towards their use. In 1994, a more extensive labour market reform, probably as important as that of ten years earlier, aimed at changing the workings of the labour market, by reducing the power of law as 5

2 against collective bargaining, by reducing the actual extent of dismissal costs for open-ended workers (more by clarifying and simplifying the procedures than by altering severance payments), and also by restricting the use of the EP-TC, which became a residual possibility. The reform was probably too ambitious and some critics argue that trade unions bought it away by accepting small wage increases which actually implied negative growth rates of real wages. Figure 1. The evolution of the proportion of employees under fixed-term contracts (Source: Spanish Statistical Office, Labour Force Survey; 2005 weights unless otherwise indicated) Percentage of total employees Both sexes Both sexes weights Females Males 1994 reform 33,7% 1997 reform Law 45/ ,2% EPA (points indicated correspond to the second quarter of each year) In 1997, in the first year of the new right-wing (Popular Party) government, and somewhat unexpectedly, a nationwide agreement was signed between the social partners, the main objective of which was to increase employment stability. After several years of continuous employment growth, and with persistently high proportions of temporary workers (recall Figure 1), the social partners appeared to be ready to provide a new impetus to employment stability. Trade unions were willing to accept that a new open-ended contract, carrying lower dismissal severance payments (at 33 days per year of seniority as opposed to 45), be created. The new contract was to be applicable to new open-ended hires of people under 30 and over 45, as well as to long-term unemployed of any age and open-ended contracts signed as a conversion of previous fixed-term ones. At the same time, the government which converted the agreement into law, passed, as a measure aimed at fuelling the agreement and helped by the availability of extensive ESF funds under the Community Support Framework, a companion law which provided incentives in the form of lower social security contributions, to be applied almost exclusively to open-ended contracts (the only exception being temporary contracts for disabled people), with a broadly coincident coverage with that of the new contract. These reductions, which were to last for 2 years, were calculated in such a way that, over a 2-year horizon, the total cost of an openended contract, including severance payment at the end of the period, was marginally lower than the cost of a temporary worker being paid the same wage. 6

3 The policy under review, although applied in some instances in the earlier period, can be thus said to have started in 1997, because it was only then when the objective of reducing the extent of temporary work became dominant. The policy combines two different elements, which were mixed up, especially at the beginning of the policy, because of the overlapping between the groups to which the two measures could be applied: a specific, so called employment-promotion open-ended contract, which, under certain circumstances (i.e. in case of redundancy for economic reasons), carries a lower severance payment, and which carries no direct budgetary cost. a reduction of social security contributions paid by employers, which has an obvious budgetary cost for government, but which is a tax expenditure, i.e. it is not a disbursement by government but rather an ex-ante deduction taken by firms when filing their Social Security records. The precise elements of the policy were to be determined on a yearly basis as part of the employment promotion programme included in the General Budget of the State. This policy was passed in 1997 as a temporary measure, to be assessed four years later. It was then converted into a permanent law in 2001, although the reductions of the social security contributions were still to be adopted on a yearly basis (the link with the general government budget makes this necessary). One important element introduced in the 2001 law was that a general gender perspective was added to the policy. Traditionally (since 1992 at least), women in occupations where they were underrepresented had been the subject of specific incentives, although the intake in these programmes was marginal. However, after 2001, all women between 16 and 45 years of age could also benefit, if hired in an open-ended contract, with a specific reduction of the social security contributions. A number of more specific groups of women (e.g. those subject to gender violence, women unemployed up to 2 years after giving birth) were also added as target groups. The 2006 reform has maintained these special target groups as they were included as part of other gender-related policies. In general, it can be said that the reason why women have been included as specific target groups has been that their labour market situation is weaker, with lower participation rates, higher unemployment rates (especially for the lower educated) and higher incidence of temporary work. It should be mentioned that while in the early years (up to 1999), the policy was financed to a significant extent by the ESF funds, the increasing surplus of the unemployment insurance system tended to become the main and then the only source of financing of this policy. This situation has a downside, however, because unemployment benefit taxes are largely paid by employers, which means that they may tend to consider the surplus their property and ask for rebates in the form of a reduction in social security contributions. As a matter of fact, the employers preferred policy would be an overall reduction of social security contributions, although they tend to settle with some other form of rebate inasmuch as the total amount is large enough. Although not directly related to the policy under review, there was a legal change in 2002 (Law 45/2002) which is significant because it altered the balance between the dismissal costs of openended and temporary workers, usually considered as the main element underneath the high use of temporary contracts in Spain. An apparently small change in the regulations of unemployment insurance, whereby an employer which would accept that a dismissal was unfair and would deposit the amount of the legally established severance payment in courts within 48 hours after 7

4 the dismissal, would not be liable for the intervening wages, i.e. wages corresponding to the period between dismissal and the final court (or previous agreement) settlement. This implied not only a reduction in severance payments of between 1 to 2 months wages but also created de facto a system of dismissals-at-will, even though not costless, but with a significant certainty about those costs. The importance of the 2002 reform is that it provides a sort of natural experiment in the reduction of dismissal costs and its labour market consequences. The number of dismissals substantially increased in 2003 and 2004, despite the strong overall employment growth. Around three of each four dismissals followed the procedure established in the 2002 law. However, the rate of temporary work did not change and the number of new hires under open-ended contracts did not experience any significant alteration in its trend. These results cast significant doubts on the view that the main factor behind the high incidence of temporary work is the dismissal costs differentials. Of course, it is so by definition, but the 2002 example means that short of a full wipeout of dismissal costs for open-ended contracts, whereby they would become temporary contracts, a partial reduction of the cost gap does not seem to be very effective. This is important because one of the rationales behind the policy under review is such a reduction of the gap in hiring costs for temporary and open-ended contracts, and the design of incentives has always tended to make calculations so that hiring on a open-ended basis would become less expensive that using a temporary contract (for an otherwise similar worker, a very important ceteris paribus condition very unlikely to be met). The policy has received a recent boost, as a consequence of the agreement between the social partners and the government reached in May 2006, and translated into a Decree-Law in June This reform will be discussed more extensively by the Spanish government representative. 2. The results of the policy Assessing a policy such as the establishment of incentives to open-ended employment is a difficult task. If the objective of the policy is reducing the proportion of temporary workers, then only that indicator should be considered. However, since that would be an indirect result of the increasing use of open-ended contracts, a more direct assessment of the immediate results of the policy itself should also be considered. 2.1 Global assessment Recalling Figure 1, it seems that, as a measure to reduce the share of temporary employees, the policy has been a failure. An important caveat should be made, though. The evolution of temporary employment in recent years has been affected by the change in the weights used by the LFS, to better take into account the huge increase of foreigners into Spain since 1995 and especially since Thus, in Figure 1, the black line represents the evolution of the percentage of temporary employees with the weights calculated in 2002, which did not take properly into account the recent inflow of immigrants. To some extent, thus, this line could be seen as a better indicator to assess the effectiveness of the incentive policies in terms of temporary work. In any case, even on those grounds, the global result of the policy cannot be considered fully successful. 8

5 It is interesting that, despite, these rather limited results, the main focus of the policy has been maintained by the recent agreement signed in June 2006 by the government and the social partners. One of the novelties of this agreement has been the emphasis on conversions, for which an extraordinary, though time-limited, programme has been implemented. 2.2 Direct policy assessment A second way of assessing the results of the policy would start from its direct results in terms not only of the total number of open-ended contracts, since the ultimate end of the policy is to reduce temporary work by increasing the number of open-ended contracts, but also of the number of contracts benefiting from the incentives provided. In this section, these issues are discussed and evidence on them is presented Total number of contracts Table 1 presents the evolution of open-ended contracts signed between 1992 and 2006 (the figure for this latter year has been extrapolated from actual figures for the first three quarters). To put the numbers in adequate context, a comparison is made between them (the inflows into open-ended employment) and the actual variation in the stocks of open-ended workers (as estimated by the LFS), which is the result of these inflows less outflows, the latter being unknown. In Table 1, the outflows numbers are the estimations which would make the numbers of inflows and stock variation consistent. This consistency is not necessarily achieved due to the differences in methods of data collection between the LFS and the administrative data on which contract numbers are based; however, consistency is not the issue here, but rather the evolution of the estimated outflows. The most interesting figures in Table 1 are those referring to the post-1997reform period. Starting in 1997, the number of contracts sharply increased, but the change in the stock did not follow, thus implying an ever increasing estimated outflow. While the initial increase of outflows in 1997 might be attributed to earlier underreporting, the increase in later years cannot. In percentage terms over the initial stock level, estimated outflows soared from 5 percent in 1997 to over 11 percent in 2001, stabilizing somewhat in recent years. On the whole, therefore, the policy of promoting open-ended contracts has met a very high success in terms of the accession of workers to such contracts. Considering the nine-year period , a grand total of almost 11 million open-ended contracts were signed, to which an estimated figure of 2 millions for 2006 should be added. Considering that the average number of open-ended workers at any given moment of time was 8.7 million, it is clear that entry into openended employment has been extremely high in Spain. The whole point of Table 1, however, is why is it that such a large inflow has not translated itself into higher stocks, i.e. how is it possible that the share of temporary workers in total employees has remained so high, under those circumstances. The only explanation is provided by Table 1: inflows into open-ended employment increased sharply, but so did outflows. In general, outflows are the sum of the natural wastage of the stock of open-ended contracts, mostly due to retirements, dismissals of workers with an open-ended status, for whatever reason, and attrition of new open-ended workers for whatever other reasons. Retirements have been more or less 9

6 constant at some 200 thousands persons per year, while dismissals have been on the rise, especially since But the numbers do not appear to turn up right. We shall come back to this point later on. Table 1. The evolution of employees under open-ended contracts in Spain, : stocks according to the LFS, inflows (contracts registered with the PES) and estimated outflows for consistency between stocks and flows (Source: LFS, PES) Time(t) Stock (first quarter of t) Change of stock between t and t+1 (1 st qtr) Inflows (contracts signed in t) Estimated outflows in t ,8-143,3 284,7 428,0 6, ,5-175,9 222,2 398,1 6, ,6-1,5 204,3 205,8 3, ,1 226,3 367,0 140,7 2, ,4 384,6 354,4-30,2-0, ,0 375,6 707,5 331,9 5, ,6 539,1 971,0 431,9 6, ,7 646, ,3 571,4 7, ,6 441, ,4 767,1 9, ,9 344, ,1 960,1 11, ,9 502, ,0 780,5 8, ,3 361, ,8 908,1 9, ,0 374, , ,0 10, ,7 390, , ,5 11, ,0 Outflows as percentage of initial stock Evolution of specific types of contracts The numbers in Table 1 correspond to all kinds of open-ended contracts. The rationale for including these total figures was that the purpose of the policy was to promote the use of such contracts, and while not all of them could benefit from the advantages established in the regulations, it could be argued that externalities across different types of open-ended contracts were probably there, so that all open-ended contracts, even those not formally affected by the policy, would benefit from the incentive policies. In any case, a full assessment of the policy requires a breakdown of the number of contracts signed by type of contract. Table 2 thus presents the distribution of contracts by type; in Table 2, the number of temporary contracts signed is also included, to provide a full image and also as a reference. In this table, up to 2001 part-time open-ended contracts are a separate category because they were so classified even though they might benefit from incentives in the same way as full-time contracts. After 2002, this anomaly was corrected and the contracts were properly classified. 10

7 On of the significant results in Table 2 is the fact that ordinary contracts (i.e. contracts with higher dismissal costs and no reductions of social charges) have been on the rise together with employment promotion contracts. An interesting shift took place between 1999 and 2001, when conversions were for a while partially excluded from the policy benefits. This implied a reduction in the number of conversions in 2000 which went hand in hand with an increase in the number of initial contracts. When conversions were again included in the policy, their number sharply increased while initial employment promotion contracts decreased. Table 2. The evolution of contracts registered with the Public Employment Services, Spain, (Source: SPEE; estimates for 2006 by author, based on information up to May and up to September) CONTRACTS OPEN-ENDED (open-ended + Employment TEMPORARY temporary) TOTAL Ordinary Part-time Conversions promotion Initial (est.may) 2006 (est.sept.) Note: employment-promotion refers to all kind of contracts subject to some kind of incentive; prior to 1997, some of these were lump-sum; after 1997, there were two kinds, as explained in the text: lower dismissal costs and reduction of social security charges. Given that the 2006 reform was passed in early June, two estimates for the 2006 numbers are presented. The first one is based on the number of contracts up to May 2006, assuming that the rest of the year would follow the same pattern as in The second one uses the information available until September, and is based on estimates for the last quarter of 2006 on the basis of the comparison between the fourth and the third quarter of Thus, the second estimate captures the effect of the 2006 reform, while the first one can be considered (with some caveats, to be explained later) the counterfactual that would have taken place without the reform. Under these circumstances, it can be argued that the reform has sharply increased the number of conversions and also to some extent the number of initial contracts, while leaving roughly unaffected the number of ordinary open-ended contracts. It is also interesting to note that the reform appears to have reduced the number of temporary contracts, although some caution should be applied here, as this reduction could be reflecting an early overshooting of temporary contracts in the early part of 2006, as it was known that subsidies to conversions would be 11

8 adopted in the final reform (the agreement was the result of more than one year of negotiations, and many of its aspects had been known, at least on general grounds, before). To complete the analysis of the distribution of contracts, Figure 2 provides a graphical representation of the percentage distribution of open-ended contracts. The growing share of conversions in total open-ended contracts is made clear by this figure, especially after 2001, when conversions were once again recognized as one of the target groups of the incentive policies. Figure 2. Percentage distribution of open-ended contracts, by type, (Source: Table2) 100% 95% 90% 85% 80% 75% 70% 65% 60% 55% 50% 45% 40% 35% 30% 25% 20% 15% 10% 5% 0% Conversions Employment promotion Initial Part-time Ordinary (est.may) 2006(est.sept.) Table 1 and Figure 2 lumped together all kinds of incentive contracts, without duly separating the elements related to lower dismissal costs and to lower social security contributions. Actually, there are no sufficiently good statistics before 2002 which would allow such a breakdown. Fortunately, recently, in the wake of the 2006 reform, the Spanish Public Employment Service has made available the distribution of contracts by separating the two elements of the policy. Table 3 provides such a breakdown, for the period Figures for 2006 are, as before, an extrapolation based on actual information up to September 2006 and based on the figures of the third and fourth quarters of The figures in Table 3 suggest a strong stability of the distribution of open-ended contracts by the different policy instruments available. Thus, roughly one-third of all contracts have not used any kind of incentives, be it lower dismissal costs or lower social security charges, with another 5 percent of contracts corresponding to higher-dismissal-costs/lower-social-charges; between 40 and 50 percent of contracts (with an increasing intensity) have corresponded to conversions, the share of which has jumped in 2006 under the impetus provided by the June reform; the remaining percent (decreasing over time) corresponds to initial lower-dismissal-costs contracts, most of which have benefited from reductions in their social security contributions. 12

9 Table 3. Distribution of open-ended contracts by use of the two elements of the incentive policies (Sources: data provided by the Spanish Public Employment Service statistical unit) (est.) ABSOLUTE NUMBERS Higher dismissal No incentive cost Incentive INITIAL Lower dismissal Incentive CONTRACTS cost 8 No incentive TOTAL INITIAL Higher dismissal No incentive cost Incentive Lower dismissal Incentive CONVERSIONS cost 7 No incentive TOTAL CONVERSIONS GRAND TOTAL PERCENTAGE DISTRIBUTION Higher dismissal No incentive 33,85 33,37 34,63 35,48 31,99 cost Incentive 4,95 5,28 5,31 4,95 4,32 INITIAL Lower dismissal Incentive 14,86 14,27 13,72 11,86 10,45 CONTRACTS cost No incentive 4,26 3,47 3,13 2,34 1,88 TOTAL INITIAL 57,93 56,38 56,79 54,62 48,65 Higher dismissal No incentive 5,16 5,48 10,04 13,82 7,62 cost Incentive 4,07 4,50 4,85 27,04 25,44 CONVERSIONS Lower dismissal Incentive 25,36 25,62 24,95 4,07 15,53 cost No incentive 7,47 8,02 3,37 0,45 1,91 TOTAL CONVERSIONS 42,07 43,62 43,21 45,38 51,35 GRAND TOTAL 100,00 100,00 100,00 100,00 100,00 Conversions provide, once again, an interesting evolution. In 2004, the regulations regarding the possibility of using the lower dismissal cost contract for conversions expired, as the deadline established in the 2001 law was reached at the end of 2004 and the new Socialist government did not extend it. Despite this restriction, however, conversions were still done, and their growth rate in 2005 was similar to that observed the year before. In fact, in 2005 the share of conversions in total open-ended contracts even increased. This suggests that dismissal costs are not a significant decision variable for employers, and that granting this incentive carries a significant deadweight loss, as firms would enter the contract even without the incentive provided. There is no similar experiment as regards the reduction of social security contributions, but there is a strong presumption that similar results would be found The duration of the various open-ended contracts One important problem is the lack of adequate longitudinal evidence on the life of open-ended contracts. Too often, it has been assumed that open-ended was the same as permanent (there is a strong tendency in the economic literature to speak of permanent employment) and therefore 13

10 not much research has been devoted to determine the extent and the causes of these outflows from open-ended employment. One exception comes from evidence based on open-ended contracts for the period (Cebrian, Moreno and Toharia, 2004). Empirical Kaplan-Meier estimates of the duration of different types of open-ended contracts, presented in Figure 3, show that the 1997 reform made open-ended contracts in general less stable; this was particularly true for employmentpromotion contracts (here encompassing both contracts subject to incentives and to lower dismissal costs) and less so for conversions from temporary contracts. More sophisticated econometric duration models confirm these differences. Figure 3. Kaplan-Meier survival profile estimates of the duration of open-ended contracts, by starting year and type of contract (Source: Cebrián, Moreno and Toharia, 2004, calculated from linked micro data contract and unemployment insurance databases provided by the Public Employment Service) Survival Ord Ordinary 1997-Conversions 1997-Empl.Prom Ordinary 1998-Conversions 1998-Empl.Prom Ordinary 1999-Conversions 1999-Empl.Prom One significant implication of this analysis has to do with the recent emphasis on conversions, seen as a double-edged instrument. On the one hand, as Figure 3 clearly shows, conversions appear to be the most stable open-ended contracts. On the other hand, providing incentives to conversions is seen as an element favouring temporary contracts, something which might have happened in Spain in early 2006, when the reform was expected to be passed with a sort of deadline for conversions. Unfortunately, no more recent studies exist on this issue, this being one of the main lacking elements for a full assessment of the incentives policies. Days 14

11 2.2.4 The cost of the policy As already argued, the policy under review has a limited budgetary cost. Its first element affects the expected cost which firms would have to face in the case of a dismissal, which obviously has no budgetary repercussion. The second one is a reduction in the social security contributions which firms have to pay when hiring employees, and this does have a clear budgetary cost, even though the fact that it is a tax expenditure (i.e. an ex ante deduction directly calculated by firms) may make it difficult to calculate very precise and detailed numbers on the actual cost of the policy. Only general, aggregate figures exist, and no figures on the actual cost for the various contracts involved are available. This is another instance for which the assessment faces important holes of information. One problem complicating the calculations is that incentives have been provided for a two-year period (the 2006 reform extends this period to 4 years). Estimations available come from a simple stock-flow model implemented by the Public Employment Service, but this requires an estimate of the survival of different types of contracts over time. On the whole, it is calculated that some 0.3 percent of GDP is spent on this incentive policies, a figure which makes it the largest amount for all active labour market policies in Spain The gender perspective As mentioned before, since 2001 women have been a specific target group of the policy under review. Although information on the different target groups can be collected, it is not particularly easy to compute and requires some tedious elaboration. Only recently, the Public Employment Service has begun publishing the breakdown of the contracts by target groups. For example, for the first nine months of 2006, women benefiting from a reduction in their social security contributions represented half of total initial contracts for women, although this percentage dropped to 15 percent when considering conversions as well. 3. The policy debate In the two preceding sections, the policy under review has been presented in terms of its main characteristics and of its main results. What are the policy issues raised by the measures adopted, given the (limited) results achieved? 1. A first question is that there is a general agreement across Spanish society that the extent of temporary work is far too large, even if one takes into account the specificities of the Spanish economy, where tourism and construction are two of the leading economic sectors. The problem is aggravated by the persistent and significant differences existing within Spain, once again not explained by economic or structural factors. This first point is very important to understand the Spanish debate regarding the policy under review. Spanish society in all its layers believes that government and the social partners have to take some action to curb this situation. 2. This issue has acquired a more significant weight in the policy debate given the strong employment and economic growth experienced by the Spanish economy over the last decade or so. Thus, in 2004 when the social partners and government started the process of 15

12 social dialogue which was to lead eventually to the 2006 agreement and labour market reform, they appointed an experts commission and asked from them a report providing a diagnosis of the temporary work situation. The report, issued in early 2005, emphasised the good situation of the Spanish labour market as well as the importance of balancing labour market flexibility with workers employment security, broadly understood as the possibility of holding a job for a long period of time, although not necessarily always the same job. 3. While there is a general agreement over the nature of the problem the existence of a high proportion of workers under a temporary contract- there is much less coincidence as regards the roots explaining it, and hence the policy recommendations to tackle it. At least three views may be mentioned. a. First, the strictly economic argument puts the emphasis on the cost differentials between a temporary and an open-ended worker, in terms of the higher dismissal costs which the latter would carry should a firm need to adjust its work force. Needless to say, this is, by definition, correct, in the sense that what defines an open-ended contract is the right to contest a dismissal at the courts and to receive a substantial severance payment if the dismissal is ruled to be unfair (the fact that very few dismissal procedures do actually reach the courts, as they are settled at en earlier stage, is not relevant here; what matters is that they could go that far) b. Secondly, the strictly legal argument puts the emphasis on the fact that the law is not duly applied by firms when using temporary contracts. According to current Spanish law, temporary contracts may only be used under duly warranted circumstances, in terms of the temporary nature of the tasks involved. If the use of temporary work is higher than justified on these temporary tasks grounds, it is because the administrative and legal controls over the use of labour contracts is inadequate. Part of the problem may be that the law still has loopholes, and part may be due to a lack of control over the labour market. One of the loopholes which has gained adepts in recent times is the increasing situation of productive decentralization whereby firms tend to subcontract part of their activities on a temporary basis, thus providing a legal justification for the use of temporary labour contracts. c. Thirdly, there is an intermediate view, which points towards a specific factor of the Spanish situation, namely the fact that during over a decade Spanish firms were allowed to more or less freely use temporary contracts without restrictions. This created a learning process which is difficult to unwind. A sort of path dependence problem arose from the 1984 reform whereby firms learned that temporary contracts were the ones to be used first and foremost, no matter under which conditions open-ended contracts could be entered. A sort of temporary work culture pervaded Spanish society, and reversing the situation is not an easy task. d. Empirical evidence on these views is mixed. All of them are problematic in terms of the actual labour market experience. The 1997 reform reduced dismissal costs for a number of workers, and the proportion of temporary work tended to decrease, although very modestly. However, the 2002 reform also decreased dismissal costs and it had no effect on the proportion of temporary work. The successive legal reforms since 1992 have reversed the situation and to a large extent have disabled the changes introduced in 1984; yet the proportion of temporary work has not decreased, although it could be 16

13 argued that it stopped increasing. In addition, the legal framework is exactly the same across the different Spanish regions, and the differences in terms of the incidence of temporary work are enormous (for example, the rate is over 50 percent in the private sector in Andalusia, and is only a few points above 20 percent in Catalonia or Madrid). As for the third view, the most challenging evidence against it is the short duration of the new open-ended contracts. Firms appear to have been using these contracts as substitutes for temporary ones. Should this be so, they would not appear to be particularly concerned about the use of open-ended contracts; rather they would prove that they know very well its possibilities and the easiness to terminate them. 4. In these circumstances, and given the healthy situation of the unemployment benefit system, which has been experiencing surpluses during a number of years now, the government and the social partners have decided that a cost approach is probably a solution worth being maintained, despite its clearly limited results: finance is there, a general agreement that the available surpluses should be returned to employers in one way or the other, and a general feeling that this cannot make things worse exists. This raises the issue, of course, as to whether the available funds could have been spent in ways more conducive towards stable jobs. The answer, given the political constraints of the Social Dialogue, is probably negative. As a matter of fact, this has not been a controversial issue in any way in Spain. 5. Provided that the policy of incentives in terms of a reduction of social security contributions is accepted as the only politically feasible policy, how should it be implemented? An interesting development has taken place in Spain. Between 1997 and 2004, the number of groups which ended up benefiting from social charges reductions sharply increased, as this incentive became a standard element of any policy directed towards any specific group, no matter how unrelated to the labour market (to provide just an example, women victims of gender violence are one of the groups who can enjoy these reductions). All of these groups enjoyed a reduction in their social charges at a rate which could vary between 20 to 50 percent and for periods which varied, with 2 years being the duration for directly labour market related groups. In these circumstances, three issues have been raised: i) whether the number of groups eligible for the reductions should be simplified and reduced; ii) whether the incentive should adopt the form of a percentage reduction as has been the case in the past or a lumpsum independent of the wage to which it is applied; and iii) what is the desirable length of the incentive provided. The 2006 has adopted the view that the number of groups should be reduced as much as possible, that the incentives should be given as lump sums and that the duration of the reductions should be extended as compared to the previous situation, from two years to three (in the case of conversions) or four (for other contracts). While this is the solution which has been adopted, the issues remain as open questions for the future. 6. The final issue refers to the evaluation of the policies. At the macro level, the main indicator is, of course, the proportion of employees under fixed-term contracts. The consistency between the evolution of stocks, as given by the LFS, and the observed flows is also an important indicator to be followed, especially in terms of the high outflows estimated in recent years. A third element refers to the budgetary cost of the policy. There are objective difficulties to collect the necessary information to publish the breakdown of the cost by type of contract, as well as other significant variables (gender, age, region, and the like). At the micro level, much remains to be done in terms of the longitudinal follow-up of workers under different types of contracts, in order to better understand why the outflows from open-ended contracts are so high. More specifically, the gender dimension would benefit from a more 17

14 detailed analysis which might allow an assessment of the actual labour market experience of women benefiting from the reduction in their social security contributions. 4. Final comments This paper has presented the policy adopted in Spain to foster the use of open-ended or indefinite contracts, as an intermediate step towards the reduction of the proportion of employees who are currently under temporary or fixed-term workers. This kind of policy has been in effect in Spain for many years but more significantly since 1997, when it became one of the main instruments of labour market policy, a situation fuelled first by the availability of ESF funds and later by the high surpluses enjoyed by the unemployment benefit system. The results of the policy have been, so far, quite limited. The extent of temporary work remains high in Spain and although the number of contracts benefiting from the policy has been very high, it has met with increasingly high outflows from open-ended contracts, the nature of which is not fully understood. In any case, this is a policy over which there has been agreement between the social partners and the government, and it has received a significant boost with the recent June 2006 labour market agreement and reform. In terms of the evaluation of the policy, much remains to be done, both in terms of the collection of aggregate expenditure figures and in terms of the longitudinal experience of workers affected by the policy. On both issues, the role of the Social Security Administration appears to be crucial as it is the agency which has access to the information required for the evaluation tasks. References Cebrián, I., Moreno, G. and Toharia, L. (2004), Are open-ended contracts permanent? Evidence from Spain, presented at the XIV Conference of the European Association of Labour Economists, Lisbon. Toharia, L. (2005), El problema de la temporalidad en España: un diagnóstico, Madrid, Ministry of Labour and Social Affairs. 18