This Selected Background Issues paper examines different aspects of the social security system in Spain. The paper describes the role of this system in the process of consolidating the government finances in the short- and medium-term, and to present possible measures to address the policy challenges of the longer term. An analysis of the finances of the social security system is presented. The paper provides a description of the main features of the pension system and projections of its financial situation in the long term.

Abstract

This Selected Background Issues paper examines different aspects of the social security system in Spain. The paper describes the role of this system in the process of consolidating the government finances in the short- and medium-term, and to present possible measures to address the policy challenges of the longer term. An analysis of the finances of the social security system is presented. The paper provides a description of the main features of the pension system and projections of its financial situation in the long term.

II. Labor Market Reforms in Spain: A Preliminary Evaluation 1/

1. Introduction

The single most important economic and social problem confronting Spain is unemployment. From relatively modest levels in the 1970s, the unemployment rate has grown to nearly one quarter of the labor force by 1994. Although the cyclical rise in unemployment during the recent recession was sharp, it has became clear that the much of Spanish unemployment is structural and persistent. 2/ This situation has prompted a number of government initiatives to reform the structure of the labor market. This paper reviews the recent reform efforts, and attempts to provide a preliminary assessment of their likely effectiveness. Particular attention is paid to the issue of dismissal costs, which are especially high in Spain. An annex to the paper explores the recent academic literature on the impact of dismissal costs on employment and unemployment.

The paper begins by providing background to the current labor market situation. Section 2 provides a review of the unemployment performance of Spain in recent years and sketches some of the important structural features of the labor market. Section 3 discusses the labor market reforms introduced by the Spanish government between 1992 and 1994. Section 4 uses recent data to attempt to get a preliminary idea of some of the effects of the reforms, though it acknowledges that it is too early to provide a full evaluation. Section 5 concludes with some speculations on the longer-term effects of the reforms and presents suggestions for further action.

2. Background

a. The evolution of unemployment

Unemployment in Spain demonstrated continuous growth from the late 1970s through the middle 1980s, with the unemployment rate climbing from under 5.0 percent in 1977 to 21.6 percent in 1985. This unprecedented rise was driven by five major factors. First, the early 1980s was a period of weak economic growth (averaging only 1.1 percent from 1981–85), and the relative stagnation was not conducive to employment growth. Second, the effects of population growth and increasing labor force participation by women produced a sharp rise in the labor force. Third, the structural transformation of the economy led to large declines in employment in agriculture and in certain basic industries (e.g., coal and steel). Fourth, the return to democracy brought with it a resurgence of trade union power an a sharp increase in the level and intensity of industrial action, which contributed to strong real wage growth and correspondingly weaker employment creation. Finally, and perhaps most importantly, severe rigidities in the labor market inhibited job creation; total employment in 1985 was 13.7 percent lower than the level in 1977 (Table 1; Chart 1). Among the most important distortions identified by recent research are wage rigidities caused by strong insider-outsider effects, restrictions on the flexible use of labor by firms, and generous unemployment benefits which reduce incentives to search for work.

Table 1.

Spain: Population and Unemployment, 1986–94 1/

article image
Source: IMF, Labor Force Survey.

End of year data except as indicated.

Annual averages.

Chart II-1
Chart II-1

SPAIN: Employment and Unemployment, 1976–94

Citation: IMF Staff Country Reports 1995, 039; 10.5089/9781451811988.002.A002

Spanish accession to the EC in 1986 sparked a major economic recovery, with growth averaging 4.5 percent per year from 1986–90. This expansion, combined with the government’s introduction of flexible temporary labor contracts in 1984, fueled an increase in employment averaging 3.0 percent per year. The growth in employment was reflected in a drop in unemployment; however, the unemployment rate did not fall as sharply due to a large increase in labor force participation (6.7 percent during the period). As a result, the unemployment rate failed to fall below 16 percent at the cyclical peak.

Employment growth turned negative in late 1991, as GDP growth initially fell below 2 percent. The economy dipped into full-fledged recession by the end of 1992 and the unemployment rate soared. The unemployment rate climbed from 15.9 percent in the second quarter of 1991 to 24.6 percent in the first quarter of 1994, a change of nearly nine percentage points (and an average increase of 0.7 points per quarter). With economic recovery, the unemployment rate declined slightly in the second and third quarters of 1994, falling to 23.9 percent in the third quarter. Despite this improvement, the annual average unemployment for 1994 will still be above 24 percent, the highest rate in any industrial country in recent memory.

b. The Structure of the labor market

The Spanish labor market has evolved into a singular amalgamation of the corporatist system of the Franco era and the social protections of a modern European welfare state. Many of the restrictions on functional and geographical mobility and high severance payments in place under the Franco regime were retained into the 1990s, and a new set of regulations was superimposed (such as generous unemployment compensation, minimum wages, and strong trade union bargaining rights). The resulting peculiar combination has had particularly negative effects on the functioning of the labor market, and has contributed to the elevated unemployment rates of the 1980s and 1990s.

(1) Hiring and contracts

The labor market in Spain has suffered from severe rigidities in the working conditions embodied in labor contracts. 1/ Traditionally, labor contracts have been notably lacking in flexibility; in the Franco era, neither part-time nor temporary contracts were widely encouraged, and permanent employees worked under strict rules regarding hiring, firing, and workplace mobility.

Today the labor market has become extremely dualistic; highly protected permanent employees coexist in the market with a large group of temporary workers, many of whom do identical jobs as their permanent counterparts at lower wages 1/ and with significantly less legal and social protection. After a short trial period (2–6 months maximum), a worker hired under an indefinite contract is considered permanent. Permanent workers are entitled to generous severance pay if they are fired or laid off, and they are generally eligible for unemployment compensation as well.

The dualism in the labor market began to gain importance in 1984, when the rules governing the hiring of temporary workers were loosened substantially as a way of increasing flexibility and boosting employment. From 1984, temporary contracts could be granted to workers for periods as short as 6 months and ranging up to 3 years. 2/ These workers, upon completion of their contracts, receive substantially less severance pay than their permanent counterparts. 3/ The short length of their contracts also gave employers much more geographical and functional flexibility in organizing their workforce, whereas permanent workers enjoyed tight legal constraints on their mobility.

As a result of the more flexible conditions of employment for temporary workers and their lower costs to employers, temporary workers became an important component of the labor force in the late 1980s and early 1990s. The different forms of temporary contracts grew to encompass roughly one third of the salaried workforce, accounting for more than 95 percent of new job contracts issued over the period.

(2) Working conditions

Working conditions in the Franco period were governed primarily by a set of Ordenanzas Laborales (“Labor Ordinances”) which regulated almost every aspect of the employee-employer relationship, from the working hours to functional assignments, to geographical mobility. The Ordenanzas were supplanted as the primary labor legislation in 1980 by the passage of the Estatuto de los Trabajadores (“Workers’ Statute”); however, many of the provisions of individual Ordenanzas were retained. In particular, the ability of managers to organize the production process continued to be severely impeded by elaborate definitions of occupational categories and restrictions on mobility between those categories. These restrictions on functional mobility were considered by the current government to be “probably the greatest flaws in the system of industrial relations … a system which in this respect was entirely the legacy of an archaic and corporatist interpretation of the ‘Ordenanzas Laborales…’”. 1/

In addition to the restrictions of the Ordenanzas Laborales, labor legislation itself maintained tight controls on geographical mobility, the length of the working day, number of permissible overtime hours, minimum vacation time, and other features of the employment relationship. Firms were required both to conduct lengthy consultation with unions and to obtain permission from the government in order to transfer workers from one location to another. If workers did not acquiesce to changes in their working conditions, or to geographical or functional mobility, they retained the right to quit with full access to severance pay (20 days pay per year of seniority), unemployment benefits, and with the right to challenge the legitimacy of the changes in the social tribunals. 2/

(3) Collective bargaining

Two aspects of the structure of collective bargaining are important to understanding the role of trade unions in the labor market in Spain. First, despite relatively low unionization rates, trade unions have a disproportionately strong impact on the determination of wages and salaries. Chart 2 illustrates this point. Despite the fact that only about 15 percent of workers belong to a trade union, more than 86 percent of all salaried workers have their wages determined by collective bargaining agreements. 3/ Trade unions in Spain are also particularly conflictive; the number of days lost in industrial action is quite high. Also shown in Chart 2 are days lost to industrial action, which peaked in 1988 when nearly 7 million work days were lost (nearly an average of one day per salaried worker). This rate is well above that prevalent in most other European countries.

Chart II-2
Chart II-2

SPAIN: Collective Bargaining and Industrial Action

Citation: IMF Staff Country Reports 1995, 039; 10.5089/9781451811988.002.A002

The second important aspect of collective bargaining concerns its structure. The evolution of bargaining since the transition to democracy has created a bargaining system which tends to perpetuate rigidity in the real wages. In the late 1970s and early 1980s, the unions, employers, and the government concluded a series of nationwide social pacts which helped to restrain real wage growth and to liberalize working conditions. Since the end of the last of these agreements, collective bargaining has been based primarily at the industry level, with industry-wide agreements automatically serving as a floor for any additional agreements at the firm level. Both theoretical research and empirical studies indicate that this type of “intermediate” level collective bargaining is particularly prone to real wage rigidity and thus contributes to high unemployment. 1/

(4) Dismissals

Among the most important rigidities in the Spanish labor market is the high cost of dismissals of workers with permanent contracts. 2/ Three aspects of dismissals are problematic: (1) Legally mandated severance pay is relatively generous; (2) the legally admissible criteria for authorizing dismissal are restrictive; and (3) the bureaucratic and legal procedures for authorizing dismissals are time-consuming and costly. The combined effect of these features is to make the dismissal of permanent workers very expensive, which in turn affects the hiring decisions of employers.

The labor laws divide dismissals into categories according to the number of workers affected (collective versus individual) and according to the justification for the dismissal (justified versus unjustified).

(a) Collective dismissals

Until the 1993–94 reforms, theoretically all redundancies were considered collective dismissals and required that the firm receive prior governmental approval via an Employment Regulation Procedure (ERP) with the Ministry of Labor. The employer was required to initiate a 30-day negotiation period with the trade unions over the characteristics of the redundancies. After the consultation period, the redundancy plan had to be approved by the relevant government labor authorities which had 30 days to rule on the plan. Employers could appeal an adverse ruling to the central Labor Ministry officials, who were given another 15 days to rule. Any dismissals initiated under the ERP regime involved a minimum severance payment of 20 days pay per year of seniority.

(b) Individual dismissals

In practice, individual dismissal procedures have been widely used for redundancies as a way of circumventing the time-consuming consultation and processes inherent in the procedures for mass layoffs. At the individual level, dismissals do not require previous negotiation, but severance pay must be made and the dismissals are still subject to appeal by the worker. Justified firing (despido procodente) requires a minimum severance payment of 20 days of wages per year of seniority up to a total of 12 months pay, while unjustified dismissal (despido improcedente) involves severance pay of at least 45 days of wages per year worked, plus the firm must pay retroactively full wages to the employee for the duration of the appeals process (up to 60 working days). The social tribunals have been notoriously favorable to dismissed workers. In 1993, of cases decided by the courts, 72.8 percent resulted in decision in favor of the worker or partially in favor of the worker. This disproportionate number of appeals resolved in favor of the workers has meant that workers have strong incentives to appeal unless they receive severance pay close to 45 days wages per year; as a consequence, the effective minimum for severance pay is closer to the 45 days associated with unjustified firings. 1/ While data are not available on the precise level of severance pay awards as a function of days pay per year of seniority, the global figures confirm the fact that payments are very high (Table 2). Chart 3 shows the evolution of severance pay in recent years. The calculation of average severance pay divided by the average wage in the economy shows that this compensation reached a peak of well over one year’s wages in 1993. 2/

Table 2.

Spain: Dissimal Costs

article image
Source: FOGASA is the wage Guarates Fund.
Chart II-3
Chart II-3

SPAIN: Average Severance Pay, 1980–94 Individual Dismissals with Agreed Settlements

Citation: IMF Staff Country Reports 1995, 039; 10.5089/9781451811988.002.A002

1/ Calculated by dividing average severance pay by average total wages of salaried workers in the economy. Assumes 261 days per year.
(5) Unemployment benefits

The unemployment benefits system, virtually nonexistent under the Franco regime, grew during the 1980s into a welfare state system comparable to that in other European industrial countries. Both the coverage of the unemployment system and the generosity of the benefits expanded significantly. Chart 4 illustrates the rise in the proportion of registered unemployed covered by benefits from a low of under 20 percent in 1983–84 to 70 percent in early 1993. When agricultural workers are excluded from the calculations, the coverage rate in 1993 reached nearly 85 percent. 3/ This increase in the coverage rate occurred primarily because of government initiatives to improve coverage, but it was also affected by the changes in the structure of employment. Government decrees extended coverage of unemployment benefits to temporary workers (1984), employees of cooperatives (1985), and to workers in several sectors previously covered by special regimes (1986, 1987). The decrease in the proportion of self-employed workers in the economy, a fall in the proportion of unemployed who have not previously worked, and a drop in the proportion of long-term unemployed 1/ also contributed to the improved coverage rate in the late 1980s and early 1990s.

Chart II-4
Chart II-4

SPAIN: Unemployment Benefits

Citation: IMF Staff Country Reports 1995, 039; 10.5089/9781451811988.002.A002

The level of unemployment benefits shows a similar positive evolution during the second half of the 1980s. This is also illustrated in Chart 4, which shows an increase in average benefits of roughly 40 percent in real terms between 1986 and 1993. As with the coverage rate, the increase in average benefits was a function both of the unemployment compensation policy of the government and of the structure of unemployment. 2/ The generosity of unemployment benefits is also reflected in the replacement ratio, which until 1993 was among the highest in Europe (Table 3). 3/

Table 3.

Spain: Replacement Ratios for Unemployment Benefits

(In percent)

article image
Source: Ministry of Labor.

3. The labor market reforms

As the unemployment rate began to climb again in 1991 and 1992, the need for a major overhaul of the labor market regulations became increasingly clear. Beginning in 1993, the government responded with a series of reforms. The 1993 budget (approved in December 1992) included provisions to tighten eligibility for unemployment compensation and to cut benefits. In December 1993, a package of reforms was proposed covering a number of aspects of the labor market, including hiring restrictions, workplace regulations, collective bargaining and dismissal costs. A number of these reforms were implemented immediately by decree; others became effective when the corresponding legislation was passed in May 1994. The 1994 budget took additional measures regarding unemployment benefits. This section presents a description of these reforms.

a. Hiring and contracts

The labor market reforms introduced several modifications in the process by which workers are hired and the contractual forms under which they may be employed. On the hiring side, the statutory monopoly over job placement held by the National Employment Institute (INEM) was officially abolished and the establishment of non-profit private employment agencies was permitted. Temporary employment agencies were also authorized for the first time.

Substantial modifications were made in the regime governing temporary contracts. The reforms ended the virtually unlimited access of firms to temporary contracts by imposing a principal of “causality” on standard temporary employment whereby temporary contracts would be reserved for jobs that were truly temporary in nature. Despite these restrictions, several avenues were left open for other temporary employment. First, current temporary contracts due to expire in 1994 may be renewed for an additional 18 months beyond the previous 3 year limit. Second, the unrestricted use of temporary contracts was retained for certain groups of workers: the long-term unemployed, workers over age 45, and the handicapped. The government has the authority to define annually (in the budget law) the extension of temporary contracts to such special, disadvantaged groups. Third, the reforms created a new apprenticeship contract for workers between 16 and 25 years of age. Firms may use such contracts to hire up to roughly ten percent of their payroll (less for large firms) at below-minimum wages and with reduced social insurance payments. In return, the firms provide training during at least 15 percent of working time.

The labor market reforms also included measures to encourage part-time employment, which is much less common in Spain than in most other industrial countries. 1/ Temporary part-time work is permitted under the same conditions as temporary full-time work. The limits on hours of work permitted in a part-time contract have been lifted, and the social contributions (and corresponding social protections) for workers employed less than 12 hours per week are now minimal.

b. Working conditions

The labor market reforms included provisions to increase functional mobility by eliminating the archaic Ordananzas Laborales and to replace them with negotiated agreements between unions and employers concerning the definition of occupational categories and the flexibility to work between categories. The reform law specified that the Ordananzas would be repealed at the end of 1994 and called for their replacement by occupational definitions and workplace conditions negotiated in collective bargaining agreements. If such collective agreements were not reached by the deadline, the Ministry of Labor had the option to either extend the relevant Ordananzas for one additional year, or to revoke them preemptively. In October 1994, the employers confeder al negotiations on occupational categories. Of 119 Ordananzas still in effect in 1994, negotiated agreements were reached to eliminate 44, and to partially eliminate 7 others. On December 29, 1994, the Ministry of Labor announced that 61 Ordananzas would be retained until the end of 1995; 9 others would be partially retained (including the 7 negotiated). Thus, a total of 49 Ordananzas were eliminated; 44 by agreement and 5 preemptively by the Ministry. 1/

In other aspects of labor market conditions, the reforms included some relaxation of restrictions on length of the working day and overtime hours. Employers may now deploy workers more flexibly, as long as the total hours do not exceed certain limits, and providing for minimum daily and weekly rest periods. The requirement for administrative approval for geographical mobility was eliminated. Government authorities do retain, however, the power to delay transfers by up to six months. 2/

c. Collective bargaining

The 1994 reform law (Lay 11/94) explicitly lays out two motivations for the labor market reforms:

  • “… the reform of the [Workers’] Statute moves in two directions: to facilitate the development of collective bargaining as the regulating force in labor relations, and introduce adaptive mechanisms which are equitably shared among the different aspects of the employment relationship.” 3/

This emphasis on increasing the importance of collective bargaining is seen in several facets of the reform. First, many of the elements of working conditions previously covered by the Ordananzas Laborales are now to be governed by collective bargaining agreements. Second, the reforms reemphasize the importance of collective negotiation in redundancies. Third, provisions are made to improve the representativeness of employers’ negotiators in the sectoral level bargaining, by requiring that firms representing more than half of the total employment in the sector be included in the process. Fourth, the reforms contain provisions to increase the decentralization of collective bargaining, by the inclusion of clausulas de descuelgue (“detachment clauses”) which permit firms to opt out of the sectoral level collective bargaining agreements. 1/ Finally, collective agreements may now explicitly specify which provisions are to have effectiveness beyond the life of the agreement. Previously, most important clauses of contract (including wage provisions, working conditions, and work schedules) were considered “normative” which meant that they were perpetuated beyond the life of the contract unless supplanted by a new agreement.

d. Dismissal costs

The recent labor market reforms did not directly modify the severance pay requirements that constitute the core of dismissal costs in the Spanish economy. Nevertheless, steps were taken that could make the dismissal process more agile and less bureaucratic, thus reducing the time costs of redundancies for firms. Three changes in this respect are particularly important. First, the range of causes for dismissal deemed “justified” has been expanded, with the intention of both reducing the number of cases of individual dismissals subject to lengthy (and costly) appeals, and of easing the conditions for justifying mass layoffs. In addition to “economic” and “technical” reasons for dismissal, firms are now permitted to present “organizational” and “production-related” causes for redundancies. 2/ Second, although that requirement of prior administrative approval for mass redundancies was not lifted, firms are now permitted to use individual dismissal procedures for redundancies affecting less than 10 percent of the workforce (up to 30 workers maximum). Third, the consultation period required for mass dismissals has been reduced. Whereas previously, firms were required to engage in a 30 day consultation period with the trade unions followed by a 30 day approval period for the government authorities, now firms may request the approval at the outset of negotiations and a decision must be issued within 15 days of the conclusion of the consultation period. Thus, the time required has been reduced from 60 days to 45 days. Furthermore, if the authorities fail to rule within the allotted time, the firm’s proposal is considered approved. 3/

e. Unemployment benefits

The system of unemployment compensation was one of the first areas to be tackled in the recent labor market reforms. Reforms were initiated in late 1992 and were further extended in the budget law of the 1993. Steps were taken to tighten eligibility requirements and to reduce both the level and the average duration of benefits.

With respect to eligibility requirements, the period of work required in order to become eligible for benefits was raised from 6 months in the previous 4 years, to 12 months in the previous 6 years. Furthermore, the ratio between the time worked and the duration of benefits, which previously was 2:1, was raised to 3:1, reducing the duration of eligibility. Whereas before the reforms, 6 months of work entitled a person to 3 months of benefits, one year implied 6 months of benefits, etc., the new regulations imply that a person who has worked less than one year receives no benefits, one with one year receives only 4 months of benefits, 2 years of work translates into 8 months of benefits, and so on. This change not only reduced the generosity of benefits, but also helped to stop the deliberate “rotation” of workers into unemployment. 1/

The level of unemployment benefits was reduced in three main ways. First, the level of benefits (as a share of previous salary) was reduced as shown in Table 3. Second, the minimum benefit level for an unemployed person without children was reduced from 100 percent of the minimum wage to 75 percent. Third, benefits were made partially subject to taxation and social security contributions. Unemployed people now make social insurance contributions at 65 percent of the rate of an employed person, and benefits have become subject to income tax withholding. Those three steps have significantly reduced the high replacement ratio which afflicted the unemployment compensation system before the reforms. As demonstrated in Table 3, the replacement ratio in the first months of unemployment has fallen from 85 percent to 77 percent for a single person at the average wage and from 94 percent to 82 percent for a typical person with a family.

4. Impact of the reforms on the labor market

The labor market reforms in Spain are much too recent to permit a full assessment on their impact on employment and wages. The principal reforms in unemployment compensation have been in place for only two years, while those governing other aspects of the labor market have been in effect for one year or less. Nevertheless, there are some monthly and quarterly data which could shed some light on likely effectiveness of some of the measures.

In addition, analysis of the remaining labor market restrictions can be used to draw conclusions concerning the adequacy of the reforms.

a. Overall impacts

The early global indicators on the evolution of wages and employment during 1994 have generally been positive. Average wage settlements have shown significant moderation, falling from 7.3 percent in 1993, to only 3.5 percent for the first 11 months of 1994. Despite a corresponding drop in the inflation (from 5.0 percent in 1993 to 4.3 percent in 1994), this translates into a decline in real wages for the first time in a decade. Employment creation has also been relatively strong. Employment growth turned positive in the second and third quarters of 1994 as the country emerged from the recession. The increase in employment, while modest, has occurred at an earlier stage in the economic recovery than in previous recessions. While the labor market reforms may well have played a role in these developments, caution should be exercised in attributing them primarily to the reforms. 1/

Despite the encouraging developments, there are some worrisome signs. As regards wages settlements, there are early indications that wage increases may rebound in 1995. The trade unions have announced a stronger emphasis on wage increases than in 1994, when the union focus was on negotiating concessions in working conditions in the context of replacing the Ordenanzas Laborales. The 1994 average wage increase was also held down by the wage freeze in the public sector, which will be raised in 1995. Furthermore, it should be noted that the improved 1994 figures cover a smaller number of wage settlements than in previous years, since many wage negotiations have been delayed by the reforms. As these (more intractable) outstanding contracts are settled, it is possible that the average increase for the year will rise. (See below for a discussion of problems in employment).

b. Hiring

The new hiring provisions of the reforms have already had a significant impact on the structure of labor contracts. Apprenticeship and part-time contracts have jumped substantially. Temporary contracts were signed in the first half of 1994 at a pace 46 percent above the rate in 1993, with over 400,000 workers hired. 1/ The new apprenticeship contracts were also widely used. Firms hired 114,000 apprentices in the first six months of 1994.

The negative side of the hiring situation is that the widespread use of new contracts may have given a once-off boost in the employment levels which will not be sustained as firms fully adjust the structure of their payrolls to the new legal framework. Between the end of 1993 and the third quarter of 1994, total employment climbed by 63.8 thousand workers. But this figure reflects transitory increases in temporary and apprenticeship contracts that will not be sustained into 1995. It is likely that the rate of 114,000 apprentices hired in the first six months of 1994 will moderate as firms will reached the legal limit on the number of apprentices per firm.

The hiring of temporary workers also dropped off sharply in the second half of 1994. It appears that firms “stocked up” on temporary workers in the months before the new regime restricted temporary contracts. Before the new restrictions took effect, firms were hiring temporary workers at the rate of 45,000 per month (roughly the same pace as in 1993), but after the new law was implemented (on May 24, 1994), the pace of temporary hires dropped to only 7,000 per month. This drop in temporary hires may begin to produce a decline in the rate of growth of employment in 1995. Indeed, it should be noted that the increase in apprentices and part-time workers has masked the fact permanent employment in Spain has continued to fall during 1994. The number of permanent workers fell by 120,000 (2.1 percent) between the end of 1993 and Q3 of 1994.

c. Working condition and collective bargaining

In the areas of working conditions and collective bargaining, the effects of the reforms are not easily measured with short-run indicators, since the changes will be felt only gradually. What can be said is that the uncertainty surrounding the reforms delayed substantially the conclusion of collective bargaining agreements for 1994; many have spilled over into 1995. The data through November 1994 indicate that there has been no significant change in the proportion of workers covered by firm level collective bargaining agreements, despite the reform provisions designed to facilitate the decentralization of bargaining. 2/

Independent of the trends in the preliminary data, there are reasons for concern that the changes implemented may be insufficient to have a significant effect on rigidities in the utilization of labor. By delegating the reform of many aspects of the employer-employee relationship to the collective bargaining process without in most cases significantly altering the underlying legislation, the unions have been given little incentive to accept increased flexibility unless in return they are granted concessions in other areas (e.g., higher wages). Expressed in game theoretic terms, by increasing the scope for bargaining without altering the default payoffs (Which are perceived as favoring the trade unions), the government has left the unions in a strong bargaining position from which to resist any changes in the outcome.

d. Dismissal costs

There has been some downward movement in severance pay costs according to data available for the first months of 1994 (Table 2). The average mediated settlement dropped from 2.37 million pesetas to 2.18 million pesetas, while the overall average dropped from 1.76 million pesetas to 1.63 million. However, these levels are still well above pre-1992 figures, and it would premature to attribute the decline to the labor market reforms.

There are early indications that the reforms facilitating individual dismissals instead of collective dismissals may be having an effect, although it may not be an entirely positive one. While during the first six months of 1994 the share of mass dismissals fell marginally (from 30.8 percent to 30.2 percent), the share of dismissals decided by social courts jumped (from 9.6 percent to 11.2 percent). Businessmen have expressed concerns that the new procedures may contribute to the “judicialization” of the firing process, as workers previously covered by mass layoff procedures are now dismissed individually and take their cases to the social tribunals. It would be premature to draw a conclusion at this time, but the early data suggest that these fears may have some justification. 1/

e. Unemployment benefits

In the area of unemployment benefits, there is clear evidence that the reforms have had a significant impact. As discussed above, replacement ratios have dropped significantly (Table 3). The percentage of the unemployed covered by benefits has also fallen. From a peak of 69.2 percent in January of 1993, the coverage rate had fallen to 53.8 percent by August of 1994. The absolute number of unemployed covered has dropped as well, despite the rise in average employment. While part of this drop is due to an increase in first-time job seekers and in the number of long-term unemployed, there has also been a clear effect of the tightening of eligibility requirements.

A further manifestation of the reform in benefits is found in the preliminary government expenditure numbers for 1994, which indicate that expenditure on unemployment benefits fell by 4.8 percent from 1993 levels (from 3.5 percent to 3.2 percent of GDP) despite the increase in unemployment. In fact, the true drop was probably even greater that the official figures report, since unemployment expenditures for 1994 included back payments for overspending in 1993.

It will take time to see if this reduction in the generosity of unemployment translates into a significant reduction in so-called “voluntary” unemployment and increases the speed at which the unemployed move back into jobs. Replacement ratios are still quite high for low and medium income workers during the first two years of unemployment, and it should be recalled that unemployment compensation is paid in addition to severance pay. Thus, a permanent worker earning average wages who loses his job will initially receive at least 77 percent of his previous salary plus an average of nearly one year in wages as severance pay! 1/

5. Conclusions

The government has taken major steps in reforming the labor market in 1993 and 1994. The measures undertaken cover virtually every aspect of the labor market, from hiring procedures and contracts, to working conditions, to dismissals. The unemployment benefit system was particularly salient in moving to tighten eligibility criteria and lower the generosity of benefits while retaining fundamental social protections. Preliminary data from the first months of the reforms show some positive signs. At the aggregate level, employment is climbing again after the sharp fall in the recent recession, and wage settlements have moderated significantly in 1994. There are also positive signs in some specific labor market areas—part-time work is rising, severance pay and expenditures on unemployment benefits are falling, and some progress has been made in increasing functional and geographical flexibility via the elimination of some Ordenanzas Laborales and relaxation of restrictions on transfers.

Despite these encouraging steps, this paper has pointed out that there are also worrying signs in the recent data as well as nagging concerns that the reforms have not been far-reaching enough to reduce substantially the unemployment rate over the medium term. The number of permanent workers in the economy is still declining, and the increase in part-time and temporary workers will likely fall off significantly in 1995, as new “equilibrium” levels are reached. Flexibility in the use of labor in the work place is still tightly circumscribed and more than half of the antiquated Ordenanzas Laboralas slated for elimination by the end of 1994 have been extended into 1995. Dismissal costs, while moderating, are still at historically high levels.

Particular concerns about the medium-term prospects for reducing unemployment center on three aspects of the labor market where reforms have thus far been either timid or counterproductive. First, the modification of the different contracts under which workers are employed has resulted in some minor improvements in the flexibility of permanent employment, but there has been a significant loss of flexibility due to the restrictions placed on temporary employment. Taken together, these changes appear to have produced a net loss in flexibility to the market as a whole. Second, little has been done to reduce dismissal costs. These costs constitute the main barrier to employment flexibility and also perpetuate high wage settlements by exacerbating insider-outsider bargaining effects. Finally, the renewed emphasis on collective bargaining in the government’s reforms, while laudable in terms of encouraging worker participation and emphasizing consultation rather than confrontation in industrial relations, will not result in moderation in wages and improvement in the level of employment unless bargaining is more fully decentralized and unless the “default” payoffs (i.e., what happens if an agreement is not concluded) are less skewed in favor of the trade unions.

It appears inevitable that further reforms must be undertaken if Spain is to reduce unemployment from 23 percent at present to below 10 percent over the medium term. The recent reforms were an important and difficult first step, but they will not be sufficient to solve Spain’s long-term unemployment problem.

ANNEX: The Impact of Dismissal Costs on Employment: A Review of the Literature

One of the most important rigidities of the labor market in Spain is the high cost of dismissals. Legally mandated minimum severance pay is relatively generous (20 days per year of seniority for permanent workers), but actual payments are much higher. In addition to these high monetary costs, employers contemplating redundancies face either a time-consuming process of governmental administrative authorization (which includes an obligatory period of negotiation with the workers), or a procedure for individual dismissals in which costly appeals to social tribunals are the norm. Although these administrative procedures have ostensibly been simplified as part of the 1993–94 labor market reforms, it is not clear that in practice the overall costs of dismissal have been significantly reduced.

Businessmen and economists who study the Spanish labor market are virtually unanimous in the opinion that these high dismissal costs act as a major deterrent to the creation of permanent jobs, and thus have been contributing factor to the high Spanish unemployment rate. It is somewhat surprising, therefore, that this widespread “common sense” opinion that high dismissal costs contribute to high unemployment is not firmly grounded in empirical studies, nor is it supported by the traditional theoretical literature on the impact of firing costs on the labor market.

This apparent conflict between the labor market literature on dismissals and the consensus opinion of experts on Spain is explained in two ways. First, there has been a confusion in the traditional literature between the direct effects of dismissal costs on hiring and firing patterns over the economic cycle (where the impact on employment is not generally large), and the powerful indirect effects on employment through the upward pressure dismissal costs have on raising insiders’ wages in an insider-outsider bargaining framework. The second reason for the apparent conflict arises in the relative absence of strong empirical work on the actual relationship between firing costs and employment.

a. The theoretical literature

Beginning with the work of Nickell (1978), a number of economists have developed theoretical models of the impact of firing costs on employment and unemployment. Nickell’s original paper presented a model of hiring and firing decisions of firms during different phases of an economic cycle. He argued that the presence of increased firing costs will indeed discourage firms contracting labor, reducing employment; however, there will be a corresponding higher level of employment during cyclical downturns. The overall impact on employment is ambiguous.

Nickell’s conclusions are echoed in a number of other theoretical studies. Gavin (1986) uses a model incorporating additional features, such as voluntary quits and explicit discounting of the future by firms. Notwithstanding the additional features of the Gavin model, his conclusion is the same—firing costs do not have an unambiguously negative effect on employment over the full business cycle. Using a dynamic control theory approach, a recent paper by Bentolila and Bertola (1990) concludes that with realistic parameters, firing costs do not have large effects on hiring conditions, nor do high firing costs reduce the average level of employment….” 1/ A more extreme case is that of Booth (1993), who uses a Nash bargaining model to argue that redundancy payments have no effect on employment levels and actually increase the utility of both workers and employers, vis-a-vis a bargain without redundancy payments.

On the other hand, the theoretical work of Lindbeck and Snower (1984) on insider-outsider effects on unemployment contrast sharply with those studies finding no overall affect of firing costs on employment. Lindbeck and Snower argue that the existence of firing costs can be the driving force generating insider-outsider effects, and furthermore, that insiders have clear incentives to maximize those costs. This effect exists even where unionization is absent (though of course unionism exacerbates the effect). They conclude that “the greater the firing-hiring costs and the greater the insider-outsider productivity differential, the greater the level of unemployment” (p. 15).

Clearly, most studies on dismissal costs fall to incorporate fully these insider-outsider effects of dismissal costs on wages, and thus ignore the primary mechanism by which these costs affect hiring. 2/ High firing costs affect employment decisions not primarily because the costs themselves as so high that they discourage hiring, but rather because they permit insiders to raise real wages while ignoring the potential competition of the unemployed who would be willing to work for lower wages. As a recent CEPR study on unemployment (1994) in Spain concludes:

  • By increasing the protection of employed workers, [firing costs] reduce the risk that these workers will find themselves unemployed. Thus, it decreases the effect of labor market conditions on wage determination. And this, in turn, leads to the persistence of unemployment.

b. Empirical evidence

In the empirical side, the evidence on the impact of dismissal costs is also somewhat confused. There have been few formal econometric studies in this area, and those that do exist vary widely in their conclusions. In a 1982 paper, Nickell finds, paradoxically, that for the U.K. an increase in unfair dismissal cases (a proxy for severance costs) actually lowers the equilibrium unemployment rate significantly. Studies of the impact of German labor market reforms designed to lower firing costs through the introduction of fixed-term contracts produce ambiguous conclusions regarding their effects on employment. 1/

For Spain, the empirical question of the impact of dismissal costs of employment has yet to be adequately studied. Studies of temporary contracts in Spain seem to show significantly positive effects on employment creation. Inferences can be made from the work on temporary contracts (Bentolila and St.-Paul, 1992; Jimeno and Toharia, 1993; Bentolila and Dolado, 1994) which suggest that because temporary contracts have had a large impact on employment and one of the main differences between permanent and temporary contract lies precisely in the level of dismissal costs, therefore dismissal costs must be important. This interpretation would lead to the conclusion that relatively large gains in employment could be achieved from reduction in firing costs. For example, Bentolila and Saint-Paul (1992) provide evidence that employment was increased substantially as the result of the expansion of temporary contracts (with low dismissal costs) in 1984. 2/ Unfortunately, the issue of dismissal costs has been addressed only indirectly in the empirical literature on Spain, so more direct evidence does not exist.

While the exact magnitude of the effect remains to be studied, it is likely that dismissal costs play a large role in perpetuating the dualism of the labor market, maintaining real wages even in the face of high unemployment rates, and in the persistence of unemployment over time. As this brief annex has shown, there is a clear need for additional research (both empirical and theoretical) which would explore more directly the link between dismissal costs and unemployment in the Spanish economy.

Bibliography

  • Bentolila, Samuel and Giuseppe Bertola, 1990. “Firing Costs and Labour Demand: How Bad is Eurosclerosis?Review of Economic Studies, 57: 381402.

    • Search Google Scholar
    • Export Citation
  • Bentolila, Samuel and Juan J. Dolado, 1994. “Labour Flexibility and Wages: Lessons from Spain”, Economic Policy 18:5499.

  • Bentolila, Samuel and Gilles Saint-Paul, 1992. “The Macroeconomic Impact of Flexible Labor Contracts, with an Application to Spain”, European Economic Review 36:10131053.

    • Search Google Scholar
    • Export Citation
  • Boletin Oficial del Estado, 1994. Número 311, (Madrid).

  • Booth, Alison L., 1993. “Layoffs with Payoffs: A Bargaining Model of Union Wage and Severance Pay Determination”, CEPR Discussion Papers, No. 843, November.

    • Search Google Scholar
    • Export Citation
  • Coricelli, Fabrizio, 1990. “Industrial Relations and Macroeconomic Performance: An Application to Spain”, IMF Working Papers, 90/93.

    • Search Google Scholar
    • Export Citation
  • Gavin, Michael K., 1986. “Labor Market Rigidities and Unemployment: the Case of Severance Costs”, International Finance Discussion Papers, Board of Governors of the Federal Reserve System, June.

    • Search Google Scholar
    • Export Citation
  • Hunt, Jennifer, 1994. “Firing Costs, Employment Fluctuations, and Average Employment: An Examination of Germany”, NBER Working Papers No. 4825, August.

    • Search Google Scholar
    • Export Citation
  • inforMISEP (European Commission), 1994. Employment Observatory Policies, various issues.

  • Jimeno, Juan F. and Luis Toharia, 1994. “The Management of Redundancies in Spain: Economic Report”, Fundación de Estudios de Economia Aplicada (FEDEA), Documentos de Trabajo, 9404, May.

    • Search Google Scholar
    • Export Citation
  • Jimeno, Juan F. and Luis Toharia, 1993. “The Effects of Fixed-Term Employment on Wages: Theory and Evidence from Spain, Investigationes Económicas, 27(3):.

    • Search Google Scholar
    • Export Citation
  • Lindbeck, Assar and Dennis Snower, 1984. “Involuntary Unemployment as an Insider-Outsider Dilemma”, University of Stockholm, Institute for International Economic Studies Seminar Papers, No. 282, July.

    • Search Google Scholar
    • Export Citation
  • Luna, Cristina and Coral Garcia, 1994. “Contratos de Trabajo Ordinarios y de Promoción de Empleo”, ES/1994/3 (Banco do España), February.

    • Search Google Scholar
    • Export Citation
  • INEM, 1994. El Contrato de Trabajo: Modalidades de Contratación, Madrid: Ministerio de Trabajo y Seguridad Social.

  • Ministerio de Trabajo y Seguridad Social, 1994. Guia Laboral-1994. Madrid.

  • Nickell, Stephen J., 1978. “Fixed Costs, Employment and Labour Demand over the Cycle”, Ecoomica 45:329345.

  • OECD, 1993. Employment Outlook, (Paris: OECD).

  • OECD, 1991. “Unemployment Benefits and Labor Market Policy”, Chapter 7 in Employment Outlook, (Paris: OECD).

  • Ramaswamy, Ramana and Robert Rowthorn, 1993. “Centralized Bargaining, Efficiency Wages, and Flexibility”, IMF Working Papers, 93/25.

    • Search Google Scholar
    • Export Citation
  • Torres Gallego, Emilio, 1994. Reforms Laboral: Textos y Comentarios, Madrid: Ciencias de la Dirección.

1/

Prepared by J. Franks.

2/

See WP/94/102 for a detailed discussion of the structural, demographic, and cyclical factors contributing to high unemployment.

1/

See Luna and Garcia (1994), INEM (1994) for a detailed treatment of the structure of employment contracts in Spain.

1/

Jimeno and Toharia (1993) suggest that temporary workers earn 11 percent less than permanent workers in similar circumstances.

2/

In 1992, the minimum contract length was extended to 12 months, and in 1993 the maximum was temporarily extended to four years.

3/

Temporary workers receive 12 days severance pay per year worked, whereas permanent workers receive a minimum of 20 days and in most cases much more (see below).

2/

In the case of collective changes, the individual worker could still resign with full severance pay even if the unions had accepted the changes. In the case on an individual change, the worker could mount a challenge in the social tribunals, where the firm would be required to demonstrate “technical, administrative, or productive” justifications for the modification.

2/

See Jimeno and Toharia (1994) for a detailed description of firing costs.

1/

The labor market experts at the employers confederation claim that the effective minimum payment in order to avoid costly litigation is above 45 days per year of seniority.

2/

The calculation is based on 261 potential paid “working” days per year (365 days minus weekends), since it is assumed that vacations and holidays are paid.

3/

Agricultural workers are often employed only seasonally and are covered by a special employment regime; hence, their exclusion is common in measuring the coverage of regular unemployment benefits.

1/

Self-employed workers who lose their jobs are not generally covered by contributive benefits, nor are those in search of their first job. Some of the long-term unemployed have exhausted their benefits. The proportion of unemployed who are looking for their first job fell from 35.3 percent in 1988 to 19.9 percent in 1993. The share of those unemployed for more than two years declined from 43.8 percent in 1988 to 28.7 percent in 1993.

2/

Contributive unemployment benefits depend upon the time the person worked before losing bar job, and upon the previous wage. Thus, average benefits are higher if higher wage workers lose their jobs, and if the proportion of people receiving contributive benefits (as opposed to non-contributive unemployment support) rises.

1/

In 1992, the share of employment in part-time work in Spain was 5.9 percent, which was the second lowest figure among 13 reporting OECD countries (next to Italy). By contrast, the rate in France was 12.7 percent; the UK was 23.2 percent, and the US 17.5 percent. See OECD (1993).

3/

Ley 11/94 as cited in Torres Gallego (1994).

1/

Sectoral level collective agreements are required to contain clauses specifying the conditions under which individual firms may opt out of the higher level agreement. If no such conditions are specified, then a firm may opt out of the sectoral contract if both the employer and workers’ representatives agree. The law specifically states that the clausulas de decuelgue are designed for firms whose economic stability would be endangered by the application of the sectoral wage increase. These requirements (that trade unions consent and that a threat to the stability of the firm exist) suggest that the scope for decentralized bargaining is still limited.

3/

Previously, if the labor authorities failed to respond in the allotted time, the request was considered rejected.

1/

Under the previous system, there were reports of employers hiring workers on the basis of a cycle of six months of work followed by three months of unemployment, followed by six more months work, etc. This tactic has been largely frustrated by the new regulations.

1/

Both the moderation in wages and the early recovery in employment may well be more the result of the sharper-than-normal decline in employment that occurred during the downturn than of labor market reforms. The rapid drop in employment in 1992 and 1993 affected permanent as well as temporary workers, and thus exposed labor market insiders to downward wage pressure for the first time in years. After such unusually severe downsizing in firms, it would also be expected that employment might recover more quickly as the economy turned around.

1/

This increase in part-time contracts over 1993 was not simply a function of the economic recovery, since the 1994 pace is also well above that of previous (non-recessionary) years.

2/

It should be acknowledged that the bulk of collective bargaining agreements take effect in January of each year. Since the new reforms were not yet fully in place in January 1994, it will not be possible to fully evaluate the degree of centralization in bargaining under the new system until data are available for early 1995.

1/

Two clarifications are in order. On the one hand, the data reported may understate “judicialization”, since they only report case resolved and there is a lag between appeal and resolution. On the other hand, it should be acknowledged that the share of cases taken to social tribunals also varies according to the economic cycle, with fewer appeals occurring in economic downturns. Thus, the increase in appeals in 1994 may be due in part to the economic recovery.

1/

Workers dismissed justifiably will have to wait 3 months after losing their jobs to claim their benefits. Those whose dismissal is ruled unjustified may collect benefits immediately. Recall however, that of those cases where the justification for dismissal is challenged in the social tribunals, nearly 80 percent of the decisions favor the employee.

1/

p. 381 (italics are original). The authors claim that despite this result, their model can help explain the dynamics of European employment in the 1980s. The model suggests that increased uncertainty in demand increases the importance of firing costs to the hiring decision; thus during the late 1970s, the increased economic uncertainty after the oil shock combined with higher dismissal costs to hold unemployment artificially high. An after effect of this situation was the low hiring rates and increasing unemployment of the 1980s. As the authors themselves acknowledge, however, this model still relies upon traditional factor affecting labor demand to drive the outcome.

2/

This is the problem with Gavin (1986); Bentolila and Bertola (1990), and other studies.

1/

See Hunt (1994). Hunt’s study does show an improvement in the speed of employment adjustment in the German labor market, but she was unable to attribute it to lower firing costs. This may be because labor costs did not decline significantly despite the reforms, or because although costs did decline, they do not impact in a major way on the hiring decision.

2/

The conclusion that the lower dismissal costs of temporary contracts produce higher employment is tempered by two features of the result. First, the analysis suggests that the sharp increase in employment after the introduction of temporary contracts was in part a temporary consequence of overshooting of the long-run equilibrium employment level. Second, the authors suggest that the jump in employment was partly due to an increase in cyclical sensitivity of employment, not because of an increase in the average employment level. Thus, the Bentolila and Saint-Paul result could be seen as largely consistent with the conclusion of the theoretical studies of Nickell and others which conclude that firing costs affect the cyclical variations of unemployment but average employment over the cycle.

Spain: Selected Background Issues
Author: International Monetary Fund