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.
Bentolila, Samuel and Giuseppe Bertola, 1990. “Firing Costs and Labour Demand: How Bad is Eurosclerosis?” Review of Economic Studies, 57: 381–402.
Bentolila, Samuel and Gilles Saint-Paul, 1992. “The Macroeconomic Impact of Flexible Labor Contracts, with an Application to Spain”, European Economic Review 36:1013–1053.
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.
Coricelli, Fabrizio, 1990. “Industrial Relations and Macroeconomic Performance: An Application to Spain”, IMF Working Papers, 90/93.
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.
Hunt, Jennifer, 1994. “Firing Costs, Employment Fluctuations, and Average Employment: An Examination of Germany”, NBER Working Papers No. 4825, August.
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, 94–04, May.
Jimeno, Juan F. and Luis Toharia, 1993. “The Effects of Fixed-Term Employment on Wages: Theory and Evidence from Spain, Investigationes Económicas, 27(3):.
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.
Luna, Cristina and Coral Garcia, 1994. “Contratos de Trabajo Ordinarios y de Promoción de Empleo”, ES/1994/3 (Banco do España), February.
Ministerio de Trabajo y Seguridad Social, 1994. Guia Laboral-1994. Madrid.
Ramaswamy, Ramana and Robert Rowthorn, 1993. “Centralized Bargaining, Efficiency Wages, and Flexibility”, IMF Working Papers, 93/25.
Prepared by J. Franks.
See WP/94/102 for a detailed discussion of the structural, demographic, and cyclical factors contributing to high unemployment.
Jimeno and Toharia (1993) suggest that temporary workers earn 11 percent less than permanent workers in similar circumstances.
In 1992, the minimum contract length was extended to 12 months, and in 1993 the maximum was temporarily extended to four years.
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).
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.
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.
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.
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.
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.
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.
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).
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.
Previously, if the labor authorities failed to respond in the allotted time, the request was considered rejected.
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.
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.
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.
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.
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.
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.
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.
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.
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.