ISBN 92-2-110907-0
ISSN 1020-5322
First published 1997

Download the document in Adobe Acrobat's Portable Document Format (PDF)

Table of Contents

Foreword

In this paper the author sets out to investigate whether the principles of labour market policy espoused in general in the member countries of the OECD are useful in the circumstances of Argentina, Chile and Uruguay. The structure of these economies is moving closer to that of OECD members and, in addition, recent changes in "governance" have expanded the role of the social partners. The author finds that the choice of forms of state intervention in the labour market in these three countries had not been independent of the labour market intervention model adopted by OECD members to confront their severe employment problems. This model centred on reconstitution of the work incentive, the relaxation of regulations constraining employers and the decentralization of collective negotiation.

The paper begins with a discussion of the OECD model of state labour market intervention. This is followed by a review of the evolution of economic and labour policies and of labour market trends in Argentina, Chile and Uruguay. State labour market intervention in the three countries in the 1990s is examined, focussing on employment policy. Finally, the scope and impact of those forms of labour market intervention are analysed and contrasted with the experience of advanced countries. The analysis suggests that in Argentina, Chile and Uruguay, even though the OECD model of state labour market intervention did shape their patterns of intervention, emulation has been only partial. It was also biased to the extent that state intervention through employment, training and income maintenance programmes has been weak, while the decentralization of bargaining and, usually, the reduction of employment protection were pursued much more decisively.

The author points out that apart from the contrasts between Argentina, Chile and Uruguay on the one hand, and advanced OECD countries on the other, there is substantial divergence in the application of the model among Argentina, Chile and Uruguay. Differences are, to some extent, the result of disparities in social and political pressure. The model's core (relaxation of labour protection and decentralization of wage bargaining) was instituted in Chile under a military regime, whilst it is being implemented in Argentina under the disciplining impact of large-scale surplus labour, but within the political requirements of democratic functioning. Even though in Uruguay the situation resembles that of Argentina, the impact of unemployment seems to have been less harsh and the political constraints stronger, so that labour reforms are more gradual and less sharp.

One significant point is that while governments in all three countries, and particularly in Argentina and Uruguay, are pressed by the need to attain more adequate employment growth rates and reduce unemployment, in all three state labour market intervention is constrained by fiscal discipline. This naturally presents a major obstacle to the full working of the OECD labour market model.

Gek-Boo Ng
Chief
Employment and Labour Market Policies Branch

State labour market intervention in Argentina, Chile and Uruguay:
Common model, different versions*

1. INTRODUCTION

2. THE LABOUR MARKET INTERVENTION MODEL

This notwithstanding, and somewhat independently of the theoretical and empirical analysis of the determinants of the diversity of unemployment and employment performances, one "standard" model of state labour market intervention developed, mainly during the 1980s. The model was followed in different degrees (only partly or marginally in some countries, more fully in others) in the hope of curbing pervasive unemployment in OECD countries, ⁽¹⁾ and it was to be viewed as a the standard for emulation in several countries of Latin America as well as of other regions. Departing from some still disputable, but only too often asserted, assumptions about how the labour market operates, this paradigmatic state labour market intervention model is oriented by three main objectives: reconstitution of the "work incentive" (allegedly undermined by employment and income security), relaxation of constraints on the managerial prerogative to hire, deploy, relocate and dismiss workers, and decentralization of bargaining structures. Given economic changes and international restrictions, fulfilment of these aims is considered to be a prerequisite for reducing unemployment and strengthening the employment, and ultimately the economic, performance. The relative emphasis on each one of these objectives, and accordingly on the different instruments of intervention, vary across countries, depending inter alia on which had been their prevailing systems of social protection. In addition, as the unemployment problem is to an important extent attributed to individual deficiencies, other supply-side forms of intervention, such as those addressed at improving the individuals' employability prospects and skill levels, have been privileged.

In this context, labour law reforms have been a major concern in OECD countries since the late 1970s, and particularly the 1980s. Labour policies and legislative reforms in OECD countries paid particular attention to the regulation of employment and income protection, to wage determination regimes, and to working time arrangements. The reforms were expected to stimulate employment growth. Following these ideas, inter alia temporary employment was made easier, regulations on dismissal somewhat relaxed, access to unemployment insurance benefits made tighter, and cuts imposed on amount and duration of unemployment compensation. Changes in working time arrangements were proposed to "share" existing employment opportunities among a larger number of workers. ⁽²⁾

At the same time, the OECD placed strong emphasis in state labour market intervention through the active labour market programmes (namely, employment services, training, subsidized private sector employment, direct employment creation in the public sector). As is common knowledge, labour market programmes have been differentiated according to whether they are "passive" or "active", in other words, to whether the government gives just some support (mainly income support to the unemployed), or undertakes some effort at the same time to relocate workers and improve their capabilities. In fact, the "active" labour market programmes characteristic of the 1980s and 1990s evolved from the earlier "active manpower policies" implemented in Western European countries from the early 1960s to face labour shortages, following the model of the Swedish programmes initiated in the 1950s. ⁽³⁾

In the 1960s, with unemployment levels close to 1 per cent in several Western European countries, state labour market intervention sought to ease the prevailing (relative) labour scarcity, considered to be at the root of spiralling "wage-cost inflation". With a view at checking wage growth in excess over productivity increases, the main purpose of labour market policy was to increase the labour force available to work and to facilitate labour reallocation to the expanding sectors and regions, and the occupations with labour shortages, away from high unemployment localized enclaves (the unemployment "pockets") and declining industries (coal mining, shipyards, etc.). Thus, labour market measures and programmes were intended to stimulate labour force participation (among others, through tax incentives to married couples, creation of part-time jobs that would facilitate women's labour market incorporation), and geographic (providing accessible housing and relocation subsidies, for instance) and industrial/occupational (offering retraining) mobility. In some countries, they were coupled with governments' active participation in the recruitment of foreign workers, who were directed to areas and jobs with labour shortages through the tight control over work permits (the so-called "guest worker" programmes). The immigration policy typical of that period aimed at increasing the available workforce and regulating its distribution so as to satisfy labour requirements.

3.2.1 The 1970s and crisis of the early 1980s

Table 3 Unemployment trends (per cent)

Table 4 GDP and unemployment trends

3.2.2 "Adjustment"

3.2.3 Labour market outcomes in the 1990s

Table 5 Employment and output in manufacturing, 1991-95. Annual rates of change (per cent)

Table 6 Youth unemployment and expenditure in active labour market programmes for young workers

Table 7 Labour force participation rates in the 1990s (per cent)

4. LABOUR MARKET INTERVENTION

In Chile, in the context of the preservation of the essential features of the ongoing economic model, democratic conditions, the state of the labour market and political alliances led to some reforms in the benefit of labour after 1990. The starting point was a legal framework that, thus far, had granted only little protection and the minimal collective rights to the workers. Employment protection (contracts and dismissals) was tightened, and collective rights were somewhat improved. These reforms took place alongside with several tripartite agreements, that ensured labour's support to the economic programme in exchange for modest reforms and minimum wage increases.⁽⁶⁶⁾ According to the government's labour policy during 1990-94, the control of compliance with labour laws was to be the sole form of state intervention in intra-firm bipartite relations (González, 1996). On the other hand, several labour market, basically training, programmes were continued and some expanded.

Finally, in Uruguay, with its lax legal system of worker protection, government efforts after 1990 concentrated in changes in wage-setting institutions, so far characterized by the predominance of tripartite industry-wide collective bargaining, pursuing aims similar to those that guided wage policy in Argentina.⁽⁶⁷⁾ New institutions and labour market programmes were also created in this period, to confront unemployment and the acute employment problems of young persons.

Given that the employment protection reforms of this period, together with other measures to reduce labour costs, were officially viewed as incentives to employment generation, ⁽⁶⁸⁾ they are discussed below as an indirect form of labour market intervention (the "indirect incentives"), after reviewing the direct forms, i.e. the active labour market programmes and unemployment compensation schemes.⁽⁶⁹⁾ Last, some changes in wage determination institutions, namely collective bargaining, and in other collective worker rights are analysed.

4.1 Active labour market programmes

In any case, it is obvious that, on the basis of different criteria, many classifications of labour market interventions may be made.⁽⁷²⁾ Labour market intervention may focus primarily in either employers or workers. Once established the policy objectives, measures with uncertain outcomes are different from those with guaranteed short term effects. It is not full proof, for instance, that improved individual employability through training will lead to a job, and even less to a positive effect on the overall unemployment rate; individual success in getting a job might mean replacement of a worker with outdated skills, who in turn might become unemployed or leave the labour force altogether. Similarly, across-the-board payroll tax rebates will not necessarily result in employment creation. By contrast, subsidies and rebates granted in exchange for effective targeted employment creation (subsidizing the wages of workers hired for newly created jobs, for instance) do subtract these workers from the counts of the unemployed in the short term, although whether this could bring about positive results of a more lasting nature is controversial.⁽⁷³⁾

The classification I follow below does not exactly match the OECD standard one, but it seems to be the best suited to organize the labour market programmes in the countries analysed in this study. It takes into account to some extent both criteria, target and nature of the incentive, but some programmes may fall within more than one category (training related to subsidized employment, for instance, may be regarded as being addressed to both worker and employers).⁽⁷⁴⁾ Four main broad categories are differentiated: state placement and counselling offices, state direct temporary employment creation, subsidized private-sector employment (including training or not), and training programmes not linked to employment generation. ⁽⁷⁵⁾

4.1.1 Employment service: Placement and counselling

In the 1990s, the Argentine Employment Service Network, that incorporated Public Employment Offices, Public Placement Agencies and other collaborating institutions, was established (Montoya, 1996). Functions of offices and agencies include: registration of workers in search of jobs and of vacancies; counselling and intermediation; local labour market studies; collaboration in the administration of unemployment insurance; personnel selection for employers; worker relocation; and assistance to the self employed and to micro-firm employers (Montoya, 1996).

Also in Chile the public placement system (the Municipal Placement Offices), created earlier, was reformed. About 25 per cent of the municipalities have a Municipal Placement Office, and there has been an increasing registration of both workers and vacancies, but it seems that this has not been reflected in the evolution of the placement record (Romaguera et al., 1995). Fulfilment of the traditional functions of Municipal Placement Offices (occupational counselling, placement, statistical records) has been impaired by scarce funding and by lack of infrastructure, equipment and skilled personnel. A new role for Municipal Placement Offices was incorporated when they were assigned some participation in the Ministry of Labour's youth training programme (González, 1996).⁽⁷⁹⁾

In Uruguay the National Employment Service established in 1974 never functioned (Melgar, n.d.). But in 1992 the Dirección General de Empleo (DINAE; National Employment Directorate) was created within the Labour Ministry, as well as a body to study, promote and coordinate active employment and training policies, the Junta Nacional de Empleo (JUNAE; National Employment Board), and the Fondo de Reconversión Laboral (FRL; Labour Reconversion Fund). The DINAE is in charge of elaborating employment policy at the national level, advising in relation to and carrying out labour migration programmes, designing and conducting placement programmes for special groups of workers, supervising private labour agencies, preparing and conducting job counselling and labour information programmes, keeping a register of the beneficiaries of retraining programmes, assisting workers who wish to become small employers, etc. The JUNAE should, among other functions, advise the DINAE, administer the FRL, and design retraining and counselling programmes (Melgar, n.d.). By the mid 1990s the DINAE counted 50 employees in Montevideo and had signed several agreements with municipalities to develop regional activities (Melgar, n.d.); to place this figure in the wider context: the DINAE belongs administratively to the Ministry of Labour and Social Security, that in 1994 had a total of 1,430 job positions (INE Uruguay, 1995a). These employment and training boards are, of course, present in the Argentine and Chilean administration structures too.

4.1.2 Direct employment creation

The coverage of Argentina's state temporary employment creation programmes is much less extensive than that of Chile's earlier schemes, i.e. those implemented as from 1975 to confront the dramatic rise in unemployment immediately following the application of the neoliberal reforms.⁽⁸²⁾ Although notoriously insufficient to alleviate massive unemployment in the 1970s and early 1980s, the Chilean programmes had reached a considerable segment of the out of work. The size of this segment declined over time, together with the fall in unemployment. The Argentine schemes of the mid 1990s are far from performing the same role. From their inception in 1975 and to their end in 1988, the employment emergency programmes in Chile, at different points, absorbed between a minimum of 4 per cent and a maximum of 13 per cent of the labour force.⁽⁸³⁾ At their peak, expenditure in emergency employment programmes represented 1.4 per cent of GDP (Scott, 1996). In Argentina, the PIT, at its highest in 1994, provided employment to the equivalent of only just over 0.2 per cent of the urban economically active population (EAP) and 2.2 per cent of the urban unemployed,⁽⁸⁴⁾ and its substitutes, Trabajar and community services, to some 2.5 per cent of the urban unemployed in 1996, ⁽⁸⁵⁾ while expenditure in the latter programmes is equivalent to only 0.03 per cent of GDP.

The Argentinean "work intensive" programmes of state direct employment creation were implemented at a time when overall state employment was - at least in theory - being downsized, both via cuts in the civil service and through the privatization of public enterprises. ⁽⁸⁶⁾Although the reduction of public employment did not respond primarily to labour policy directives but was intended to contribute to reduce the fiscal deficit, the measures taken to limit hirings, stimulate voluntary exit from the public sector, and relocate redundant state employees, among others, could be regarded as labour market programmes and merit a brief digression. Laws and Executive decrees formulated in Argentina as from 1989, with application to the federal administration exclusively, established limitations on recruitment, facilitated employee relocation and introduced incentives to retirement and resignation;⁽⁸⁷⁾ imposed a "freeze" on vacancies; made dismissal somewhat easier; enforced compulsory retirement if legal requirements such as retirement age had been met; prohibited to hold more than one job within the public administration and to pay overtime; terminated temporary contracts; and provided specific training to foster shifts to private activity (Orlansky, 1993 and 1996a). This notwithstanding, actual reductions in state employment seem to have been due exclusively to privatizations. ⁽⁸⁸⁾ State employment was affected by substantial cutbacks also in Uruguay. In 1990, new rules on civil service employment were passed, inter alia encouraging the internal relocation of civil servants, severely restricting new recruitment, and establishing incentives to voluntary resignation (Correa Freitas, 1990); this resulted, as we have seen, in a -20 per cent drop in the number of state employment between 1989 and 1992 (even though state employment was to increase somewhat later).

4.1.3 Subsidized employment

Between 1994 and 1996, the Argentine government produced three programmes to subsidize directly plain employment creation in the private sector (two still ongoing in 1996), each one of a more limited scope than the programmes of direct public employment creation, but some of them as short lived. The Programa de Empleo Privado (Private Employment Programme, PEP) was created in 1994 to replace the state funded PIT above described, on the grounds that PIT's funds had been distributed discretionally by municipal majors and council members.⁽⁸⁹⁾ The state paid part of the wages of participant workers, with and without previous training, who should be registered with the Employment Service Network, and be hired by private firms for at least 4 months. The initial monthly average of 1,108 participants in 1994 had more than doubled in 1995, but the figure still was exceedingly small; by the first half of 1996 the number of participants had gone down to a negligible 49 per month, as a newer programme (Programa de Empleo Privado para PyME), of similar characteristics but confined to small and medium enterprises (with up to 100 employees), created in 1995, was given priority instead. The latter's coverage also is smaller than the earlier PIT's. ⁽⁹⁰⁾ Finally, Forestar subsidizes employment to be created specifically in forestry; the state offers the minimum wage for a maximum of six months, and covers the cost of the liability insurance policy that corresponds to work injuries.⁽⁹¹⁾ In 1996 the two ongoing programmes totalled some 20,000 beneficiaries on average per month (MTSS, 1997), equivalent to less than 0.2 per cent of the urban labour force and 1 per cent of the urban unemployed.⁽⁹²⁾

In addition to these programmes where the state funds the wages of participating workers, the so-called "promoted" temporary employment contracts exempted, partially or totally, from payment of social security contributions, and introduced in the National Employment Law of 1991, were also meant to stimulate employment creation in the private sector, this time by means of the tax rebates.⁽⁹³⁾ Among other restrictions, the law stipulated that they could be used only on the condition that collective lay-offs had not taken place in the last year, and that workers hired through these contracts should be in excess of the pre-existing employment volume. In 1995, special advantages were offered to small firms through a new law. ⁽⁹⁴⁾ Of the four modalities, three are targeted to specific groups.⁽⁹⁵⁾ "Employment promotion" (Fomento del Empleo) is addressed to unemployed workers, including those expelled from the public sector due to administrative rationalization, and exempts employers from paying 50 per cent of payroll taxes; ⁽⁹⁶⁾ Práctica Laboral para Jóvenes offers work practice to young workers (up to 24 years old) who have technical or professional skills; and Trabajo-Formación is targeted to unskilled young workers looking for their first job; both are exempted from payroll taxes except those for the health care scheme and the national employment fund. The fourth, facilitating temporary contracts in the event of the launching of new activities, is not targeted, and the tax rebate is equivalent to 50 per cent of social security contributions. ⁽⁹⁷⁾ Between 1992 and 1996, the number of "promoted" contracts increased almost fivefold, but still amounted to some 5,000, on average, per month.⁽⁹⁸⁾ Non-targeted temporary contracts, used in the event of initiating new activities, account regularly for the major proportion of all contracts under promoted modalities (50 per cent in 1996; 54 per cent of the accumulated total 1992-96), followed by "employment promotion" (43 per cent and 36 per cent, respectively; MTSS, 1997). The contracts that involve some training played a really minor role.⁽⁹⁹⁾

The new labour code in force as from 1995, introduced, together with the part-time contract and the trial period, the apprenticeship contract (minimum 3 months, maximum 2 years) for unemployed workers aged between 14 and 25, that implies substantial non wage labour cost reductions,⁽¹⁰⁰⁾ and a special temporary contract (minimum 6 month, maximum 2 years) applicable to workers older than 40, disabled workers, ex soldiers in the Malvinas/Falklands war, and, notably, women, with a 50 per cent rebate on payroll taxes except those for health care, and not entitled to severance pay at termination.⁽¹⁰¹⁾

Two other instruments created to subsidize private-sector employment involve some form of training. The main programme of subsidized training for young workers is the Proyecto Joven (Youth Project), created in 1993 and co-financed by the Interamerican Development Bank. It is targeted to low-income young workers (older than 16), that lack working experience, who did not complete secondary education, and who are unemployed, underemployed or out of the labour force. The programme is intended to provide skills (at semi-skilled level) that are actually demanded by the private sector; training is followed by a work practice period in a firm; ⁽¹⁰²⁾ state/IDB funds cover all training costs and the scholarships and subsidies granted to participants. ⁽¹⁰³⁾ The other programme is Aprender ("Learn"), that works through the apprenticeship contracts given legal status in 1995; firms are expected to train unemployed workers between 14-25 years old; the government pays their health and work injury insurance if the contract is to exceed 6 months.⁽¹⁰⁴⁾ In April 1996 almost 24,000 persons registered at the Proyecto Joven sixth opening, while in the first semester of the same year the monthly registration with Aprender only averaged approximately 1,600.

In Chile, the programmes of subsidized private-sector employment are linked to training schemes. Two programmes, Chile Joven and an apprenticeship programme (Aprendices), were established in 1991 and 1988, respectively. Chile Joven is comparable to its Argentine counterpart (Proyecto Joven, above described). In fact, probably the Chilean programme helped shaping the Argentinean one. Besides, to alleviate the problems brought about by the decline of coal mining, a re-training programme, with a component of subsidized employment, for these miners was started in 1992; it followed an agreement between the government and the miners' union permitting early retirement (after 25 years service) of underground miners. ⁽¹⁰⁵⁾ Chile Joven, co-financed by the government and the Interamerican Development Bank, coordinated from a state unit, and run through private training institutes, is targeted to young unemployed workers, preferably between 15 and 24 years old, who are out of the schooling system (Romaguera et al., 1995). It has four sub-programmes: training and work experience in firms, work training to become self employed, training for young workers, and a kind of apprenticeship programme added in 1994.⁽¹⁰⁶⁾ As in Argentina, it is meant to provide the skills effectively demanded by the private sector. In five years, over 115,000 persons participated in this programme, the bulk of which, some 70 per cent, under the "training and work experience in firms" sub-programme (Paredes et al., 1996). The apprenticeship programme (Aprendices) is designed to train persons aged 20 or less on-the-job, and does not provide a subsidy but offers a tax rebate to firms that hire these young workers; it is not a "dual" programme as the work practice in a firm is its only component, that is, it does not include a stage of formal training; its scope is almost negligible, totalling some 560 participants in 1994 (Herrera and Ruiz-Tagle, 1996; Mizala and Romaguera, 1996).⁽¹⁰⁷⁾

The DINAE in Uruguay, through the FRL, initially provided orientation and short training courses, as well as an additional benefit, only to unemployed workers that were receiving unemployment compensation, and attempted to find them a job; firms that hired these workers would obtain a 90 day payroll tax exemption on these contracts, but half of this rebate was to be given back to the FRL. In 1994, some 1,400 workers were enrolled in the retraining courses, a mere 2.5 per cent of the potential beneficiaries, but up to 1995 not one firm had applied to the tax incentive (Melgar, n.d.). Later, training supported by the FRL (now funded by worker and employer contributions) was extended to employed workers too ⁽¹⁰⁸⁾ Opción Joven, targeted to young people and co-financed by the state and the Interamerican Development Bank, was modelled along lines similar to the Argentine and Chilean programmes. It is targeted to unemployed workers, preferably women, between 15-24 years old, with no more than secondary education, and its objective is to generate skills appropriate to the actual private sector requirements; its continuation, Projoven, was initiated in 1996. ⁽¹⁰⁹⁾ They include orientation, training, work practice in enterprises (the state subsidizing only the contributions to social security), scholarships for entrepreneurial training, and support to job obtention (Melgar, n.d.; Lasida and Pereira, 1997).

4.1.4 Subsidized training not linked to employment

But in Chile the state's main effort in the area of subsidized training (apart from the general educational system) is focused in training programmes for employed adults, via offering tax rebates to firms that provide training to their workers (the Programa Empresas, created in 1976).⁽¹¹³⁾ It absorbs almost 70 per cent of training expenditure. In 1995 over 430,000 workers completed training under this programme in Chile (Herrera and Ruiz-Tagle, 1996). Argentina also has a scheme (with a much more limited role in the occupational training system than the Chilean programme) based on tax allowances to firms that sponsor training, the Crédito Fiscal, meant to promote investment in training infrastructure; at its inception in 1980 it was applicable exclusively to manufacturing activities; it was extended to all economic activities and given renewed impetus in 1995.⁽¹¹⁴⁾ As we have seen, inclusion of these programmes in the category of labour market policies is dubious. In a way they belong to the broader state sponsored educational systems, that include also the direct provision of professional training in public-sector schools and state subsidies to private schools that provide occupational training.

Also Uruguay has a system of state provided technical education of long standing, including the Consejo de Capacitación de Personal (Council of Personnel Training, COCAP) and the Centro de Capacitación Profesional (Center for Professional Training), the first one to train workers to be employed in the export industries (Melgar, n.d.). Besides, the Instituto Nacional de la Juventud (National Institute for the Youth, INJU) was created in 1991 to foster youth employment. The Centro de Capacitación Profesional is also targeted to the youth, and trains some 400 young persons per year (Melgar, n.d.). The orientation of these institutions is to improve employability: INJU's programmes include a scholarship programme (set in 1992) to help acquiring those work attitudes and behaviour expected to help young workers from low-income households to get jobs; some 340 persons obtained scholarships in 1993-94. The INJU has an information centre for the youth, provides vocational orientation, and advises in relation to small firm projects. In 1994, two two-year experimental programmes, one for training young workers and the other for developing entrepreneurial skills, co-financed by the state and the Interamerican Development Bank, were devised. The first is Opción Joven, discussed above. The second programme, to promote entrepreneurial activity among young persons, is to provide training, technical assistance, funds and monitoring during two years (Melgar, n.d.).

4.2. Unemployment compensation

The three existing schemes to date (1997) were designed exclusively for wage earners, dismissed without a fair reason or as a result of collective lay-offs; in Argentina also workers whose temporary contract has ended are eligible, and in Uruguay, those in temporary lay-off.⁽¹¹⁹⁾ Access is confined to workers registered with the social security scheme, and require a minimum length of worker contributions.⁽¹²⁰⁾ Workers employed in certain economic activities are not eligible for the unemployment insurance benefit: rural and household workers in Argentina and Uruguay; bank and state employees in Uruguay; construction - protected by an older, separate scheme - and public administration employees whose job ended due to rationalizations, in Argentina.⁽¹²¹⁾

Benefits are linked to the latest actual wage in Argentina and Uruguay (but with a top limit in terms of minimum wages), and only to the minimum wage in Chile. In Argentina and Chile, the duration and/or amount of the benefit vary according to the length of contributions to social security. The benefit ceases to be paid in Chile if a job offered by the SENCE, with a wage above 50 per cent of the last job's remuneration, is rejected, and in Uruguay if a "convenient" job offer is rejected without a legitimate cause.⁽¹²²⁾ A cursory glance at table 8 suggests that the unemployed are better-off in terms of amount of benefit in Uruguay than in the other two countries, while the Chilean scheme provides the lowest compensation, although it is the most extensive in terms of potential coverage. Effective coverage is very low in the three countries, only marginally better in Uruguay.

4.3. Indirect incentives

Employment protection reform was one cornerstone of post 1990 Argentine labour policy; the labour costs reductions as well as the greater freedom to terminate employment that it would entail were regarded and publicized by the government as potential incentives to employment generation. Lay-off compensation had been increased in 1989, when the three-minimum-wages-per-year limitation on length of service compensation was eliminated; this change preceded the economic reforms and was against the "spirit" of the later labour law modifications. The Employment Law of 1991 reintroduced a ceiling but higher than the earlier one, as it now was three total collectively agreed earnings, exclusive of seniority benefits. In this context, the "promoted" temporary contracts, examined above, that reduced or eliminated the cost associated with dismissal, and the formalization in 1995 of the three-month trial period (no compensation is required if the employment relationship terminates at the end of this period) were supposed to help expand the number of available jobs. ⁽¹²³⁾ Besides, from 1995, small firms have been allowed to alter dismissal regulations, in any direction, through collective agreements and were favoured by a shorter advance notice period. The generalized rebate of social security payroll taxes (the reduction varying throughout the country according to zone coefficients determined by poverty level and distance to the capital city), ⁽¹²⁴⁾ launched in 1993, was meant also to stimulate the generation of employment, and as we will see the major government effort in terms of labour market expenditure, in this case, forgone revenues, centered in this area.⁽¹²⁵⁾

Before the reforms of the 1990s, labour protection in Chile was extremely weak, and the change towards enhancing protection was associated with democratization. As discussed above, the former erosion of labour rights had been one central component of the military's economic strategy, and it had implied the total transformation of individual and collective labour legislation (Paus, 1994). The civilian government undertook as from 1990 modest legislative reforms in some crucial areas, in the context of decreasing unemployment levels and a positive economic performance. In contrast to Argentina's labour law reforms flexibilizing the employment contract, in Chile the maximum period for fixed-term contracts was reduced from two years to one. Moreover, the possibility of dismissal without a cause ("deshaucio", incorporated in the labour code of 1979) was eliminated, and the earlier clauses that had been intended to prevent labour's collective action at the workplace were erased from the list of fair causes of dismissal. However, workers now may be dismissed due to "modernization" and "economic needs of the firm" as well as if they are considered not suitable to the job's requirements, and in these cases, after one year of employment in the firm, they are entitled to compensation. The cost of lay-off compensation was increased, benefitting in fact only the workers with over five years employment in the firm, as the change approximately doubled the ceiling to compensation (one month wage per year of service) from five months wages to 330 days.⁽¹²⁶⁾ Still, after the labour law reforms in both countries, employment protection in Argentina continues to be stronger than in Chile, and labour costs higher. On the other hand, the retirement scheme in Chile continues to be financed only out of worker contributions.

In Uruguay employment protection had traditionally been permissive; protection against unfair dismissal only establishes a maximum of six month wages to length of service compensation (most issues are regulated via case law), and norms on temporary contracts are minimal (Marshall, 1993).⁽¹²⁷⁾ No substantive changes have been introduced in the 1990s in relation to contracts and dismissals in the private sector.

4.4 Wage setting institutions

Since 1991 the Argentine government's wage policy has been directed to promoting decentralized bargaining and fostering the heterogeneity of wage increases, in line with productivity growth differentials. Formerly these aims had been assigned some role in the anti-inflationary package but, once inflation subdued, they were pursued in connection with other expectations, such as positive employment effects. Decentralization was pushed forward by successive legal measures (1991, 1993, 1995 and 1996). ⁽¹²⁸⁾ In the government's view, decentralization permits a better matching between wage increases and the financial situation of individual firms. Particular emphasis has been placed on the benefits of decentralization for the smaller enterprises (specifically addressed to in some of the legislative measures), expecting that it will lead to greater differentiation in employment and working conditions between the small and the larger firms. On the other hand, the very role of collective bargaining was changed when it was legally permitted to negotiate collectively conditions less favourable to workers than those stipulated in the national labour code, of general application. Until 1995, legislation and case law established that the clauses "most favourable" to the workers were to prevail over all others. The regulations of 1995 on employment and labour relations in small firms made possible collectively-agreed modifications in any direction, the issues encompassing holidays, the 13th wage, employment termination and internal worker mobility. The new law also allowed separate bargaining for small firms; it established that representation of small firm employers in industry-wide negotiation is mandatory, and that collective agreements should include specific regulations for small firms,⁽¹²⁹⁾ unless small firm agreements are negotiated separately, The Executive's decrees of 1996 (application suspended), among other aspects reinforced decentralization, declared agreements no longer valid once the term for which they had been signed expires and, in relation to the small enterprise, the prevalence of small firm over industry-wide agreements.

A similar stand in relation to collective bargaining was adopted in Uruguay; the government considered decentralization to be more suitable to enhance competitiveness in the context of the ongoing process of regional integration. Once again, however, the steps taken to expand decentralized negotiation have been less extreme than in Argentina. Following a few failed attempts in this direction as from 1990, the government finally withdrew from the tripartite wage councils⁽¹³⁰⁾ in 1992 (Rodríguez, 1996; Informe RELASUR, 1995); this withdrawal was intended to encourage bipartist bargaining in the hope of, in this way, fostering decentralization.⁽¹³¹⁾

In Chile, the labour law reform of 1991 permitted multi employer negotiation, but, still in 1993, enterprise agreements represented 99.6 per cent of all agreements (Chile, 1995), and a project that might have facilitated the extension of multi employer bargaining, sent to Parliament by the Executive, was rejected; employers also opposed the extension of coverage and of the issues to be negotiated collectively (González, 1996). ⁽¹³²⁾ It is worth noting here that, whereas in Argentina and Uruguay the terms agreed collectively are legally extensive to all workers in the firm or industry, irrespective of trade union membership (erga omnes clause), in Chile, in spite of the extension of coverage in 1990 to non union members at the expense of paying a fee to the union, still in 1995 collective bargaining covered only as little as 15 per cent of all the wage earners entitled to negotiate collectively in 1995. ⁽¹³³⁾ In this period, as we have seen, also several comprehensive, centralized tripartite agreements were reached, that influenced the government's minimum wage policy, a fact that was to have repercussions on the wage structure.

Other collective worker rights were also modified. Limitations on the right to strike in essential services were imposed in Argentina at the beginning of the 1990s, but no other major change, apart from those affecting collective bargaining, was made (Feldman, 1995). Instead, in Chile, certain trade union rights were strengthened, ⁽¹³⁴⁾ including the lifting of some of the severe restrictions on the right to strike (such as elimination of the 60-day limit to the duration of the strike) that had been introduced during the military regime. In any case, the right to strike in Chile still continued to be, even after the reforms, exceedingly limited: substitution of strikers is permitted explicitly (this is prohibited e.g. in Argentina). Further, labour's position showed few signs of improvement: union density in Chile did not increase between 1990 and 1995 (in relation to all wage earners, it was 19.8 per cent in 1990 and 19.3 per cent in 1995; González, 1996). Finally, in Uruguay, the right to strike, restored in 1985, has not been reformed, although the government that started in 1990 had planned to legislate on the definition of legal and illegal strikes and on other procedural aspects in connection with the right to strike, as well as on lock-outs and protection against employer reprisals to striking workers. These reforms met with opposition and, until 1994, no new labour legislation had been passed (Foreign Labor Trends - Uruguay, 1993 and 1994). The unionization rate in Uruguay has been falling (29 per cent in 1987, 21.4 per cent in 1993), particularly in the private sector, but trade unions continue to play a significant social and political role. ⁽¹³⁵⁾ By contrast, in Argentina, the evidence suggests that trade unions' socio-political influence weakened considerably, even though union density did not decline. The unionization rate in 1994, relative to all employed wage earners, was estimated to have been even higher than in 1991 (about 47 per cent and 44 per cent, respectively). ⁽¹³⁶⁾

5. LABOUR POLICY EFFECTS

5.1 Active labour market programmes

Table 9 Comparative expenditure in labour market programmes GDP shares, circa 1996 (per cent)

Official documents in Argentina often include, within the category of labour market expenditure, estimates of forgone government revenues due to the elimination of or rebates to social security taxes.⁽¹⁴²⁾ This latter appears to be, by large, the most substantial item of expenditure in labour market programmes: income forgone due to social security rebates (the official estimate, for 1996, of the revenues that would have originated in payroll taxes at their previous level) is equivalent to approximately 1.2 per cent of GDP and, if included in the total, would represent 90 per cent of expenditure in all the labour market programmes.⁽¹⁴³⁾ As we know, their impact in relation to the objectives posited is, at best, uncertain (and, in fact, between 1993, the year when the tax rebate started to be implemented, and 1997, at the time of writing, the effects appear to have been nil).⁽¹⁴⁴⁾ This notwithstanding, by mid 1997 yet another project, this time to fully eliminate payroll taxes, is circulating in Parliament.⁽¹⁴⁵⁾

Whilst in Argentina the composition of expenditure in labour market programmes shows, first, a marked preference for indirect incentives and then, if these are not counted, for training (45 per cent of total direct active labour market expenditure envisaged in the 1996 national budget) followed by incentives to private employment (36 per cent), in Chile, once the direct employment creation schemes ceased, the emphasis lies in training programmes, particularly in training for employed adults: 69 per cent of public expenditure in occupational training went to the Programa Empresas, 21 per cent to Chile Joven, 7 per cent to retraining of coal miners and the residual to the apprenticeship and scholarship programmes.⁽¹⁴⁶⁾ By contrast, out of 19 OECD countries, in nine the state does not spend anything on training for employed workers, and in other five expenditure on this group is below 10 per cent of training expenditure. ⁽¹⁴⁷⁾ On the other hand, approximately similar proportions of expenditure in labour market programmes have been assigned in Argentina and Chile to schemes targeted exclusively to young persons (table 6). ⁽¹⁴⁸⁾

Coverage of labour market programmes (against either the targeted population, the labour force, or the unemployed) gives some indication on the impact of state labour market expenditure. If the programme of firm training for employed adults in Chile were to be counted, the labour force proportions reached by active labour market programmes in Argentina and Chile would be comparable, but, if that programme is disregarded, Argentina is much better placed; ⁽¹⁴⁹⁾ and clearly Argentina is better off also if we consider, instead, how substantial is the segment of the unemployed population benefitted by programmes directed specifically at the unemployed or at least susceptible of leading to some impact on unemployment (tables 10a and 10b). ⁽¹⁵⁰⁾ However, coverage of labour market programmes addressed to the youth, relative to the unemployed young population, is similar in both countries.⁽¹⁵¹⁾ Whether the unemployment level or other criterion is chosen for evaluating the success of labour market intervention depends on the policy objectives set in each country, and in Chile purposes other than reducing unemployment appear to have been privileged once unemployment subsided to a more "reasonable", although still quite significant, rate.

As we have seen, the limited scope in Argentina of direct temporary job creation in the public sector contrasts with that of earlier programmes in Chile. Even though the labour force percent reached by these programmes in Argentina is close to the coverage of equivalent schemes in several OECD countries at the end of the 1980s (between 0.1 per cent and 0.5 per

Table 10a Beneficiaries of training and employment programmes in relation to the labour force (per cent)
Beneficiaries: averages per month

In sharp contrast to the advanced OECD countries, few studies on the potential impact of active labour market programmes on the individual workers' likelihood of improving their employment situations have been made in the Latin American countries, and most of them in Chile. In this country, several surveys monitored the occupational trajectory of workers who graduated from the Chile Joven training programme (that combines training and in-firm work

Table 11 Comparative impact of labour market programmes: Selected OECD countries and Chile
Labour force percent of participants in labour market programmes*

In Argentina, perhaps because many programmes were short lived and the newer programmes have not lasted enough as yet, the analyses of the impact of labour market programmes are few, and, as in Chile, tend to be centered in the IDB sponsored project of youth training (Proyecto Joven). In this case, evaluations revealed that the programme was focused in the target group, as 83 per cent of the participants were less than 30 years old, most (91 per cent) had less than complete secondary education, and 49 per cent had been unemployed at the time of enrolling and another 39 per cent underemployed; moreover, most participants came from poor or low-income households (92 per cent).⁽¹⁶⁰⁾ In both Proyecto Joven and Aprender, in-firm work practices took place principally in the service sector (59 per cent and 88 per cent, respectively; over one half of the apprentices were employed in financial and business services, and 24 per cent in social and personal services).⁽¹⁶¹⁾ We know little about whether workers trained under these programmes improved their employment prospects vis-à-vis non participants of similar characteristics; one study monitored the situation of a sample of graduates from Proyecto Joven, finding that sometime later 51 per cent were in a job (comparing favourably with the control group in the case of men but not of women), less than one half of whom in a formal job (Moura Castro, 1997; Mitnik, 1997).⁽¹⁶²⁾ Follow-up studies of how workers employed temporarily through the "promoted" or other modalities fared once the contract had ended are not available, but the fact that the hiring rates of workers with fixed term contracts (including those under "promoted" modalities) generally exceed their separation rates suggests that some might have remained in employment.⁽¹⁶³⁾ This relative lack of investigation is perplexing as, officially, strong emphasis has been placed on the programmes' "effectiveness", following Word Bank guidelines. ⁽¹⁶⁴⁾ For its part, in Uruguay it was reported that, out of 20,000 workers participating in employment and training programmes in 1995 and until mid 1996, 66 per cent later obtained a job. ⁽¹⁶⁵⁾ Specifically in relation to Opción Joven and Projoven, the training programmes targeted to the young, evaluations concluded that they were focused in the target groups, and that about one half of those who completed training had obtained a job later, 15 per cent of whom only for 20 hours a week or less (Lasida and Pereira, 1997).⁽¹⁶⁶⁾

5.2 Unemployment insurance

Table 12 Argentina: Composition of the unemployed population and entitlement to unemployment compensation, 1995.
Buenos Aires (per cent)

The unemployed workers that in 1994 were receiving compensation in Uruguay represented between some 3 per cent (according to the household survey in urban areas) and 16 per cent (an estimate that results from contrasting the number of unemployment compensation recipients registered at the Banco de Previsión Social with the urban unemployed). ⁽¹⁷²⁾ Low coverage partly results from the fact that wage earners employed in agriculture, the public sector, domestic service and banking are not entitled to unemployment insurance compensation. In addition, as Melgar (n.d.) points out, the distinctive economic structure of unemployment insurance beneficiaries, as compared with that of all unemployed workers who had been in a previous job and were eligible for compensation, suggests that non registered employment plays an important role in explaining lack of access to the unemployment insurance scheme in Uruguay: more formalized sectors such as manufacturing and construction are over-represented among unemployment benefit applicants (table 13). In any case, since the 1980s the increase in the number of unemployment compensation recipients has been substantial; according to Melgar (n.d.) this expansion is not correlated with changes in the unemployment rate but with rising social security registration.

Table 13 Uruguay: Economic distribution of unemployed workers and applicants to unemployment compensation, 1994 (per cent)
Activities eligible for unemployment compensation only

5.3 Indirect incentives: Employment protection reforms

In Argentina, initially the use of "promoted" temporary employment contracts was only modest; ⁽¹⁷³⁾ it was estimated to have represented just some 1.5 per cent of all new recruitment in the first half of 1994 (Feldman, 1995). By the first half of 1997, however, concomitantly with the steady fall in open-ended employment, a very significant increase in "flexible" contracts was underway (December 1995 = 100; June 1996: total = 100.4; open-ended = 97.4; fixed term = 144; trial = 168.5; March 1997: total = 103.5; open-ended = 92.6; fixed term = 234.1; trial = 782.8). ⁽¹⁷⁴⁾ Therefore, since the time of the labour law reforms, there has been an ongoing process of replacement of workers with permanent employment by temporary wage earners. Between early 1996 and 1997 permanent contracts went down from 94 per cent to 83 per cent of total employment; the obverse was that the share of workers employed under "trial period" contracts increased almost seven times, and that of workers with fixed term contracts doubled (MTSS, 1997). In manufacturing, this took place in the wider context of continued employment declines in spite of the expansion of output until the crisis of 1995 and after resumption of manufacturing growth in 1996-97 (table 14). Argentina's negative manufacturing employment performance, as we have seen, resulted primarily from the effects of trade liberalization combined with domestic currency appreciation. Economic transformations, and their impact on investment and productivity, naturally were more powerful determinants of employment trends than changes in labour market legal regulation. On the other hand, the effects of the promotion (via temporary contracts entitled to labour cost rebates) of female and youth employment (among other groups) are uncertain. Between 1992 and 1995 unemployment among young workers increased somewhat less than did the average rate (EPH, INDEC, in MTSS, 1996b), and between May 1995 and October 1996, once the apprenticeship contract had been introduced, the unemployment rate of workers aged between 15 and 19 fell more than the average,⁽¹⁷⁵⁾ but this could be attributed to other factors too. Women's unemployment declined more than male's after the creation of low cost contracts for women in 1995,⁽¹⁷⁶⁾ but the behaviour of labour force participation rates and, particularly, the distinctive employment structure of the female labour force could have played a role in explaining this development.

In Chile, during a period (1991-95) of rapid expansion, employment-output elasticities in manufacturing fell; this was due mainly to the fact that in 1994-95 the elasticities became negative; by the mid 1990s manufacturing employment was systematically lagging behind output growth (table 14). Such fall cannot be attributed necessarily (as neoliberals would) to a slowing down in recruitment following the rise in the cost of dismissal and to the somewhat less permissive regulations on temporary contracts.⁽¹⁷⁷⁾ The growth rate of productivity accelerated: in 1990-95 manufacturing productivity increased at an annual rate of 3.5 per cent, above the annual average of 0.8 per cent in the 1980s; the acceleration was more notorious in 1994-95, with a rate of 6.5 per cent per year. With expanding investment in manufacturing, the gradual process of technological change, led by the most export intensive industries, that started by 1989 (Agacino at al., 1995), may have finally provoked a switch to a more labour-saving growth path.

Table 14 Employment performance before and after the labour law reforms:
Medium term employment output elasticities* and employment trends**

5.4 Wage policy: Bargaining structures

One possible approximation to the study of the impacts of changes in the regulation of collective bargaining is to look at trends in wage inequality. And, given the short time span since the regulatory reforms were made in the countries studied here, it seems premature to seek for the expected wider effects on employment and growth performances.⁽¹⁷⁹⁾ In any case, in 1990-1995, labour costs in manufacturing rose faster than productivity in Chile while the opposite was the case in Argentina.⁽¹⁸⁰⁾ Improvement in the labour market situation, and the fact that labour demand in manufacturing was rising until 1994, perhaps in combination with some strengthening of labour's rights, probably is behind the gain in wages in Chile. The reverse occurred in Argentina, that is, manufacturing employment fell and unemployment expanded. This suggests that, at least in this period, labour market conditions played a much more crucial role in the determination of the average wage than bargaining structures, and that one could hardly expect changes in negotiation levels to have any employment effects.

To study the effects of institutions on wage inequality in each country over time, it seems more appropriate to compare the dispersion of wage increments rather than that of wage levels; governments fix increments, and trade unions and employers bargain over rises. In fact, trends in levels and increments do not always coincide, especially in the short period. Growing dispersion of increments does not necessarily lead to a more unequal structure, and, even if wage increases were to be equal for all, this would automatically generate an increasing dispersion of wage levels.

What does the evidence on wage dispersion say on potential effects of the official encouragement to decentralized negotiation in Argentina and Uruguay? ⁽¹⁸¹⁾ Consistently with the expansion of decentralized bargaining in Argentina following the wage policy change as from 1991, between 1991 and 1993 the dispersion of wage increases in manufacturing jumped (Marshall, 1995),⁽¹⁸²⁾ showing the effects of reduced coordination. In 1991-95, the dispersion of wage increments was visibly larger than in 1987-91, a period where wage negotiation had been allowed (coefficients of dispersion: 1987-91 = 0.11, and 1991-95 = 0.23). ⁽¹⁸³⁾ And in Uruguay too, state's withdrawal from negotiation since 1992 and the incipient decentralization of bargaining appear to have affected the dispersion of manufacturing wage increments in the short period (the coefficients of dispersion, rather low, of wage increments are 0.08 and 0.11, for 1989-92 and 1992-95, respectively).⁽¹⁸⁴⁾ These results suggest that state promotion of decentralized bargaining led to the (expected) less coordination in wage determination and more diversity in wage change (although, judging by the more thorough study of the Argentine case [Marshall, 1995], with little indication of the wished-for association with differential productivity growth).

6. FINAL REMARKS

Looking specifically at labour market programmes, it is clear that in Argentina, Chile and Uruguay their implementation is fundamentally distinct from the actual workings of model variants in OECD countries, particularly in the most advanced. First, as we have shown, governments' commitment in terms of expenditure and coverage is clearly much weaker than in the OECD. Second, the analysis revealed that there is a preference for indirect incentives (Argentina) and for measures addressed to employed adults (Chile), both of which received less state support in the OECD. Third, even though unemployment insurance could hardly be labelled as generous in these countries, it has been questioned in the light of presumed negative effects on the work incentive that not only could not reasonably be expected in countries where the requisites for access in fact exclude more than 90 per cent of the unemployed but neither were demonstrated to dominate in OECD countries with much more liberal schemes. On the other hand, if we consider the model guidelines in relation to the removal of obstacles that hinder employers' freedom of recruitment, deployment and relocation of workers, and to the decentralization of wage negotiation, these seem to have been followed much more closely (again, particularly and earlier, in Chile).

Apart from the contrasts between Argentina, Chile and Uruguay on the one hand, and advanced OECD countries on the other, there is substantial divergence in the application of the model among Argentina, Chile and Uruguay. Differences are, to some extent, the result of disparities in social and political pressure. The model's core (relaxation of labour protection and decentralization of wage bargaining) was instituted in Chile under the unconstrained force of a military regime, whilst it is being implemented in Argentina under the disciplining impact of large-scale surplus labour, but within the restrictions imposed by the political requirements of democratic functioning. Even though in Uruguay the situation resembles that of Argentina, the impact of unemployment seems to have been less harsh (changes over time were less dramatic) and political constraints stronger, so that labour reforms are more gradual and less sharp. In Chile, once the military were replaced by the civilian government, unemployment declined and the unions were somewhat strengthened, labour market intervention reversed partially in some respects. In any case, within Latin America, Chile continues to be the paradigmatic case of the neoliberal model of state labour market intervention and, although in a more favourable position than formerly, trade unions are still weak.

One significant point is that while governments in all three countries, and particularly in Argentina and Uruguay, are pressed by the need to attain more adequate employment growth rates and reduce unemployment, in all three state labour market intervention is constrained by the policy decision to maintain fiscal discipline. In Argentina, for instance, social and political restrictions have obstructed the substantive reform of the taxation system that would make it possible to increase genuine government revenues, by way of imposing higher and progressive taxation on personal income as is usual in the advanced countries, and this has placed the burden of fiscal control on state expenditure.⁽¹⁸⁵⁾ Notably, the policy preference in Argentina for across-the-board indirect incentives to stimulate investment and employment at the cost of reduced government revenues, as is the case with the rebates on payroll taxes, clearly conflicts with the aim of controlling the fiscal deficit, while the state self restrains its responsibility for a more active intervention in employment regulation. But active labour market intervention seems to have resulted from the compromise between, on the one hand, (ceteris paribus) constraints on expenditure and, on the other, the need to manage social conflict and avoid the potential unrest that could be exacerbated by pervasive unemployment -- an aspect that becomes ever more demanding as election times draw nearer. This compromise took the form of limited but widely publicized activity, involving little direct spending, through the proliferation of programmes against unemployment, each one short lived and with small actual incidence, instead of that of extending the entitlement to unemployment compensation to the labour force segments excluded from access. In contrast to unemployment insurance schemes that guarantee income support to all those who are entitled (for whom it is a lawful right), programme participation is a particularistic benefit granted by the federal or local governments, that could be withdrawn equally easily, and the criteria for determining the expected number of beneficiaries are often obscure. Employment qua anti poverty programmes, apart from being oriented by a kind workfare ideology, give governments the possibility of more a discretionary manipulation of benefits.⁽¹⁸⁶⁾ This less universal scheme of income support through ad hoc programmes for particular groups, defined in terms of very specific characteristics, fits better with the preoccupation about the fiscal deficit: expenditure is more confined, and can be more easily adjusted to conjunctural circumstances; programmes can be created and as well eliminated, enlarged or contracted, at the Executives' whim, as generally they do not require Parliamentary approval; and funds administered by local governments might be diverted more easily away from their original destination to other purposes.

In Chile, civilian governments since 1990 have opted for continuing with the (still timid) emphasis of labour market intervention on training, in particular for employed workers, and this is consistent with the fact that enhancing external competitiveness is crucial for the prevailing growth strategy. It has been possible to do so for as long as labour demand has been on the rise, but developments by the mid 1990s (the slowing down of employment growth) open up an interrogation as to the forms of state labour market intervention in the near future, as does the uncertainty as to how solid is the growth base.⁽¹⁸⁷⁾ Promotion of the new "privatized" unemployment insurance scheme and the attempts to minimize employer severance pay insinuate that the switch to indirect incentive policies expected to favour employment growth may be given renewed impetus. Training policies helping to rise the competitiveness of manufacturing should be crucial to Argentina's growth strategy since 1991, and no doubt they were incorporated into the government's agenda of policy objectives, but this aim seems to have been overshadowed by the more pressings need of fostering employment creation and reducing unemployment, given that the very continuity of the ongoing economic strategy might be jeopardized by the devastating social consequences of long lasting unemployment.

In Uruguay, with high structural unemployment that affects very particularly the youth and shows no sign of receding despite economic expansion, the main traditional mechanism of absorption of excess labour, state employment, has been inhibited; and, as in Argentina, labour substitution seems to be one crucial impediment to employment increases in the private sector. The void left by the retrenchment of state employment was not filled as yet by new forms of state intervention; these are being set up, but the process seems to be slow and so far with small potential impacts.

On the other hand, whereas decentralization of wage negotiation was consolidated in Chile well before the 1990s, in Argentina, and to some extent also in Uruguay, substantial state effort has been geared to strengthen the role of bipartist and decentralized collective bargaining in the determination of a wider array of labour standards than what was customary in these countries, standards that until the 1990s had been regulated by the labour legislation of general application. In Argentina, although this has given motivation for repeated conflict in state-trade unions relations, decentralization did advance substantially and general labour protection was curtailed, both processes being facilitated by the impact of large-scale unemployment in weakening labour; nevertheless, the advance has taken place along successive "stalemate" situations, particularly at times of approaching elections. Once again, progress in the same direction has been more gradual in Uruguay, where political and trade union opposition was able to play a more important role. Whether bargaining decentralization will lead to consequences beyond greater inequality and weaker unions, such as general positive economic effects, is, at best, uncertain.

Active labour market programmes, particularly direct or subsidized job creation, offer to a segment of the unemployed some, albeit temporary, relief. But, as we have seen, they alone cannot solve the unemployment problem. In the same way as programmes for training the poor cannot reduce poverty unless enough jobs for the workers that were trained are created at the same time, circumscribed employment programmes generally end by sending back the participants to unemployment. The policy combination to combat unemployment requires a sound diagnosis of the latter's main causes, on which further research is indispensable, and the state's firm and more direct involvement in employment generation.