FOREWORD 3. Foreword

Transcription

1 FOREWORD 3 Foreword The OECD Council decided to open accession discussions with Chile on 16 May 2007 and an Accession Roadmap, setting out the terms, conditions and process for accession, was adopted on 30 November In the Roadmap, the OECD Council requested a number of OECD Committees to provide it with a formal opinion. Accession discussions are currently ongoing. The Employment, Labour and Social Affairs Committee (ELSAC) was requested to review Chile s labour market and social policies in order to provide a formal opinion on the degree of coherence of Chile s policies with those of OECD member countries. In light of the formal opinions received from OECD Committees and other relevant information, the OECD Council will decide whether to invite Chile to become a member of the Organisation. This report, prepared as part of ELSAC s accession review, highlights some of the key labour market and social policy challenges facing Chile. The formal opinion of ELSAC mentioned above will be sent separately to the OECD Council and the findings set out in the present report are without prejudice to the subsequent discussions and decision of the Council concerning the accession of Chile to the Organisation. Chile has made impressive progress in recent years: two decades ago Chile s GDP per capita was 18% that of the United States; by 2006, it had reached 29%, above the level in Mexico and close to that of Poland. Furthermore, the national poverty rate has fallen dramatically from almost 39% of the total population in 1990 to less than 14% in But there is still some way to go as the jobs that are created in Chile are often characterised by low productivity, low pay and poor working conditions, and compared with OECD countries informal employment and income inequality remains high. Female labour force participation, albeit increasing, is still well below the OECD average, and access to jobs for lowskilled youth and certain other groups remains difficult. Chile s pension system is widely regarded as a good practice in international comparisons, and recent reform has further widened its appeal

2 4 FOREWORD as the introduction of a minimum pension addresses the issue of insufficient pension contribution among a large group of Chilean workers. Chile s social policy approach is innovative, as exemplified by the Chile Solidario programme, but the time has now come to further invest in active social policies for the working-age population. The review was prepared by Willem Adema, Anders Reutersward and Veerle Slootmaekers, assisted by Katherine Latour, and under the overall supervision of the Head of the Social Policy Division, Mark Pearson and the Head of the Employment Analysis and Policy Division, Stefano Scarpetta. The report was prepared over Summer 2008 (the most recent data concern that period), and was discussed by ELSAC on 4 November The analysis thus concerns the pre-financial crisis environment.

3 TABLE OF CONTENTS 5 TABLE OF CONTENTS List of Abbreviations...11 Assessment and Recommendations...15 CHAPTER 1. KEY TRENDS: STRONG ECONOMIC GROWTH BUT INSUFFICIENT JOB CREATION A favourable macroeconomic environment Chile has large human resources that are not well utilised...35 Women and youth face difficulties in entering the labour market Job creation in low-productivity sectors The quality of jobs...44 Sub-contracting and worker dispatching...44 The role of informal employment Significant improvements in health, education and housing Poverty has been much reduced, but the income distribution remains very unequal Income inequality in Chile is closely linked with wage inequality Regional diversity Conclusions...60 Annex 1.A1. Measurement Issues on Income Distribution...61 Bibliography...63 CHAPTER 2. TOWARDS MORE EQUAL JOB OPPORTUNITIES Introduction Important issues of labour legislation have yet to be resolved...67 Employment protection legislation (EPL)...67 Sub-contracting and temporary-work agencies (TWAs)...71 Working time Industrial relations and collective bargaining...75 The minimum wage is relatively high Labour taxation and informal employment The Labour Inspectorate (Direccion del Trabajo, DT) and labour courts An atypical unemployment insurance programme (UI) The public employment service...90 Active labour market programmes (ALMPs) are relatively insignificant...91

4 6 TABLE OF CONTENTS 8. Job-related training and life-long learning...92 Strengthen SENCE s controls and speed up the development of skill certifications Special issues concerning employment of women Conclusions...97 Annex 2.A1. The Presidential Advisory Commission Work and Equity...99 Annex 2.A2. Measuring Employment Protection Legislation (EPL) by the OECD scoring method Bibliography CHAPTER 3. REDUCING POVERTY IN THE WORKING-AGE POPULATION Introduction Social policy Overall spending and redistributive effect Education, health and housing policies An integrated approach towards supporting the poor: Chile Solidario Conclusions Bibliography CHAPTER 4. THE NORMALISATION OF CHILE S PENSION SYSTEM Introduction The Chilean pension system Poverty among elderly and redistribution of pension spending The basic solidarity fund The mandatory private pension system Disability and survivor coverage: reduce cost-shifting while maintaining efficiency Conclusions Annex 4.A1. Background Data to Chile s Private Pension System Annex 4.A2. The Chilean Pension Market: Competition, Individual Choice and Financial Risk Management Bibliography...180

5 TABLE OF CONTENTS 7 List of Tables Table 1.1. Employment/population ratios by gender for selected working-age groups in OECD and accession countries in Table 1.2. Labour force status of the Chilean population, Table 1.3. Youths in 2006: unemployment, labour force participation and the proportion who neither study nor work...40 Table 1.4. Profile of women between years old by per capita income quintiles, Table 1.5. Employed persons by economic sector in Table 1.6. Employment by economic sector in Chile and OECD countries, Table 1.7. Relative wages in selected economic sectors...43 Table 1.8. Activity rate, informality and earnings in the population by education, Table 1.9. Housing conditions in 1990 and Table Poverty trends...54 Table Poverty and inequality indicators, mid-2000s...55 Table Distribution of household market income...57 Table Distribution indices for wages and other household market income, Table Regional differences, Table 1.A.1. Comparing data on income distribution...62 Table 2.1. Incidence of fixed-term and other temporary labour contracts in Table 2.2. Employed persons by effective working time...74 Table 2.3. Trade union membership in Table 2.4. Wage adjustments resulting from collective agreements...77 Table 2.5. Labour inspections...82 Table 2.6. UI benefits in June Table 3.1. Poverty and its severity has fallen since Table 3.2. Public spending on education and health is the most important in redistributing resources towards the poor in Chile Table 4.1. The Basic Solidarity Pension (PBS): evolution of payment rates and expected coverage Table 4.2. Distribution of pensioners by payment methods Table 4.A1.1. Most outlays on pension payments are through annuities rather than programmed withdrawals Table 4.A1.2. AFP portfolio characteristics and yields and default age rules for clients List of Figures Figure 1.1. Chile s annual GDP growth compared with OECD and selected Latin American countries...34 Figure 1.2. Chile's exports...35 Figure 1.3. Population by age class...36

6 8 TABLE OF CONTENTS Figure 1.4. Evolution of employment/population ratios by gender and age groups, Figure 1.5. Reasons for inactivity among women, Figure 1.6. Employed persons in Latin America not contributing to social insurance...45 Figure 1.7. Informality among employees by age...48 Figure 1.8. Health indicators...50 Figure 1.9. Education outcomes...51 Figure Households owning their main dwellings...52 Figure Change in the Lorenz curve: a measure of declining income inequality..58 Figure 2.1. Measuring employment protection legislation (EPL)...68 Figure 2.2. Minimum wage as a percentage of the average wage in OECD countries and Chile...79 Figure 3.1. Public social and education spending has increased in Chile, but remains low in international comparison Figure 3.2. Public investment in health and education in Chile is limited compared with OECD countries Figure 4.1. Publicly-mandated pension spending in Chile is close to the OECD average Figure 4.2. The income position of the elderly is relatively good in Chile Figure 4.3. Replacement rates in Chile are relatively close to the OECD average Figure 4.4. Pension coverage is increasing but remains low in international comparison Figure 4.5. The contribution density among female workers is relatively low Figure 4.6. Disability benefit replacement rates in Chile are on par with other OECD countries Figure 4.A1.1. Evolution of the minimum wage, average wage of contributors to the mandatory pension system (AFP), and minimum pension payments Figure 4.A2.1. Administrative charges declined until Figure 4.A2.2. AFP profits seem to be at a higher level than in the 1990s Figure 4.A2.3. Most AFP members have investment portfolios with intermediate risks Figure 4.A2.4. Foreign investment by AFPs has grown rapidly over the last decade Figure 4.A2.5. Real rate of investment returns have trended up since the late 1990s List of Boxes Box 1.1. Labour market and social policy recommendations for Chile...31 Box 1.2. Alternative measures of labour informality in Chile...46 Box 1.3. Poverty line in Chile...56 Box 2.1. Severance pay and unemployment insurance in Austria...70

7 TABLE OF CONTENTS 9 Box 2.2. Labour taxation and mandatory contributions...80 Box 2.3. Chile s unemployment insurance (UI, seguro de cesantía)...85 Box 3.1. Public financing of social support in Chile: budgetary prudence and little redistribution in taxation Box 3.2. Anti-poverty policy development and coherence in Chile Box 3.3. Conditional family support in Mexico and holistic service delivery in some other OECD countries Box 4.1. Reforming the pre-1981 PAYG scheme into a private-funded pension system...148

8

9 LIST OF ABBREVIATIONS 11 LIST OF ABBREVIATIONS AFP ALMPs APS AUGE CAN CASEN CCT CEPAL CLP DIPRECA DT EITC ENCLA EPL FONASA FOSIS GDP ICT INE Administradora de Fondo de Pensiones (Pension Fund Management Companies) Active Labour Market Programmes Aporte Previsional Solidario (Pension Solidarity Complement) Acceso Universal con Garantías Explícitas en Salud (Regime of Explicit Health Guarantees) National Accreditation Commission Encuesta de Caracterización Socioeconómica Nacional (National Characterisation Socio-economic Survey) Conditional Cash Transfers Comisión Económica para América Latina y el Caribe (Economic Commission for Latin America and the Caribbean) Chilean Pesos Dirección de Previsión de Carabineros de Chile (Social Security for the Police) Dirección del Trabajo (Labour Inspectorate of the Ministry of Labour and Social Welfare) Earned Income Tax Credit Encuesta Laboral (National Labour Survey) Employment Protection Legislation Fondo Nacional de Salud (National Health Fund) Fund for Solidarity and Social Investment Gross Domestic Product Information and Communication Technologies... Instituto Nacional de Estadisticas (National Statistics Institute)

10 12 LIST OF ABBREVIATIONS INP INSP ISAPRE IUSA JUNJI MdH Mideplan MINVU MPG NEET OMIL OTEC PASIS PAYG PBS PISA PW SENCE SAFP SENAME SIMCE SMEs Instituto de Normalización Previsional (Chilean Social Insurance Agency) Instituto Nacional de la Salud Pública (National Institute of Public Health) Instituciones de Salud Previsional (Private Health Providers) Individual Unemployment Savings Accounts Junta Nacional de Jardines Infantiles (National Childcare Authority, part of the Ministry of Education) Ministerio de Hacienda (Ministry of Finance) Ministerio de Planificación (Ministry of Budget and Planning) Ministerio de Vivienda y Urbanismo (Ministry of Housing and Urbanism) Minima Pensión Garantía (Minimum Pension Guarantee) Neither in Employment nor in Education or Training Oficinas Municipales de Intermediación Laboral (Municipal Employment Intermediation Offices) Organismo Técnico de Capacitación (Technical Training Organisation) Pensiones Asistenciales (Social Assistance Pension) Pay As You Go Pensión Básica Solidaria (Basic Solidarity Pension) OECD Programme for International Student Assessment Programmed withdrawals (Regular pension payments) Servicio Nacional de Capacitación y Empleo (National Employment and Training Service) Superintendencia de Administradoras de Fondos de Pensiones (Supervisory body of Pension Fund Management Companies, see above) Servicio Nacional de Menores (National Service for Young People) Sistema de Medición de la Calidad de la Educación (System for the assessment of quality in education) Small and Medium-sized Enterprises

11 LIST OF ABBREVIATIONS 13 SNED SOFOFA SPR TWA UF UI UTM VAT National Education Evaluation Service Sociedad de Fomento Fabril (Chilean Federation of Industry) Ficha de Protección Social (Social Protection Record) Temporary work agencies Unidad de Fomento (Unit of account which adjusts the value of the Chilean Peso on a daily basis) Unemployment insurance Unidades tributarias mensuales (Unit for monthly tax payments) Value-added Tax

12 ASSESSMENT AND RECOMMENDATIONS 15 Assessment and Recommendations Towards more inclusive growth in Chile Chile has started to catch up to OECD standards of living Chile has enjoyed an impressive economic performance over the past two decades. Apart from a cyclical slowdown during , real GDP increased by nearly 6 % per year over the period Despite relatively strong population growth up from 12.1 million in 1985 to 16.6 million in 2007, real GDP per capita rose by 4.3% per year over the past two decades. This impressive performance has resulted in a significant catching up with the OECD countries: two decades ago Chile s GDP per capita was 18% that of the United States; by 2006, it had reached 29%, above the level in Mexico and close to that of Poland. The national poverty rate has also been reduced dramatically, from almost 39% of the total population in 1990 to less than 14% in 2006, with the share of individuals living in extreme poverty down to close to 3% in 2006, from about 10% in but labour utilisation remains low and inefficient in a segmented labour market Despite this significant progress, Chile still has a long way to go in catching up with average OECD living standards and in reducing inequalities. One of the main requirements to sustain economic growth, and to ensure that all benefit from it, is to foster the creation of more and better jobs. Labour utilisation is still relatively low with few job opportunities for certain groups. Although female labour force participation has risen steadily over the past two decades, at 39% the female employment rate is 33 percentage points below the male employment rate and well below the OECD average of 57%. At the same time, youth participation rates have declined, and only about a quarter of youth (15-24) is in employment, compared with 44% on average in the OECD. As elsewhere, this in part reflects efforts to promote further education. But for many youth with low skills, access to jobs remains difficult in Chile.

13 16 ASSESSMENT AND RECOMMENDATIONS The jobs being created in Chile are often characterised by low productivity, low pay and poor working conditions. International comparisons are difficult, but indicators suggest that informal employment remains significant (albeit receding and lower than in most of Latin America). About one-fifth of all Chilean employees either did not have a formal labour contract or did not contribute to social security in In addition, many self-employed workers, who account for one-fifth of total employment, should probably be included in any broad count of informal employment as only 22% of them make social insurance contributions on a voluntary basis. Informality is most common in the low-productivity segment of the labour market, where low-skilled employees are often hired without a contract by small firms that are themselves informal. These workers generally lack old-age, unemployment or health insurance and are not covered by the labour law provisions (minimum wage, employment protection or occupation health and safety standards). They are often poorly paid, with limited access to training and career advancements..leading to high levels of inequality Persistent segmentation in the labour market is one of the key factors behind the marked inequalities in earnings and income in Chile. Measured against the standard OECD relative poverty benchmark (equivalised disposable income less than 50% of the median income), Chile s poverty rate at 16.4% of the population would put it at the bottom of the OECD ranking, similar to the relative poverty rates observed in Mexico, Turkey and the United States. In terms of income inequality, Chile also fares worse than all OECD countries, with a Gini coefficient of 0.53 compared with an OECD average of To a large extent, income disparities in Chile are closely linked to wage inequalities that are not sufficiently smoothed by the tax and benefit system. The macroeconomic policy framework is sound and is serving the economy well, but structural barriers to job creation persist The macroeconomic policy setting in Chile, combining rules-based fiscal management, inflation targeting and a flexible exchange rate, has played a major role in promoting strong and sustained economic growth. In particular, a sound fiscal framework enhanced in 2006 with the fiscal responsibility law has allowed the authorities to avoid a pro-cyclical stance in a context of booming world prices for Chile s main export, copper. Chile has managed to reduce public indebtedness, and has become a net external creditor. Recently, world copper prices have started to fall and a significant slowdown in the

14 ASSESSMENT AND RECOMMENDATIONS 17 world economy unfolding from the financial crisis could potentially put further downward pressure on copper prices in the near future. However, it is too early to assess the extent of the correction in copper prices and the impact on Chile s fiscal and macroeconomic position. Sound macro-economic policy has helped Chile to avoid the raw material curse that affects many developing countries whose economies are heavily dependent on primary commodities. Despite being the world s biggest producer and exporter of copper, Chile has succeeded in diversifying its economy, with significant increases in the export shares of other than for copper products, especially in the agro-food sector. But a number of structural barriers persist that limit the creation of productive jobs. Over the past two decades, economic growth in Chile has been largely driven by physical capital accumulation and productivity growth. Employment growth has barely matched the increase in the working-age population and thus labour utilisation has not improved significantly. And while employment has remained fairly stable in the productive manufacturing sector, its share in total employment is well below what could be expected given Chile s GDP per capita. Job creation has largely taken place in generally low-productivity areas and sectors dominated by SMEs such as retail trade, hotel and restaurant and personal services. Promoting the creation of more productive jobs requires a comprehensive strategy that focuses on increasing the creation of new firms, fostering the expansion of successful ones, and improving the productivity performance of existing businesses. Much can also be done to improve human capital. Together with a judicious expansion of social programmes, this will also generate a more equitable distribution of the benefits of growth. The OECD Reassessed Jobs Strategy provides a good framework to promote more and better jobs in Chile The 2006 Reassessed OECD Jobs Strategy 1 provides a comprehensive policy framework for boosting jobs and income in OECD countries. The 1. The OECD Jobs Strategy was originally formulated in 1994 and was aimed at reducing high and persistent unemployment. While the key recommendations have been found to be useful in this respect, the policy focus has broadened and the 2006 Reassessed OECD Jobs Strategy (RJS) puts more weight to the objective of promoting labour market participation and employment, and taking into account concerns about low incomes of certain groups. The RJS has four main pillars: i) set appropriate macroeconomic policy; ii) remove impediments to labour market participation as well as job search; iii) tackle labour- and product-market obstacles to labour demand; and iv) facilitate the development of labour force skills and competencies. Within each of these pillars, the Reassessed Jobs Strategy contains specific on-binding recommendations for the OECD countries. All countries need to ensure that each of

15 18 ASSESSMENT AND RECOMMENDATIONS Strategy identifies a number of common principles that are also relevant to Chile: beyond ensuring stable macroeconomic conditions, where Chile has already achieved impressive results, the Reassessed Jobs Strategy stresses the need to establish a simple, transparent and not overly cumbersome regulatory environment in which firms can find the right incentives to invest, hire and train more workers (formally) and ultimately promote productivity and output growth while workers have incentives to search for jobs that match their capabilities and to invest in improving their skills and competences. Previous OECD reports (OECD Economic Surveys: Chile, 2005 and 2007) provide a comprehensive review of the key policy challenges to improve the business environment in Chile. But much can also be done in the labour market to enhance job opportunities for all. First, there is a need to improve the balance between employment security and labour market flexibility. Related to this is also the need to price the low-skilled back into the formal sector. While further investment in education and training is required to boost the level of human capital and its distribution, policy should also review existing taxes, social security contributions and wage floors that may price some low-skilled workers out of jobs. Second, Chile also needs to take additional measures to remove existing barriers to labour force participation and access to formal-sector jobs of under-represented groups, including women, youth and more generally low-skilled workers. Improve the balance between employment security and flexibility The recent report by the Presidential Advisory Commission for Work and Equity includes many suggestions for a large reform package of labour market and social policies to help generate more and better jobs, which are in line with the OECD Reassessed Jobs Strategy and the related international debate on flexicurity. However, the detailed policy design and implications of an appropriate flexicurity strategy for Chile may often be different from that required in many OECD countries in view of Chile s historical development and current position. In particular, Chilean policy makers will have to pay attention to the development of administrative capacity and quality to make reform of employment and training services or in-work benefits for the low-income population a success. At the same time, in some cases, the regulatory environment should be relaxed in order to permit effective enforcement since this is the problem with many existing labour laws. the four pillars is solid. However, within each pillar there may be scope for individual countries to use different policy combinations to achieve successful outcomes, taking into account policy interactions and country circumstances and objectives.

16 ASSESSMENT AND RECOMMENDATIONS 19 Employment protection legislation contributes to the dualism in the labour market Chile s employment protection legislation (EPL) for regular contracts is not overly strict by OECD standards. Until 2007, however, it was coupled with unregulated temporary work agencies (TWAs) and sub-contracting employment arrangements. These latter two atypical forms of labour contract have grown significantly in Chile: for example, in 2006 over 40% of employers reported using sub-contractors, and one-third of them declared that their work was part of the principal activity of the enterprise. Subcontracting has led to a fragmentation of enterprises: the so-called phenomenon of 100 firms within a firm. For mother firms or subcontractees, this has given flexibility and has lowered labour costs, although, if sub-contracting is widespread, it may lead to an inefficient firm structure, and generate relatively high administrative and organisational costs. For the workers concerned, typically low-skilled, the result has been precarious jobs with no opportunity to organise and bargain collectively and little or no investment in training. Taken together, sub-contracting and TWAs have contributed to a further element of segmentation in the labour market, not only between informal and formal-sector workers but, among the latter, between those with a regular and those with an atypical contract. the 2007 reform is welcome, but implementation of new subcontracting legislation is difficult. The 2007 reform restricted sub-contracting to jobs that are separate from the firm s own work process. In addition, the client firm must give certain guarantees, e.g. in the event of a sub-contractor s failure to pay wages, so that the client firm performs a role that in many OECD countries would be normally undertaken by the Labour Inspectorate. TWA jobs are now only allowed for specific tasks and for limited periods of time (three to six months). This new law is welcome as it will reduce uncertainty and inequality among workers in the formal sector, but its implementation faces some difficulties. While in some sectors it has already had an effect (e.g. banks no longer sub-contract to have cashiers at the till), some larger firms have challenged the powers of the Labour Inspectorate in complex court cases, and the Supreme Court has partly sustained such complaints. Moreover, attempts to enforce the new rules in the mining sector faced such strong legal and political opposition that the government had to suspend implementation, at least temporarily. Despite these setbacks, the Chilean government should continue its efforts to introduce and enforce the new sub-contracting regulations.

17 20 ASSESSMENT AND RECOMMENDATIONS Reform of EPL for regular workers should also be considered, especially concerning severance payments Reducing the extent of sub-contracting and use of TWAs may push jobs into the informal sector, unless steps are taken to make formal employment more attractive. It is therefore important to also reconsider elements of the EPL for regular workers. While some aspects of EPL for regular workers are relatively loose from an international perspective notice periods and administrative procedures others are quite costly and subject to abuse, e.g. severance pay. Severance payments can be as much as 11 months wages in Chile, compared with three to four months in many OECD countries. The high costs associated with severance payments provide employers with incentives not to formalise employment relationships. Moreover, as an income-support scheme for the unemployed, severance pay has a number of drawbacks. With present rules, it concerns only those dismissed for economic reasons from indefinite-duration jobs, a group representing only a small proportion of those at risk of unemployment. Moreover, it offers only limited pooling of unemployment risk: the risk is pooled only among workers in a given firm but not across firms. Moreover, severance pay does not protect workers against the risk of a long spell of unemployment, and firms compliance with severance pay rules tends to be fairly low, even for formal workers with indefinite contracts. Indeed, there is evidence suggesting that in Chile employers often avoid paying the full amount of severance payments by reaching an agreement with workers, or simply by refusing to pay. Non-compliance also creates a burden on labour courts and government budgets. which could be phased out while increasing the importance of both IUSAs and the Solidarity Fund... Befitting its experience with private individual pension accounts (see below), Chile has introduced a unique unemployment compensation system based on individual unemployment savings accounts (IUSA), with a complementary Solidarity Fund for low-income clients dismissed for economic reasons. Employers contribute to both IUSAs and the Solidarity Fund, and are partially compensated for the associated increase in labour cost by a reduction in severance payments. Benefits are paid from IUSAs regardless of the reason for the separation. Because they are pre-paid, IUSAs do not increase the direct cost of employment adjustments for employers, as is the case with severance payments (even if they are only partially enforced). In

18 ASSESSMENT AND RECOMMENDATIONS 21 addition, since IUSAs are owned by workers, they are unlikely to have a great effect on incentives to search actively for a suitable job. Compared with traditional unemployment insurance schemes, IUSAs are therefore less demanding in terms of job counselling and activation, a particularly important factor in Chile given the limited resources available in the labour offices. The Chilean government should complete the reform by phasing out severance pay and further increasing the role of unemployment benefits. While this may seem a bold step, Austria achieved a similar result in 2003 by converting its severance-pay system into individual saving accounts, while continuing to implement unemployment insurance and unemployment assistance. But if a full phasing-out of severance payments proves too controversial in Chile, a second-best option would be to lower its level considerably to a maximum of three to four months, as observed in many OECD countries. and in particular reinforcing the solidarity component of the dual unemployment benefit scheme The subsidised Solidarity Fund provides supplementary income support to claimants who are unemployed and have exhausted their own IUSA funds, thus allowing for some pooling of the unemployment risk. Benefits under the Solidarity Fund are paid at fixed replacement rates and require registration with the municipal employment service. In theory, this dual system of individual saving accounts backed-up by a Solidarity Fund, which is similar to a more traditional unemployment insurance programme, can combine the benefits of both types of schemes. However, overly strict eligibility criteria have effectively curtailed access to the Solidarity Fund so far. A recent reform is addressing this issue by making the Solidarity Fund s benefit scale more generous and applying it as a rule to all terminations of indefinite-duration contracts for economic reasons. Moreover, those terminating fixed-term contracts would also be eligible, though with a much less generous benefit scale. Recipients would be obliged to seek jobs and register with the public employment service as unemployed. It is still too early to assess whether these changes will increase access to the Solidarity Fund, but, if they do, stronger efforts by the public employment services would be required to enforce the job-search requirement in order to offset the adverse effect of more generous benefit entitlements on search activity.

19 22 ASSESSMENT AND RECOMMENDATIONS Strengthening the role of the social partners in setting wages and working conditions could also contribute to flexicurity Labour relations in Chile are generally confrontational and marred by lack of trust, a factor that is becoming increasingly problematic for the development of flexicurity in Chile. This is partly due to the limited coverage of unions and business associations. For example, union membership accounts for only 11% of the workforce and is concentrated in big firms and certain sectors, particularly mining (42% of employees are unionised). At the same time, the weak and unpredictable enforcement of labour regulations fosters a confrontational climate. In this context, the phasing out of severance payments in favour of a more widespread and generous unemployment benefit system could weaken one of the main sources of disputes. Promoting less confrontational and more co-operative labour relations, while simplifying labour regulations and strengthening their enforcement, is essential to reduce costs to both parties and spread benefits to a wider group of workers, rather than to a relatively small group of insiders. The government intends to develop new legislation to promote collective bargaining in enterprises and to make it richer in content. It would thus encourage more bargaining over matters other than wages, and it would seek to strengthen unions representativeness in enterprises. A procedure would be established for strengthening union representations on issues related to working hours. The government has also offered to support training of trade union representatives in enterprises. Moreover, as part of its effort to develop more effective industrial relations, the government has begun to implement a significant reform aimed at improving the court system for labour cases, combined with optional fast-track procedures for resolving conflicts in meetings with the Labour Inspectorate. These initiatives all go in the right direction of promoting collective bargaining and creating a clearer and less confrontational environment for social dialogue. But the government has good reason to consider which further institutional developments might be needed in order to promote dialogue and consultation with employers and unions as well as to strengthen trust between the social partners. The minimum wage is relatively high Expressed as a proportion of the average wage, the legal minimum wage was raised substantially between the mid-1990s and By mid-2008, it was a little over 45% of the average wage while the OECD

20 ASSESSMENT AND RECOMMENDATIONS 23 average is 40%. OECD Economic Surveys of Chile (2003) found evidence that the increase until the early 2000s had contributed to a rise in unemployment, a result that still appears relevant. Moreover, despite the fact that informal work arrangements are sizeable among the low skilled, the minimum wage tends to operate as a strong pay signal for the informal sector as well. This implies that hikes in the minimum wage can have distributional implications that go beyond the formal sector: the wages of the low paid might increase in both segments of the economy, but their employment prospects might also decline in both sectors. The Chilean authorities are aware of this risk so they reimburse half of the cost of the minimum wage to employers who have to pay the minimum wage to apprentices and participants in the similar Jovenes Chile Solidario scheme who are not yet 25 years old. Chile s minimum wage is lower for those aged under 18 (and those over 65, as well as domestic workers). However, under-18 youth are often in school, while the minimum wage is likely to affect particularly the year-olds, whose employment rate at 26% is extremely low in Chile today. A weak school-to-work transition process is one of the biggest failings in Chilean policy. Reducing youth wages would be a price well worth paying if it increased their employment rates. The Chilean authorities should consider increasing the age for receiving the full minimum wage to 25, and pay lower rates (increasing with age) to younger people, as is the case in some OECD countries, e.g. the Netherlands. and in future its level should be considered in view of the planned introduction of in-work benefit payments The income tax burden is not high in Chile. Income tax is charged on incomes higher than about 1.4 times the average wage, with an initial marginal tax rate of 5% rising to 40% (for those with wages in excess of about 15 times the average wage); social security contributions (towards pension, health, unemployment and work injury insurance) represent about 26% of the wage. The government envisages introducing in-work benefits (in the form of cash transfers) in 2009 as a way of supporting those in low-paid jobs. Such benefits have proven to be effective in raising labour supply among primary earners and adults in sole-parent families in many OECD countries that have introduced them, but the evidence also indicates that basing such policies on household earnings can reduce the labour supply of second earners in couple families. Compared with schemes based on household earnings, in-work benefits based on individual earnings are more expensive as they cover more workers at a given income threshold. However, given the extremely low rates

21 24 ASSESSMENT AND RECOMMENDATIONS of female employment in Chile, it would be unfortunate if incentives for women (usually the second earner in the family) to work were reduced. Despite the extra cost, a system based on individual earnings would likely be more efficient in the Chilean situation. Whatever the basis for setting up in-work benefits, their introduction argues in favour of avoiding further significant hikes in the minimum wages to minimise dis-employment effects among the low skilled. which, if based on householdearnings, should be tied in with childcare supports. Low female labour supply has been one of the main reasons why discussion of public support for childcare has increased in Chile. Public support has been pledged to expand capacity, with new facilities over the period (compared with facilities created over the period). Nevertheless, public spending on childcare and pre-school education is low at about 0.1% of GDP in 2007, compared with an OECD average of about a quarter percent of GDP. A further increase in childcare supports would therefore be justified if public finances permit this since all the evidence suggests it would lead to rising female employment. If Chile opts for introducing a household-earnings-based in-work benefit for low-income workers, it would be important to integrate the benefit rules of such a scheme with childcare support, for example, by paying higher childcare support rates in case where both adults are in employment for at least 30 hours per week. Increased benefits to the unemployed will require an effort to enhance employment services and job-related training. An appropriate flexicurity package in Chile, which eliminates severance pay but includes more generous unemployment benefits, childcare supports and in-work benefits, would need to be complemented by enhanced employment services and activation measures in order to maximise job-search activities and minimise moral hazard. However, delivering such programmes effectively would pose a major challenge for Chile, especially at the local level. To facilitate this, a merger of the municipal employment service offices with the national employment and training agency (SENCE) should be considered. Resources devoted to public employment services amount to only about 0.1% of GDP, compared with 0.4% on average across the OECD countries. With gradually improved unemployment compensation, more clients are likely to register

22 ASSESSMENT AND RECOMMENDATIONS 25 with the employment service and there will be scope to introduce at least some elements of the activation strategy implemented in many OECD countries, in particular job counselling, but also referral to training. Steps in this direction have been taken in recent years, but the relevant public institutions have very modest resources by OECD standards and they have limited experience of implementing targeted programmes for job seekers. It would be important to monitor the performance of the various labour market programmes. Use of private and non-profit organisations to deliver employment services on a competitive basis needs to be considered, given the lack of administrative capacity in the public sector. Examples of the private sector playing such a role in the employment services market can be found in Australia, the Netherlands and the United Kingdom. Another important challenge is to develop institutions for lifelong learning. Most of the SENCE-sponsored training until now has concerned employees in enterprises, which are obliged to pay a 1% payroll tax or spend the corresponding amount on training, whether in-house or in approved training centres. The range of training courses available to job seekers is restricted, and the possibilities of expanding them will probably remain limited in the near future. But in the longer term, a broad policy package called Chilecalifica co-ordinated by the Ministries of Economy, Education and Labour aims to develop institutions for lifelong learning, covering general as well as vocational training for adults. A national certification system for job skills, still incomplete, is expected to increase efficiency in the labour market and in the markets for training in the future and this effort should be speeded up, if possible. Of course, this must be done in the context of an education system that gives all youth good-quality education. The 2004 OECD Review of Education Policy also encouraged the Chilean authorities to take further steps to improve the technical/vocational focus in education and strengthen linkages between education and enterprises through apprenticeships and the extension of vocational secondary education in industrial fields. These recommendations are still valid today. Invest more in active social policies for the working-age population In its quest for more inclusive growth, social policy is moving away from merely alleviating basic needs to improving the quality of supports Chilean social policy has been more active and innovative than in many OECD countries. In the recent past, public social policy has focused on removing shantytowns and helping low-income households to acquire inexpensive accommodation, extending primary and secondary education

23 26 ASSESSMENT AND RECOMMENDATIONS to all children, and making health care affordable to all. But notwithstanding the sizeable decline in poverty, income disparities remain enormous, and these outcomes have triggered a search for policy options that could generate more inclusive growth. Chilean policy is moving on from an emphasis on extending capacity towards an increased focus on improving the quality of dwellings, neighbourhoods, healthcare and education and social services more generally. Chile s innovative approach has a strong focus on integrated service delivery Rather than focusing on income support to the working-age population, the Chilean social policy response through its innovative Chile Solidario framework connects low-income families, of whom many are sole-parent families, with support networks. Through sustained and regular psycho-social support, social workers help clients to assert their rights to a range of potentially available services such as healthcare, childcare, education and other family support services. The most recent Chilean social policy initiative Chile Crece Contigo (Chile Grows with You) is aimed at mothers and children until entry in pre-school age, and is similarly broad in nature, providing a range of supports touching all relevant aspects of the family environment, health, cognitive development and education. and involves a variant of mutual obligations that helps clients make use of services From the client perspective, Chile Solidario is to be the gateway to a range of networks and services (and some financial support). Client s active participation is key to success. By means of a small participation bonus (USD 10 to 24 per month, depending on the length of participation), authorities provide clients with a small financial incentive to sign up to Chile Solidario, and commit themselves to a wide variety of actions including, for example, vaccinations for children, participating in medical check-ups and/or training or education. The Chilean variant of mutual obligations in benefit receipt ties participation to access to services, not to receipt of financial transfers on a comprehensive basis.

24 ASSESSMENT AND RECOMMENDATIONS 27 but capacity constraints have contributed to limited success in getting clients in employment. With limited spending on income support and little success in raising employment, Chile Solidario has not been very effective in reducing poverty until now. The emphasis on developing access to services, which supports people in helping themselves, is an innovative approach that has borne dividends, but which has its limits. The existing service infrastructure in Chile has been inadequate to help many of the Chile Solidario clients into paid employment, which is a condition for any sharp reduction of poverty. Chile needs to continue its investment in the infrastructure of service provision, e.g. training of social workers and employment service staff, especially at the local level. Chile has a good track record in targeting help on those who need it most, but further extensions pose new challenges The Chilean system is good at identifying needy clients. The Social Protection Record (SPR) covering about half the population has provided authorities with a good profile of low-income households, and is increasingly used to target education, healthcare, housing and social support. However, the SPR intends to measure long-term household income capacity. The SPR does not focus on current income streams, which only account for about 10% of the SPR score, and does not allow for phasing out or gradually withdrawing income support. However, as SPR scores are increasingly used to assess eligibility for cash transfers, policy has either to adapt the SPR so as to accurately measure current income changes, or design a supplementary tool alongside it to measure short-term income changes. Clients should also be asked to give officials permission to verify such information with the relevant tax and/or social insurance authorities. The introduction of basic pensions has strengthened the retirement income system The system of individual accounts brought much-needed transparency to Chilean pension saving The famous mandatory private pension system restored confidence in pension saving in Chile. Collective pension schemes which no longer operated in a fair manner were replaced by a system of individual accounts

25 28 ASSESSMENT AND RECOMMENDATIONS operated by pension fund management companies (AFPs). This involved administrative costs that are relatively high compared with private occupational pension systems, but by linking individual contributions to benefit payments, the system introduced much-needed transparency. and the privately managed pre-funded system has performed well overall. The pre-funded Chilean system has performed well over the years; it has deepened Chilean capital markets, assisted economic development, and rates of return on investment have been higher than originally envisaged. Furthermore, the pre-funding of the system helps spread the costs of dealing with ageing populations. The mandatory private pension system has, however, failed to achieve some of its objectives. Competition in the pension market is limited, and this has contributed to administrative fees being higher than necessary. Pension reform introduced in 2008 aims to generate price competition by having AFPs compete for a share of the market (new entrants) rather than trying to stimulate competition within the market. Lack of coverage has proven to be the Achilles heel of the private pension system requiring the introduction of a basic pension Arguably the biggest shortcoming of the private pension system is that it failed to achieve coverage for the vast majority of Chilean workers. There are an increasing number of workers with patchy employment profiles, and contribution densities are particularly low among low-income workers with low levels of educational attainment, young workers, female workers, and the self-employed. The Chilean authorities have responded by introducing a basic solidarity pension as of right in At the same time, financial incentives have been introduced to make pension saving more attractive for women and young workers, and promote the development of a third tier of voluntary pension saving. With the development of a basic pension entitlement, mandatory second-tier pension saving and fiscally supported voluntary pension saving, Chile s pension system has moved towards the three-pillar pension norm, which is often held as an international benchmark.

26 ASSESSMENT AND RECOMMENDATIONS 29 there are plans to extend coverage to the self-employed, but success is not guaranteed. Mandatory coverage of pension contributions will be extended to groups of the self-employed by While the intention is sensible, there are concerns that these initiatives may not be successful. For example, the implicit marginal tax rate of 30% on contributions may be a deterrent for many low-income workers to save for retirement and some of the self-employed may stop reporting income to the tax authorities rather than pay contributions towards pension insurance. The disability system is innovative, but new measures need to be closely monitored to see if they achieve their stated goals. The AFP scheme includes provisions for disability insurance, and the involvement of private commercial enterprises in the disability assessment process seems to have curtailed the inflow into the programme. This success, however, is at the expense of the public purse, since unsuccessful applicants are likely to seek recourse to publicly-financed minimum income benefits. Recent reform is likely to increase the cost of the private disability programme as coverage is extended to groups of the self-employed, and costs may rise further because of institutional changes, with AFPs no longer having incentives to keep costs down. The implementation of the disability reform needs to be closely monitored to ensure the right people receive the assistance to which they are entitled. Recent disability reform may be more effective if experiencerating for employers were introduced. Recent reform also made employers responsible for disability contributions. However, employers are likely to pass on the cost of this mandate to their workers in the future. Moreover, the OECD experience suggests that shifting responsibility for contributions to employers will not per se make them more willing to invest in disability-preventing workplace measures. Experience-rating of disability premiums is likely to be more effective in terms of encouraging employers to invest in preventive measures by employers, as the experience in the Netherlands shows. Moreover, the Chilean employment-related occupational accident and diseases scheme involves experience-rating on a sectoral basis. Policy may

27 30 ASSESSMENT AND RECOMMENDATIONS consider extending experience-rating to enterprises for both the employment-related and private disability schemes to encourage employers to invest more in preventive measures. Chile should invest more in employment and active social policies. Chilean social policy for the working-age population is at a crossroads in that, for the first time, all the main political parties are debating how to deal with the in-work poverty that affects large parts of the working-age population. If adopted, the recent proposal for in-work benefits (see above) could cost as much as half a percent of GDP three times the Chile Solidario spending (not counting the basic pension). Strengthening public support for the working-age population would be appropriate. Compared with an OECD average of just over 20% of GDP, Chile spends much less on social protection: about 9% of GDP (not including education). Most of current social protection outlays are towards pensions and/or health, while other social support (including family benefits and housing) only amounted to 1.5% of GDP. As public spending on social support is low, the redistributive power of the Chilean tax/benefit system is relatively limited. If Chile wants to achieve its objectives of more and better jobs and a reduction in inequality, it should spend more on social protection, and particularly on labour market and social policies. There are several promising areas for investment with potentially high social returns including Chile Solidario, in-work benefits, employment and training services and labour market programmes as well as childcare support, all policies that promote labour force participation and self-sufficiency. Comfortable public finances lay the groundwork for expanding labour market and social programmes. But in the future, if the fiscal situation deteriorates, financing these programmes may imply some reallocation of resources and/or increases in taxes that are currently relatively low from an international perspective.

28 ASSESSMENT AND RECOMMENDATIONS 31 Box 1.1. Labour market and social policy recommendations for Chile In the context of its policies to enhance job opportunities for all, the Chilean government is invited to consider the following items as part of its strategy to improve the balance between employment security and flexibility, generate trust between the social partners, and the development of an inclusive and active social policy: Intensify efforts to introduce and enforce the new sub-contracting regulations. Phase out the existing system of severance pay while simultaneously increasing the dual unemployment benefit scheme. Reinforce the solidarity component of the dual unemployment benefit scheme to raise coverage and the pooling of the unemployment risk. Consider merging municipal employment services with the national employment and training agency. Monitor the implementation of the new legislation to promote collective bargaining and consider developing organisations of consultation and dialogue with the unions and employers to improve trust between social partners. Improve the effectiveness of industrial relations, promote bargaining on other issues than wages; enhance the effectiveness of the Labour Inspectorate; and continue to simplify the judicial process regarding labour cases. Increase the age for receiving the full minimum wage from 18 to 25, and prescribe lower rates (increasing with age) for younger workers and further develop a system of apprenticeships. Invest more in the capacity of employment and training services, and the expansion of active labour market programmes, and enhance the application of the national certificate for training. Consider whether it would be feasible to sub-contract the market for employment and training services to private actors. Enhance the effectiveness of in-work benefits by keeping design simple while targeting strictly defined groups of workers. To promote female labour supply, design of the in-work benefit scheme should be linked with childcare supports, for example, by differentiating payment rates with parental employment patterns. Strengthen the redistribution within the Chilean tax/benefit system, proportionally increase allocations for social support within public budgets, and invest more in active social policies, and reinforce the Chile Solidario gateway to further increase access to such support. Re-design the Social Protection Record or introduce a measure alongside it to properly account for changes in current household incomes and facilitate a gradual withdrawal of income support if need be. Clients should also be obliged to report changes in their income situation to the social services (and give permission for verification of such information with relevant authorities). Monitor carefully the reform that schedules the extension of mandatory private pension coverage to the self-employed, and be ready to make changes if outcomes are not in line with expectations. Monitor closely the recent disability reform to ensure the right people receive the assistance to which they are entitled. The Chilean authorities should also consider introducing experiencerating of disability premiums to stimulate employers to invest more in preventive measures.

29

30 CHAPTER 1. KEY TRENDS: STRONG ECONOMIC GROWTH BUT INSUFFICIENT JOB CREATION 33 Chapter 1. Key Trends: Strong Economic Growth but Insufficient Job Creation Chile s unequal income distribution is linked with relatively low employment and a segmented labour market, two factors that also contribute to the difficulties in extending social protection to the whole population. Despite strong and sustained economic growth, job creation has been insufficient to give opportunities to under-represented groups and to cope with a growing working-age population. For women and youth, in particular, employment rates are low by OECD standards. The manufacturing sector s share in total employment declined in the 1990s, after which it has been mostly stagnant at a low level for a country at Chile s development stage. Informal employment is less widespread than elsewhere in Latin America, but still sizeable by OECD standards. The recent job creation has occurred partly in sectors that employ many high-skilled workers, but also in the areas where jobs tend to be low-productive and precarious.

31 34 CHAPTER 1. KEY TRENDS: STRONG ECONOMIC GROWTH BUT INSUFFICIENT JOB CREATION 1. A favourable macroeconomic environment Chile has sustained high rates of economic growth for over two decades, permitting it to gradually reduce its income gap relative to the OECD area. Despite a cyclical slowdown during , real GDP increased by nearly 6% per year over the period. Chile is now one of the richest countries in Latin America in terms of GDP per capita. Chile s strong performance over the past two decades has been in sharp contrast with the general Latin American trend until recently, characterised by slow growth overall but also by very volatile performance (Figure 1.1). Sound macroeconomic management over the years, and institutional reforms of monetary and fiscal policy around the turn of the century, made Chile more resilient to terms-oftrade shocks and the financial crisis that then affected the region. These are all good news, but there is no room for complacency. With GDP per capita at only 29% of the United States level in 2006 (as measured in purchasing power parities), Chile still has a long way to go in catching up with the more advanced OECD economies. % Figure 1.1. Chile s annual GDP growth compared with OECD and selected Latin American countries Chile Argentina, Brazil and Mexico OECD Note: The OECD growth line does not include Mexico and Turkey. Source: World Development Indicators. In recent years, the country s already strong public finances have been further nourished by the upswing in the price of copper, the country s main export (Figure 1.2). This has permitted a steady improvement in the trade balance rising from negative figures in the 1990s to a surplus of 13% of GDP in 2007 and, also, an upward trend in public social spending. However, as shown in Chapters 3 and 4, the latter remains relatively low by most internationals standards.

32 CHAPTER 1. KEY TRENDS: STRONG ECONOMIC GROWTH BUT INSUFFICIENT JOB CREATION 35 Figure 1.2. Chile s exports Percentage of GDP 45% 40% 35% 30% 25% 20% 15% 10% 5% Primary Manufacturing Other mining Copper 0% Source: Calculations based on data from Banco Central de Chile ( 2. Chile has large human resources that are not well utilised The estimated population size in 2008 is about 16.8 million, following an increase by one-third over the past two decades. The total fertility rate has been a little lower than the replacement rate of 2.1 children per woman since about 2000, but the working-age population will continue to grow for a considerable time because the biggest age cohorts are still no more than about 20 years old (Figure 1.3). Net immigration is positive but relatively modest, though it is possibly increasing as Chile is attracting workers from neighbouring countries. 2 Relative to the size of the working-age population, both employment and the effective labour supply is lower than in most OECD countries, especially for women and youths. Only for the year age group does Chile record an employment rate similar to those of most OECD countries (Tables 1.1 and 1.2). Many more jobs would have to be created not only to give more opportunities to women and youths, but also to accommodate demographic pressures. Projections suggest that the working-age population, around 11 million in 2008, will increase by over 1.2 million over the coming ten years, not counting international migration. 2. Reliable statistics about international migration do not exist. Estimates for the 1990s suggest a net immigration of around persons per year (0.1% of the population). See INE and CEPAL (2004).

33 36 CHAPTER 1. KEY TRENDS: STRONG ECONOMIC GROWTH BUT INSUFFICIENT JOB CREATION Figure 1.3. Population by age class Panel A. Millions of persons Source: Instituto Nacional de Estadisticas (INE). Panel B. Population by age and gender in 2005 and 2050 Percentage of total population in each group MEN WOMEN in 2005: 16.3 Total population (in millions) in 2050: 20.7 in 2005: 14 Old age dependency ratio (65+ in % 20-64) in 2050: 40 Source: Adapted from Population Division of the Department of Economic and Social Affairs of the United Nations Secretariat, World Population Prospects: The 2006 Revision and World Urbanization Prospects.

34 CHAPTER 1. KEY TRENDS: STRONG ECONOMIC GROWTH BUT INSUFFICIENT JOB CREATION 37 Table 1.1. Employment/population ratios by gender for selected working-age groups in OECD and accession countries in 2006 Percentages 1 Men Women Youth Older Iceland 89 Iceland 82 Iceland 73 Iceland 85 Switzerland 85 Denmark 73 Netherlands 64 New Zealand 70 New Zealand 82 Norway 72 Australia 64 Sweden 70 Mexico 82 Sweden 72 Denmark 64 Norway 67 Japan 81 Switzerland 71 Switzerland 63 Switzerland 66 Denmark 81 Canada 69 New Zealand 59 Japan 65 Australia 79 New Zealand 68 Canada 59 United States 62 Netherlands 79 Finland 67 United Kingdom 57 Denmark 61 Norway 79 United Kingdom 67 United States 54 Korea 59 United Kingdom 78 United States 66 Austria 54 Estonia 58 United States 78 Netherlands 66 Norway 53 United Kingdom 57 Ireland 77 Australia 66 Ireland 48 Australia 56 Spain 77 Estonia 65 Mexico 45 Canada 56 Austria 77 Austria 64 Sweden 44 Mexico 55 Canada 77 Portugal 62 Germany 44 Israel 55 Sweden 77 Slovenia 62 Spain 43 Finland 55 Greece 75 Germany 62 Japan 41 Ireland 53 Korea 75 Ireland 59 Finland 41 CHILE 53 Portugal 74 Japan 59 Portugal 36 Portugal 50 Czech Republic 74 Russian Federation 58 Slovenia 35 Germany 49 Luxembourg 73 France 57 Russian Federation 33 Russian Federation 47 Germany 73 Czech Republic 57 Estonia 31 Netherlands 47 CHILE 72 Spain 54 Turkey 31 Czech Republic 45 Slovenia 71 Luxembourg 54 Czech Republic 28 Spain 44 Finland 71 Belgium 54 Korea 27 Greece 42 Italy 71 Korea 53 Israel 27 France 41 Estonia 71 Slovak Republic 52 Belgium 26 Austria 36 Turkey 68 Hungary 51 CHILE 26 Hungary 34 France 68 Poland 48 Slovak Republic 26 Slovak Republic 33 Belgium 67 Greece 48 Italy 26 Italy 33 Slovak Republic 67 Italy 46 France 25 Slovenia 33 Russian Federation 66 Israel 46 Luxembourg 25 Luxembourg 32 Hungary 64 Mexico 43 Greece 25 Belgium 30 Poland 61 CHILE 39 Poland 24 Turkey 30 Israel 56 Turkey 24 Hungary 22 Poland Countries are ranked from highest to lowest employment rates in each column. Source: OECD Labour Force database.

35 38 CHAPTER 1. KEY TRENDS: STRONG ECONOMIC GROWTH BUT INSUFFICIENT JOB CREATION Table 1.2. Labour force status of the Chilean population, Labour force/population Employed/Population Unemployed/Labour force Both genders Men Women Source: Instituto Nacional de Estadisticas (INE) ; Labour Force database.

36 CHAPTER 1. KEY TRENDS: STRONG ECONOMIC GROWTH BUT INSUFFICIENT JOB CREATION 39 Women and youth face difficulties in entering the labour market Over the past 12 years, the labour force participation rate remained stable around 81% for men (Table 1.2), similar to the OECD average, while the corresponding rate for women increased from 31% to 39%. But the female unemployment rate has also increased significantly, and thus the gender difference has declined much less in employment rates than in labour force participation. Chile s female employment rate is also low compared with Argentina, Brazil, and Uruguay (World Bank, 2007a). Nevertheless, the trends suggest a considerable improvement for prime-age and older women since the mid For men, the same holds only for the year age group (Figure 1.4). The situation is particularly worrisome for large groups of young labour market entrants, and, despite some recent improvement, for prime age women. For prime-age men, the employment rate is rather stable at around 90%. It is also notable that 16.5% of people older than 65 are participating in the labour market, often reflecting insufficient pensions. Figure 1.4. Evolution of employment/population ratios by gender and age groups, Females Males Source: Instituto Nacional de Estadisticas (INE). The difficulties faced by the year age group are illustrated by the fact that only 5.6% of them combine education with work, while around 19% are for various reasons neither in employment nor in education or training (noted as NEET hereafter; Table 1.3). Although several OECD countries face a similar challenge, the latter ratio in Chile is almost twice the OECD average, with only Turkey and Mexico showing higher NEET rates (OECD, 2008a, Figure 1.1). This situation probably has several explanations, including both the structure of secondary and higher

37 40 CHAPTER 1. KEY TRENDS: STRONG ECONOMIC GROWTH BUT INSUFFICIENT JOB CREATION education, which is more easily combined with work in some countries than in others, and a significant incidence of teenage pregnancies. Table 1.3. Youths in 2006: unemployment, labour force participation and the proportion who neither study nor work Age, gender Unemployment rate (%) Participation rate (%) Neither study nor work (%) Men Women Men Women Men Women Source: Based on Mideplan (CASEN 2006). In Chile, the family conditions typically seen as obstacles to female labour force participation concern young women more than in most OECD countries. Chilean mothers gave birth to their first child at an average age of only 23.7 years in 2006, compared with around 30 years in many OECD countries. Teenage mothers have recently accounted for about 15% of all births, with substantially higher incidence in low-income groups and poor regions (INE, 2006). Some 26% of Chilean households consist of single parents with children, of which sole mothers represent 22 percentage points including one-third of the poor households and one-fifth of those not classified as poor (CASEN 2006). About half of the sole mothers are not in the labour force. For Chilean women in general, labour force participation is strongly correlated with education attainment and family income. In the year age group, women in the top income quintile had on average over five years longer education than those in the bottom quintile, while the labour force participation rate was 74% in the top quintile compared with only 36% in the bottom quintile in 2006 (Table 1.4). To some extent, this reflects a wage structure that makes it worthwhile for well-educated parents to work and pay for child care, but less so for those with little education. But in addition, low-income families have on average more than twice as many children as high-income families have (despite a lower average age reported for the heads of poor households). Poor households are also more likely to include persons not belonging to the

38 CHAPTER 1. KEY TRENDS: STRONG ECONOMIC GROWTH BUT INSUFFICIENT JOB CREATION 41 nuclear family, most often a grandparent a situation that involves additional care responsibilities for some working-age women, but which can also have positive effects, as shown by above-average rates of labour force participation and education enrolment for mothers living with their own parents (Larrañaga, 2009). Low-income women are over-represented in rural areas, reducing their access to jobs. (In the bottom quintile, 22% of women live in rural areas compared with 5% in the top quintile.) Table 1.4. Profile of women between years old by per capita income quintiles, 2006 Per capita autonomous income quintile I II III IV V Total Urban residence 78% 85% 89% 93% 95% 88% Age Years of education Family headed by women 36% 32% 29% 29% 29% 31% Children children less than 18 years old children less than 6 years old Hours worked Employees without formal contract 50% 34% 26% 22% 14% 26% Labour force participation 36% 46% 55% 64% 74% 55% Note: Employees without formal contract are counted as a percentage of all female employees (i.e. excluding employers, self-employed persons and unpaid family workers). Source: Based on Mideplan (CASEN 2006) and OECD calculations. Figure 1.5. Reasons for inactivity among women, 2006 Other reasons 17.9% Illness 7.5% Childcare and household activities 39.0% Retirement 15.1% Student 20.5% Source: Based on Mideplan (CASEN 2006) and OECD calculations.

39 42 CHAPTER 1. KEY TRENDS: STRONG ECONOMIC GROWTH BUT INSUFFICIENT JOB CREATION Asked about the reasons for not seeking jobs, women most often mention care of children or elderly relatives and household activities, but a range of other reasons also play a role (Figure 1.5). Among employed women, those in lower income quintiles are over-represented in the informal economy, particularly domestic services. Cultural factors can also be important, as suggested by a relatively high inter-correlation between spouses educational attainment Job creation in low-productivity sectors The recession of the late 1990s led to both absolute and relative declines in industrial employment, including mining, manufacturing and public utilities, as well as in agriculture. As measured in absolute numbers, this cyclical decrease was reversed after But in relative terms, the manufacturing and agricultural sectors continued to shrink until 2005 or 2006, when they stagnated at, respectively, 13% and 12% of total employment (Table 1.5). With currently almost two-thirds of total employment in the service sector, Chile s de-industrialisation process appears about as advanced as in many old industrialised economies even though the agricultural employment share is still as high as 12%, which is more comparable to Mexico, Portugal or Poland (Table 1.6). 4 Table 1.5. Employed persons by economic sector in Sector Percentage distribution of total employment Accumulated absolute change (%) Primary Industry Mining Manufacturing Utilities Construction Services Commerce and tourism Transports, communications Finance and business services Social and personal services Total Source : Instituto Nacional de Estadisticas (INE). 3. De Ferranti et al. (2004) find a correlation coefficient of 74.5%. A similarity in education levels within couples can be considered as a cause of social segmentation, contributing to the inequality of household incomes. 4. The World Bank (2008) reported similar shifts of employment from manufacturing to services since the early 1990s for most of the Latin American region, where manufacturing represented on average 15% of employment in To a large extent these employment shifts were directed towards relatively low-productive activities.

40 CHAPTER 1. KEY TRENDS: STRONG ECONOMIC GROWTH BUT INSUFFICIENT JOB CREATION 43 Table 1.6. Employment by economic sector in Chile and OECD countries, 2007 Percentage of total employment United Kingdom Sweden Australia Denmark Ireland Korea Italy Chile Mexico Portugal Poland (2005) Primary Industry Construction Services Countries are ranked by service sector shares. Source: Instituto Nacional de Estadisticas (INE) and OECD Labour Force database. The employment growth recorded in the service sector has partly concerned relatively high-skilled and well-paid jobs, for example in finance, whose growth has been facilitated by increased education attainment in the young population. Nevertheless, much of the observed labour flows appear to reflect the supply pressures created by population growth and a continuing, albeit slow, relative decline of agricultural employment. Significant employment shifts have thus been directed towards sectors with modest average wages, for example in commerce, tourism and construction. In these areas and in the manufacturing sector, relative wages have reportedly declined since the 1990s (Table 1.7). Table 1.7. Relative wages in selected economic sectors Overall average = 100. Sectors ranked by average wage in 2005 Mining Utilities Financial services Transport, comm. Manufacturing Commerce Construction Source : Instituto Nacional de Estadisticas (INE).

41 44 CHAPTER 1. KEY TRENDS: STRONG ECONOMIC GROWTH BUT INSUFFICIENT JOB CREATION 4. The quality of jobs Sub-contracting and worker dispatching Many Chilean firms sub-contract part of their production and services to other firms or use workers dispatched from temporary work agencies (TWAs). Available statistics do not show the number of workers concerned, but the 2006 ENCLA survey (Encuesta Laboral) indicated that 41% of the formal enterprises including over two-thirds of the big firms used sub-contractors and almost as many used some TWA workers. Both phenomena have often been found to concern tasks within client enterprises own work processes as well as more separate tasks. As discussed in Chapter 2, the legal and labour market implications are complex and have recently been subject to much controversy. Sub-contracting, in particular, has been considered to cause an excessive fragmentation of enterprises and to exacerbate the segmentation and opacity of Chile s job market. Available sector-level statistics about employment and wages, such as those in Tables 1.6 and 1.7, must therefore be treated with caution. The role of informal employment Informal work is less prevalent in Chile than elsewhere in Latin America, as shown by the number of individuals who contribute to social insurance systems (Figure 1.6). But it is still relatively widespread by the OECD standards. In many countries, informality is often related to a desire on the part of both employers and employees to avoid paying taxes and social contributions, or it may reflect other conditions in the business environment such as a high minimum wage, rigidities in labour law or administrative costs associated with formality. Tax enforcement may be less of a problem in Chile: OECD (2007a, Chapter 4) found that only about 11% of the potential value-added tax base appeared to have been undeclared in 2005, while undeclared work concerned a substantially higher percentage of the workforce. Analysis based on Chile s 2006 CASEN household survey suggests that much of the informal work is associated with ingrained patterns of labour market segmentation, being typically concentrated in small firms with low-skilled workers in low-paid jobs. Tax evasion also plays a role sometimes among high-skilled groups as do more transitory situations, especially among youths who may be waiting for better jobs.

42 CHAPTER 1. KEY TRENDS: STRONG ECONOMIC GROWTH BUT INSUFFICIENT JOB CREATION 45 Figure 1.6. Employed persons in Latin America not contributing to social insurance Percentage of employment in urban areas Self-employed Salaried Source: Adapted from World Bank (2007b). As an alternative measure of informality, the proportion of employees without labour contracts increased in the 1990s, but began to decline after This proportion has tended to converge with the share of employees not contributing to pension insurance, the two situations being often combined; both were around 20% of the employees in 2006, corresponding to about 15% of total working-age employment in 2006 (Box 1.2). 5 The share of self-employment (including employers and those without employees) in total employment has been relatively stable since the early 1990s, being 22% in 2006 according to the CASEN surveys, or some percentage points more according to labour force surveys. About three-fifths of them were own-account workers without employees, amongst whom barely 20% contributed to pension insurance compared with 34% for those who were employers. Counting all forms of employment, it can thus be estimated that a little more than 30% of the employed working-age workforce did not contribute regularly in Some 80% of those without contract did not contribute to social insurance, and vice versa. In addition, AFP surveys indicate that at least 5% of the contributing employees paid too little.

43 46 CHAPTER 1. KEY TRENDS: STRONG ECONOMIC GROWTH BUT INSUFFICIENT JOB CREATION Box 1.2. Alternative measures of labour informality in Chile As discussed in the OECD Economic Survey of Chile (OECD, 2007a), labour informality is a broad concept that is difficult to measure. There is no consensus in the academic literature or in the policy world over what constitutes the informal sector, even among OECD countries. The table below documents three possible definitions for employees: 1) employees without signed contract; 2) employees not contributing to social security; and 3) employees without contract who do not contribute to social security, along with 4) self-employed and 5) self-employed persons who do not contribute to social security. 1 The first three ratios are calculated over all employees and over total working-age employment; the last two only over the latter denominator. The different rates for the first three definitions suggest a non-negligible number of employees without labour contracts who are still making social security contributions. Women are over-represented amount informal employees, but under-represented among the self-employed. One reason for the higher informality rate for female is that about half of the female domestic servants and unpaid family workers have no signed contracts. Incidence of informality by alternative measures Employees without contract Employees not contributing to social security as % of all employees Employees without contract and not contributing to social Men 19.1% 16.4% 14.0% Women 25.9% 23.3% 20.5% M+W 21.8% 19.1% 16.5% as % of total employment Selfemployed persons in general Selfemployed persons not contributing to social insurance Men 14.4% 12.3% 10.5% 23.0% 16.7% Women 19.9% 17.9% 15.7% 20.6% 16.3% M+W 16.6% 14.5% 12.6% 22.1% 16.6% Note: The employees counted here include those in the private and public sectors, household servants and the military. Self-employed groups include employers and those counted as own-account workers (irrespective of the number of their employees). Source: Based on Mideplan (CASEN 2006) and OECD calculations. 1. These figures refer to employees in the private and public sectors including domestic work and the military. Self-employment as defined here covers both employers and those without employees (own-account workers). It should be noted that social security is voluntary for the self-employed, while a few employees have contracts that do not refer to the Labour Law (called honorarios), allowing the employer to decide whether or not to contribute to social security. Note also that any individual can make contributions to the health institutions or the pension fund, and the self-employed can choose varying payment intervals. In any given month, the proportion reported by the pension administration to contribute is therefore lower than the survey-based figures used here (i.e. they suggest a higher informality rate).

44 CHAPTER 1. KEY TRENDS: STRONG ECONOMIC GROWTH BUT INSUFFICIENT JOB CREATION 47 Youths of both genders are disproportionally present among informal employees (Figure 1.7), especially when they have low education, while persons aged 55 or more are over-represented among the self-employed (not shown). Women are over-represented among informal employees (see Table in Box 1.2 and Table 1.8), but this gender difference almost disappears if one does not count domestic work (Figure 1.7). If domestic work is excluded, the wage premium for a formal contract is only about 20% for women compared with 30 to 40% for men; in other words, the male-female wage gap appears smaller in informal than in formal jobs. Table 1.8. Activity rate, informality and earnings in the population by education, 2006 Gender Years of education Labour force participation rate Lack of contract: % of employees Self-employment: % of the employed Contract Job earnings No contract Self-employed Men 94% 20% 25% Less than 8 88% 30% 36% % 25% 27% % 15% 19% >12 93% 13% 21% Women 59% 26% 22% Less than 8 40% 48% 34% % 39% 26% % 23% 19% >12 77% 15% 15% Note: The figures for employees in this table include the military and domestic services. Job earnings = average hourly income in pesos from the main job for people working at least 30 hours a week. Source: Based on Mideplan (CASEN 2006) and OECD calculations. Informality of any type is most common in micro-firms, including persons working alone and those with a few employees and/or family workers. Almost 40% of employees in firms with up to five workers have no signed contract, compared with 7% in firms with over 200 workers. The linkage with firm size is similar in most economic sectors, resulting in the highest overall incidence of informality in agriculture (where 40% are self-employed and 31% of the employees lack contracts) followed by social services, commerce and manufacturing. While informality increased in the 1990s it shows no clear pattern with respect to business cycles. Regional variations are also moderate, apart from those that depend on the size of the agricultural sector. All told, a large part of the observed informal employment represents a fragile segment of the labour market. Low-skilled employees are often hired

45 48 CHAPTER 1. KEY TRENDS: STRONG ECONOMIC GROWTH BUT INSUFFICIENT JOB CREATION informally by small firms that are themselves informal. For the workers concerned, often young, informal employment is probably a second-best solution in many cases. However, self-selection probably plays a role as well. There is some evidence that white-collar workers who work informally often gain from this situation e.g. earning more take-home pay with the same gross income while the predominant low-skilled groups typically earn less in both net and gross terms when they work informally (Puentes, 2007). Furthermore, panel surveys show that employees without contracts often expect their informal status to be temporary, an expectation largely confirmed by data about transitions to formal jobs. 6 Figure 1.7. Informality among employees by age age age age Females age age age Males Note: The data used here do not cover the military and domestic workers. Female informality rates appear significantly higher when the about domestic workers are included, as shown in the table in Box 1.2 and in Table Persons without signed labour contract. Source: Calculations based on Mideplan, various years. 6. Over half of the employees who lacked contracts in one year had contracts two years later. About one-third of the self-employed made the transition to formal employee jobs over the same periods. Conducted in 2002, 2004 and 2006, these surveys referred to the about 5.6 million persons affiliated to pension insurance thus not all the employed (over 6 million), but many more than the monthly contributors (about 3 million).

46 CHAPTER 1. KEY TRENDS: STRONG ECONOMIC GROWTH BUT INSUFFICIENT JOB CREATION Significant improvements in health, education and housing Traditionally, Chile s social agenda has mainly focused on providing access to services, especially in the areas of education, health care and housing, and this with quite some success. The share of the population without health insurance came down from 12.2% in 1990 to 5.1% in 2006, education enrolment has increased substantially and the previous housing deficit was much reduced. Despite such achievements, there is still a need to improve the quality of available provisions and to reduce access barriers for disadvantaged groups. As measured by conventional health indicators such as life expectancy and infant mortality, Chile is probably the best performer in Latin America and it also compares relatively well with some OECD countries (Figure 1.8). Further efforts are required in education because the overall stock of human capital remains relatively low. The reported illiteracy rate for adults decreased from 5.7% in 1992 to 3.6% in 2006, 7 but Chile s population aged years still had only ten years of schooling on average, compared with an OECD average of 12 years. Enrolment is nearly complete in primary and lower-secondary education, while enrolment rates at upper-secondary and higher levels have increased dramatically since 1990, especially for youths from low-income households (Figure 1.9, Panel A). Income-related differences in higher-education enrolment can depend partly on students performance at lower levels, but they must also be attributed to a difference in families financial capacity and its impact on the quality of education received at lower levels. PISA scores for 15-year-olds reveal that Chile s student performance remains relatively far below the OECD average (Panel B). In contrast with the general experience in OECD countries, the performance differs too much between rather than within the public, subsidised private and fully private sectors of the education system. Based on the 2007 SIMCE national test for language and mathematics performance, fully private schools covering 6% of the students, mainly from the top income quintile outperform the subsidised private schools that cover 46% of the students, mainly from middle-income groups. The subsidised private schools in turn outperform municipal (fully public) schools, which educate 47% of the students who come mainly from the lowest quintiles (Panel C). 7. Source: UNESCO Institute for Statistics.

47 50 CHAPTER 1. KEY TRENDS: STRONG ECONOMIC GROWTH BUT INSUFFICIENT JOB CREATION Figure 1.8. Health indicators Panel A. Life expectancy at birth (years) Panel B. Child mortality Child mortality is defined as the probability of dying by the age of 5 per live births. Source: World Health Organisation and World Bank.

48 CHAPTER 1. KEY TRENDS: STRONG ECONOMIC GROWTH BUT INSUFFICIENT JOB CREATION 51 Figure 1.9. Education outcomes Panel A. School enrolment rates by income level, Income quintile Pre-primary education Primary and lowersecondary education Upper-secondary education Higher education I II III IV V Total Source: Based on Mideplan (CASEN 1990 and 2006). Panel B. Student performance: average PISA score for science in OECD members and accession countries, Horizontal line indicates the OECD average. Source: OECD PISA Panel C. Education performance by school type: SIMCE scores for secondary education, Language Mathematics Municipal Private subsidised Fully private Municipal Private subsidised Fully private Primary and lower secondary education Upper secondary education The scores of language vs. mathematics cannot be compared because they are rescaled each time to yield a national mean of 250 and a standard deviation of 50. Source: Adapted from the Spanish version, Ministry of Education (SIMCE database).

49 52 CHAPTER 1. KEY TRENDS: STRONG ECONOMIC GROWTH BUT INSUFFICIENT JOB CREATION There is also still room for improvement in the housing area, although a range of housing subsidies introduced in the 1990s has led to a sharp quantitative and qualitative improvement of the housing stock. An officially estimated housing deficit defined as the percentage of households without their own dwellings or whose dwellings did not fulfil certain minimum requirements declined from 30% in 1990 to under 10% in 2006 (Table 1.9). The proportion of owner-occupied dwellings increased from 61% to 69% over the same period (Figure 1.10). Strikingly, the increase in home ownership mainly concerned low-income groups, a fact that appears to reflect the limited market for rental accommodation apart from relatively expensive homes. 8 The share of owner-occupied dwellings acquired with the help of public subsidies has recently been over one-third from the first four quintiles and one-sixth in the highest quintile. Housing deficit 1 Table 1.9. Housing conditions in 1990 and 2006 Percentage of households concerned Dwelling shared with unrelated households Dwelling shared with extended family 2.5 or more persons per bedroom Poor building materials 2 Lack of WC Households who lack dwellings of their own or have dwellings of unacceptable quality. 2. Perishable, not water-tight, etc. 3. Lack of WC connected to sewer or sceptic tank. Source: Based on Mideplan (CASEN 1990 and 2006). Figure Households owning their main dwellings % I II III IV V Income quintile Source: Based on Mideplan (CASEN 1990 and 2006). 8. Rental housing with contracts is common mainly in the highest income quintiles (CASEN 2006). Low-income households not owning their homes more often rent informally or live in borrowed dwellings.

50 CHAPTER 1. KEY TRENDS: STRONG ECONOMIC GROWTH BUT INSUFFICIENT JOB CREATION Poverty has been much reduced, but the income distribution remains very unequal Years of sustained growth in Chile have resulted in a substantial decline in the absolute poverty headcount ratios (as defined nationally) between 1990 and 2006 (Table 1.10, Panel A). 9 Poverty went down from 38.6% to 13.7% of the total population, which corresponds to around 2.76 million persons moving out of poverty. The drop in the incidence of poverty was particularly sharp during The percentage of people living in extreme poverty (indigence) fell by 9.8 percentage points to 3.2% in Although children and youth have benefited the most from this reduction (Table 1.10, Panel B), their poverty rate remains double that observed for other age groups. Persons aged 45 or more show the lowest levels of poverty. However, these two last-mentioned observations may reflect the use in Chile of a household income concept per capita instead of equivalent units: large households look poorer when all household members are given the same weights. Notwithstanding Chile s economic success, the country faces a difficult problem of inequality. Table 1.11 (left column) ranks countries according to the best known index of income inequality, the Gini coefficient. 10 With a Gini coefficient of 0.53, the degree of inequality in Chile is far above that of OECD member countries including those with the highest income inequality, Turkey and Mexico. Chile also shows high relative poverty, with around 16.5% of the population having an equivalised disposable income of under 50% of the median for the entire population in the mid-2000s more than in all OECD countries except Mexico, Turkey and the United States (right panel). Although this measure allows a benchmarking of countries performance, this is done on an essentially arbitrary basis; it is less informative about the minimum standard of living that is considered acceptable in a country (Box 1.3). 9. These poverty headcount ratios are the shares of the population living below the national poverty and indigence lines, which correspond to certain consumption baskets. 10. The Gini coefficient is defined as the area between the Lorenz curve (which plots cumulative shares of the population, from the poorest to the richest, against the cumulative share of income that they receive) and the 45 line, taken as a ratio of the whole triangle. The values of the Gini coefficient range between 0, for perfect equality (i.e. each share of the population gets the same share of income), and 1, for perfect inequality (i.e. all income goes to the individual with the highest income).

51 54 CHAPTER 1. KEY TRENDS: STRONG ECONOMIC GROWTH BUT INSUFFICIENT JOB CREATION Table Poverty trends Panel A. Evolution of poverty and indigence, Indigence Not indigent poverty Total poverty Persons % Persons % Persons % Panel B. Percentage of population in poverty by age, 1990 and 2006 With respect to the total number of persons in the relevant age group Age group % 22% % 20% % 11% % 14% % 9% 60 and older 21% 8% Total 38.6% 13.7% Note: In-house domestic servants and their families are excluded. The official income concept provided by Mideplan is used (household market income per capita). Source: Adapted from the Spanish version of Mideplan (2007a).

52 CHAPTER 1. KEY TRENDS: STRONG ECONOMIC GROWTH BUT INSUFFICIENT JOB CREATION 55 Table Poverty and inequality indicators, mid-2000s 1 Gini coefficient of income inequality Relative poverty rates at 50% of median income 2 Denmark Denmark Sweden Sweden Luxembourg Czech Republic Austria Austria Czech Republic Norway Slovakia France Finland Iceland Belgium Hungary Netherlands Finland Switzerland Netherlands Norway Luxembourg Iceland Slovakia France United Kingdom Hungary Switzerland Germany Belgium Australia New Zealand Korea Germany Canada Italy Spain Canada Japan Australia Greece Greece Ireland Portugal New Zealand Spain United Kingdom Poland Italy Ireland Poland Korea United States Japan Portugal CHILE Turkey United States Mexico Turkey CHILE Mexico OECD average OECD average Countries are ranked in increasing order of the respective indicator. 1. The income concept used is that of disposable household income in cash, adjusted for household size. 2. Poverty rates are defined as the share of individuals with equivalised disposable income less than 50% of the median for the entire population. Source: OECD Income Distribution questionnaire, and Larrañaga (2009).

53 56 CHAPTER 1. KEY TRENDS: STRONG ECONOMIC GROWTH BUT INSUFFICIENT JOB CREATION Box 1.3. Poverty line in Chile There is no standardised procedure for measuring poverty, even among OECD countries. Often, a relative income measure is used. For example, a threshold of 60% of median income is used as a benchmark for at-risk-of-poverty at the EU level, while the poverty line used in the United States is closer to 40% of median income. Poverty measurement also allows for variations in absolute criteria, such as the poverty line of USD 1 or USD 2 per capita per day used by the World Bank. The procedure chosen depends on the interpretation of the minimum needs, which can be considered absolute or relative with respect to the standard of living in the country. In Chile, as in most Latin American countries, the national poverty line is measured on the basis of a basic food basket. This is the cost-efficient food basket that satisfies the nutritional requirements of household members, and which defines the indigence or line of extreme poverty. After adding up the cost of other commodities, such as clothing, transport, and housing, the poverty line is obtained. The calculation is based on the 1987 consumption spending of a reference group of households (adjusted for inflation). Table A shows the official poverty and indigence lines for the period Table A. Official poverty and indigence lines, In Chilean pesos of November each year Poverty line Urban zone Rural zone Indigence line Urban zone Rural zone Note: The official income concept provided by Mideplan is used (household market income per capita). Source: Adapted from the Spanish version of Mideplan (2007a). Poverty lines are a key factor in the allocation of social spending and hence deserve critical assessment. Larrañaga (2009) calculates an alternative poverty line based on the OECD income concept and estimates the impact of various poverty lines on the incidence of poverty through microsimulations (cf. Table B). 1 These estimates show that the incidence of poverty is quite sensitive with respect to the value of the poverty lines. The reference line is Chilean pesos, which is comparable to the level actually used in the country (based on the Chilean official income concept). An increase of 25% in the reference value of the poverty line causes an increase of more than 50% in the incidence of poverty in 2006, whereas a decrease of 25% in the poverty line almost halves the percentage of poor. Since there are many households whose income hovers around the poverty line, changes in the line have a large impact on the poverty rate. This is also reflected in the changes in poverty rate over time across the different poverty lines. Although poverty clearly declined over time (independently of the poverty line used), the percentage change decreases when the value of the line increases. This is because higher poverty lines are associated with a lower density in the income distribution function. 1. See Annex 1.A1 for a detailed discussion of the methodological issues on income measurement.

54 CHAPTER 1. KEY TRENDS: STRONG ECONOMIC GROWTH BUT INSUFFICIENT JOB CREATION 57 Table B. Percentage of poverty, for alternative poverty lines Household disposable income per equivalent adult Poverty lines in Chilean pesos of November Source: Larrañaga (2009). 7. Income inequality in Chile is closely linked with wage inequality The distribution of household market income remained virtually unchanged between 1990 and 2003 (Table 1.12). Household market income covers the income generated with the household s own means and includes labour income, capital income (rents and interest), benefit payments from contributory pensions, and other private income (such as transfers from relatives not living in the household and donations). In 2006, there was for the first time a visible improvement in the income distribution. Deciles 2 to 9 increased their share in household market income at the expense of the 10 th income decile, whose share of income declined from 41.5% in 2003 to 38.6% in Table Distribution of household market income By decile of household market income per capita, Decile I II III IV V VI VII VIII IX X Total Note: The official income concept provided by Mideplan is used (household market income per capita). Source: Adapted from the Spanish version of Mideplan (2007b).

55 58 CHAPTER 1. KEY TRENDS: STRONG ECONOMIC GROWTH BUT INSUFFICIENT JOB CREATION The evolution of the income distribution can also be analysed through the Lorenz curve, which plots the cumulative share of income received by people in each decile of the distribution. A distribution is less unequal than another if one Lorenz curve lies above the other (or closer to the 45 degree or equal distribution line). Figure 1.11 shows the distance between the Lorenz curve of the year 2006 and those of 1990, 1996 and In each case, the distance is positive over the entire income distribution, indicating a strictly lower level of income inequality in 2006 compared with previous years. The combination of such Lorenz dominance and higher average income tells us that the country is both richer in 2006 compared with 1990 and 1996 and less unequal in terms of income. Figure Change in the Lorenz curve: a measure of declining income inequality Distance Distance Distance Note: The OECD income concept is used (household disposable income per equivalence unit). Source: Larrañaga (2009). The largest component of household market income is wages, representing around 80% of the household s market income across all income deciles (Mideplan, 2008). 11 In general, the wage distribution is more unequal than other components of household market income (Table 1.13). The ratio between the wage incomes of the 10% richest households and the 10% poorest households is 35 to 1, whereas the same ratio is only half as big for the rest of the household market income. With little change in the wage distribution over the past 16 years, labour income continues to be the major determinant of the income distribution. 11. Household surveys generally do not capture capital income well, nor the income of the wealthiest sectors of the population.

56 CHAPTER 1. KEY TRENDS: STRONG ECONOMIC GROWTH BUT INSUFFICIENT JOB CREATION 59 Table Distribution indices for wages and other household market income, 2006 Ratio decile 10/ decile 1 Ratio quintile 5/ quintile 1 Ratio decile 10/ deciles 1+2 Labour income Other autonomous income Total autonomous income Note: The official income concept provided by Mideplan is used (household market income per capita). Total market income is divided into Labour income and Other market income. The latter includes capital income (rents and interests), payments from contributory pensions, and other private income (such as transfers from family members and donations). Source: Adapted from the Spanish version of Mideplan (2007b). 8. Regional diversity Income and earnings discrepancies also have a clear regional dimension. Chile s population is highly concentrated in a few big cities and regions, with 40% of the population living in the main metropolitan area, Santiago. Different types of economic activity are often concentrated by region, e.g. mining in the north, traditional agriculture and agro-industry in the centre, and forestry and fishing in the south. There are also significant regional differences in living standards, with the highest poverty rates recorded in some of the regions south of the capital (Maule, Bío Bío, Araucanía, and Los Rios), but not in the less-populated extreme south (Table 1.14). High poverty rates are often associated with low employment rates, which, in turn, tend to be low in regions marked by a high incidence of employees without labour contracts. Interactions between regions can be expected to increase in a climate of economic growth. However, OECD (2005) found that the speed of convergence between Chilean regions had been slower than expected in view of the long period of strong development. Geographic distances evidently play a role, as evidenced by the above-average performance of the Santiago area, which offers a large potential for job mobility and a richer choice of jobs than other regions. Significant investments in transport infrastructure are underway that will contribute to the extension of local labour markets in and around Santiago and other larger cities, and there is undoubtedly a potential for further improvement in this direction.

57 60 CHAPTER 1. KEY TRENDS: STRONG ECONOMIC GROWTH BUT INSUFFICIENT JOB CREATION Table Regional differences, 2006 Population size (15+) Poverty rate 1 Employed/ Population Informality rate 2 Job earnings 3 Arica Parinacota % 52.5% 21% Tarapacá % 49.1% 18% Antofagasta % 55.9% 16% Atacama % 54.8% 15% Coquimbo % 48.0% 26% Valparaíso % 51.9% 22% Libertador % 39.9% 18% Maule % 51.2% 24% Bío Bío % 45.6% 21% Araucanía % 48.6% 21% Los Rios % 47.7% 23% Los Lagos % 53.8% 19% Aysén % 58.5% 14% Magallanes % 55.1% 14% Metropolitana % 57.0% 19% Total % 52.2% 20% The poverty rate is the percentage of households with an income below the official poverty line. The official income concept provided by Mideplan is used (household market income per capita). 2. Informality rate is the percentage of employees without a signed contract. 3. Job earnings represent the average monthly income from the worker s main occupation in Chilean pesos. Source: Based on Mideplan (CASEN 2006). 9. Conclusions Chile s income distribution is among the most unequal in the world. This must largely be attributed to a low level of employment and to a segmented labour market. Each of these two factors are also important causes of the difficulties faced in extending effective social protection to the whole population. Chile needs to create more and better jobs. Many women and youths, in particular, are suffering from the shortage of adequate employment opportunities. The downsizing of the manufacturing sector has not been matched by a corresponding job creation in other relatively high-productive sectors. This outcome may depend partly on geographic factors with long distances within the country and between it and potential markets but it must also be seen in the context of the general business climate and the labour market. OECD experience suggests that labour market reforms can contribute to substantially better outcomes.

58 CHAPTER 1. KEY TRENDS: STRONG ECONOMIC GROWTH BUT INSUFFICIENT JOB CREATION 61 Annex 1.A1. Measurement issues on income distribution Statistics on the distribution of household income and poverty can only be compared across countries if similar measurement methodologies are used. To facilitate comparison among its member countries, the OECD works with a standard income concept. The data are collected through a network of national experts, who apply common conventions and definitions to collect data from different national data sources and supply detailed cross-tabulations to the OECD. The OECD basic income concept has three key features: 12 It is based on household disposable income net of taxes and in cash (i.e. excluding items such home production or imputed rents); It refers to the distribution among individuals living in private households, where the income of the household is attributed to each of its members, irrespective of who in the household receives that income; and Household income is adjusted to reflect differences in household needs with a common but arbitrary equivalence elasticity (i.e. the square root of household size), implying that a household s economic needs increase less than proportionally with its size. The official statistics provided by the Chilean Ministry of Planning (Mideplan) are collected through the National Socioeconomic Characterisation Survey (CASEN). 13 It is a multi-topic household survey with a sample size of households in the year 2006, and has been implemented every two or three years since The CASEN survey is the traditional source for statistics on income distribution, poverty and the incidence of social spending in Chile. But CASEN s income concept differs in three aspects from the income concept used by the OECD: 1) it is expressed in per capita terms instead of equivalent units; 2) it includes imputed rental for own-housing; and 3) it excludes contributions made to private social security funds (Larrañaga, 2009). In Table 1.A.1, the Chilean 12. For an overview of the income distribution statistics and a detailed description of the underlying methodological issues, see OECD (2008b). 13. Encuesta de Caracterizacíon Socioeconómica Nacional.

59 62 CHAPTER 1. KEY TRENDS: STRONG ECONOMIC GROWTH BUT INSUFFICIENT JOB CREATION official statistics on inequality and poverty used by Mideplan are compared with the indicators following the OECD guidelines, as constructed by Larrañaga (2009). Although the statistics are not strictly comparable, the income distributions show a similar trend. Table 1.A.1. Comparing data on income distribution Ratio decile 10/decile 1 Gini coefficient Larrañaga Mideplan Larrañaga Mideplan Note: The Gini coefficient is defined as the area between the Lorenz curve (which plots cumulative shares of the population, from the poorest to the richest, against the cumulative share of income that they receive) and the 45 line, taken as a ratio of the whole triangle. Source: Larrañaga (2009) and Mideplan (2007b). It is also worth mentioning that the income data of the CASEN survey are corrected for under-reporting. The methodology used consists of multiplying each of the income components (with the exception of income from capital) by a constant factor that adjusts the data to the level of income reported in the national accounts. The adjustment factor is constant for all recipients of the respective income component, assuming that under-reporting is constant throughout the distribution of the variable. The only exception is property income, for which the entire difference between reported and actual income is imputed to households from the two richest deciles.

60 CHAPTER 1. KEY TRENDS: STRONG ECONOMIC GROWTH BUT INSUFFICIENT JOB CREATION 63 Bibliography De Ferranti, D., G. Perry, F. Ferreira and M. Walton (2004), Inequality in Latin America: Breaking with History?, World Bank, Washington DC. Gasparini, L. and L. Tornarolli (2007), Labour Informality in Latin America and the Caribean: Patterns and Trends from Household Survey Microdata, Working Paper, No. 0, CEDLAS, Universidad Nacional de La Plata, La Plata, Argentina. INE Instituto Nacional de Estadísticas (2006), Fecundidad en Chile: Situación reciente, Santiago. INE (Instituto Nacional de Estadísticas) and CEPAL (Comisión Económica para América Latina y el Caribe) (2004), Chile: Proyecciones y Estimaciones de Población. Total País, Santiago. Larrañaga, O. (2009), Inequality Poverty and Social Policy, Recent Trends in Chile, OECD Social, Employment and Migration Working Papers, OECD, Paris. Mideplan (2007a), Serie Análisis de Resultados de la Encuesta de Caracterización Socioeconómica Nacional (CASEN 2006), No. 1, La Situación de Pobreza en Chile, Ministerio de Planificación, Santiago, June. Mideplan (2007b), Serie Análisis de Resultados de la Encuesta de Caracterización Socioeconómica Nacional (CASEN 2006), No. 2, Distribución del Ingreso e Impacto Distributivo del Gasto Social 2006, Ministerio de Planificación, Santiago, June. Mideplan (2008), Políticas Sociales, Documento para la OECD, Santiago, July. OECD (2003), OECD Economic Surveys: Chile, Paris. OECD (2006, 2007b, 2008a), OECD Employment Outlook, Paris. OECD (2007a), OECD Economic Surveys: Chile, Paris. OECD (2008a), OECD Employment Outlook, Chapter 1, Paris.

61 64 CHAPTER 1. KEY TRENDS: STRONG ECONOMIC GROWTH BUT INSUFFICIENT JOB CREATION OECD (2008b), Growing Unequal, Paris. Puentes, E. (2007), Occupational Choice and Wage Distribution in a General Equilibrium Model. Evidence from Chile: , Mimeo, University of Chicago. World Bank (2007a), Expanding Women s Work Choices to Enhance Chile s Economic Potential, Chile Country Gender Assessment, Washington DC. World Bank (2007b), Informality: Exit and Exclusion, World Bank Latin American and Caribbean Studies, Washington DC. World Bank (2008), Job Creation in Latin America and the Caribbean: Recent Trends and the Policy Challenges, Washington DC.

62 CHAPTER 2. TOWARDS MORE EQUAL JOB OPPORTUNITIES 65 Chapter 2. Towards More Equal Job Opportunities To create more and better jobs, Chile needs to pursue its recent efforts to develop a strategy for work and equity. While partly successful, this policy endeavour is facing many obstacles, including not only limited resources and administrative capacity but also a weak social dialogue and many conflicting interests in a segmented labour market. The OECD Reassessed Jobs Strategy and the international experience of policy compromises for flexicurity are therefore relevant. This chapter considers the principal issues of labour market policy in Chile from such a perspective.

63 66 CHAPTER 2. TOWARDS MORE EQUAL JOB OPPORTUNITIES 1. Introduction Chile s growing human resources are not as efficiently used as they could be. Economic growth has been driven largely by improved productivity in capital-intensive sectors, especially mining, while job creation has been too limited and occurred partly in relatively low-productive service sectors, contributing to the continued inequality of opportunities. The twin problems of under-employment and inequality have recently been much discussed in Chile and several options for labour market reform have been considered, but the challenges are complex and several specific issues have proved controversial. In an effort to reconcile flexibility for employers with income security for workers, a Presidential Advisory Commission for Work and Equity presented a package of possible reforms in April 2008 (see Annex 2.A1). But despite extensive analysis and consultations, many of this Commission s suggestions had been found too contentious to form the basis of unanimous recommendations, being complicated by conflicting interests of employers, trade unions and other actors in a fragmented labour market. In this situation, there is a particular need to explore the possibility of a broad policy package of the flexicurity type that can satisfy the objectives of different labour market groups. The OECD Jobs Strategy and the related international debate about flexicurity is therefore relevant to Chile, although many detailed implications will be different in a developing country compared with more advanced economies. First presented in 1994, the OECD Jobs Strategy was reassessed in 2006 with the help of new empirical research. Econometric results, covering all member countries in the period since the early 1990s, leave no doubt that both macroeconomic policy, structural reforms in the labour market and policies to strengthen competition in product markets have a strong bearing on employment, productivity, job creation and other indicators of labour market performance. The principal issues of relevance in the labour market area are also important in Chile, as developed in the following. Both demand and supply-side measures need to be considered. To stimulate labour demand, public policy must first ensure macroeconomic stability and a business climate that is attractive to enterprise and job creation on market conditions. Many specific labour market institutions and their functioning in practice are important in this regard. Chile s labour supplies have the potential to grow substantially with increasing demand, as shown in the previous chapter. A supply-side response should be facilitated by policies to make the labour market flexible and transparent and to reduce obstacles to labour force participation, especially for women and youths. To

64 CHAPTER 2. TOWARDS MORE EQUAL JOB OPPORTUNITIES 67 enhance the quality of jobs and the equality of opportunities, the existing segmentation of the labour market should be reduced as far as possible. OECD experience shows that equity goals can and should be compatible with economic efficiency, which is best served by a well-integrated labour market where everybody can compete on equal terms. It is useful to consider the goals of more jobs, better jobs and equal opportunities together because they depend partly on the same set of institutional arrangements in the labour market. In the following, the principal instruments of labour market policy are discussed in turn from such a perspective, beginning with labour legislation, industrial relations, labour taxation and questions about informal employment. Subsequent sections are devoted to the Labour Inspectorate, unemployment insurance, employment services, job-related training and some special issues motivated by the low employment of women. 2. Important issues of labour legislation have yet to be resolved After the end of military rule, when many labour rights had been suppressed, new labour regulations were adopted that correspond to OECD standards in most respects. But the necessary social dialogue about such reforms has not been very effective, a fact that continues to make their implementation more difficult than it should be. The OECD (2003) observed that a series of Labour Law amendments from 2001, which sought to strengthen enforcement and gave additional rights to trade unions, had been met with much hostility from employers. Many of them had then responded by reducing their recruitments, choosing instead to rely increasingly on sub-contractors a phenomenon that has later become subject to legal regulation as well. In many respects, the application of the Labour Law has continued to be a source of discord in the social dialogue, for example concerning employment protection legislation, the allocation of working time and rules about sub-contracting. Employment protection legislation (EPL) Chile s EPL is not extraordinarily rigid by OECD standards. As measured by the OECD s scoring method it is more flexible than in many member countries, notably in Continental Europe (Figure 2.1 and Annex 2.A2). However, the rigidities that exist can have potentially more severe effects in Chile for two reasons. First, as a middle-income country with a relatively young population, Chile needs more flexible labour markets in order to sustain high growth and strong job creation. Second, Chile s EPL imposes relatively big differences between different types of labour contract, contributing to a market segmentation that exacerbates an already-high level of inequality in incomes and opportunities.

65 68 CHAPTER 2. TOWARDS MORE EQUAL JOB OPPORTUNITIES Figure 2.1. Measuring employment protection legislation (EPL) Average scores 0-6 from lowest to highest strictness Indefinite-duration jobs (collective or individual) Regulation of temporary employment Source: OECD (2004), Employment Outlook, Chapter 2, Annex 2.A; Chile s Labour Law as applicable in Just as in many European countries, the Labour Law in Chile establishes an element of discrimination, favouring insiders with indefinite-duration contracts at the expense of outsiders with more precarious jobs. Empirical results in OECD countries have shown that this reduces labour turnover in indefinite-duration jobs, which makes the labour market less dynamic and contributes to the difficulties faced by many youths and women who would like to have stable employment. In Chile, as noted below, the nature of a labour contract is also important for the right to unemployment compensation. Employers in Chile can usually dismiss indefinite-duration workers for economic reasons ( needs of the enterprise ), a criterion that is widely used by employers. If a dismissed worker has over one year s tenure, the employer must pay one monthly wage for every year of service as a severance benefit, up to a ceiling at eleven monthly wages. Dismissal procedures are relatively simple, and there are no additional procedures for collective dismissal of many workers. But it is not possible to dismiss individuals for lack of skills. As a result, an enterprise facing major business-related problems can dismiss many workers relatively easily; but if there is no indisputable economic justification, an employer may find it difficult to lay off just one or a few workers. A fixed-term contract can only last one year, but it is possible to accumulate two contracts and there are no

66 CHAPTER 2. TOWARDS MORE EQUAL JOB OPPORTUNITIES 69 restrictions on the types of work concerned. Contracts can also be signed for specific tasks and services and for seasonal work. The relatively high severance pay has been much criticised by employers, but proposals to reduce it have been rejected until now. As explained further below, employers who pay severance benefits can deduct their previous contributions to the worker s individual accounts under the unemployment insurance (UI). However, these deductions account for a modest part of the total severance pay in typical cases, implying that the employer s direct cost of dismissal for economic reasons is still high, and so a source of rigidity and distortions. The severance-pay system can also be regarded as unfair because it concerns only workers with indefinite-duration contracts and 12 months tenure who are dismissed for needs of the enterprise a group that has been roughly estimated to represent a little more than 6% of all formal and informal employees who become unemployed. Even in this small group, a significant proportion probably receive no severance pay or less than they are entitled to, as employers often persuade workers to accept a reduced amount or simply refuse paying. 14 Workers with average or higher incomes are over-represented among the recipients. 15 Severance pay represents a potentially high direct cost of dismissal for employers, or at least a source of uncertainty about the full labour costs. At the same time, the limited coverage and the poor enforcement mean that the programme does not offer very effective protection against a loss of income. In OECD countries where such programmes exist, the maximum severance benefit is usually lower, often three or four monthly wages. As part of a flexicurity package of labour market reforms, Chile should consider abolishing the severance-pay regulations in connection with its planned improvements of the UI programme (see further below). Such a reform was implemented in Austria in 2003 (Box 2.1). If the severance-pay programme is not abolished, the maximum amount should be substantially reduced so as to limit the disincentive for employers to sign indefinite-duration contracts. 14. Based on analysis of a panel of Social Protection Survey (ESP) respondents, Escobar (forthcoming) finds that 6.4% of all employed persons who became unemployed in certain periods were theoretically eligible for severance pay. But only about one-fifth of them (1.25% of the whole group) recognised having received it. 15. Different types of distortions are probably common. First, employers face an incentive to dismiss workers before 12 months tenure in order to avoid paying the benefits, with economically unjustified mobility as a result. Second, weak enforcement means that workers dismissed after 12 months cannot always defend their rights. Third, employers with limited profit incentives (e.g. in the public sector) may be persuaded to record voluntary quits incorrectly as dismissals giving a right to severance pay.

67 70 CHAPTER 2. TOWARDS MORE EQUAL JOB OPPORTUNITIES Box 2.1 Severance pay and unemployment insurance in Austria Individual accounts replaced severance pay by employers Austria converted its severance payment system to individual accounts in Employers direct payments to workers were then replaced by individual accounts, to which employers must contribute 1.5% of the payroll during the contract duration. If a worker leaves before three years of tenure, the individual account balance is transferred to the new job. But if the contract is terminated after more than three years, the worker has the options of receiving a severance payment from the account or transferring the balance to the new position. At the end of a career, the accumulated balance can be used for a pension. These individual accounts thus provide workers with a degree of security while allowing job mobility. The general coverage of all employees and all types of mobility means that the risk of distorting effects on labour market behaviour is relatively low. Unemployment insurance and unemployment assistance are unrelated to the individual accounts Austria also implements a UI programme for all employees and some of the self-employed. Financed by employer and employee contributions (each paying 3% of payroll), it pays benefits during certified unemployment that usually cover 55-60% of the previous income, with maximum durations of 20 to 52 weeks depending on the length of the contribution period. If the claimant is still unemployed after exhausting the UI entitlement, an unemployment assistance benefit can be paid for an additional 52 weeks, amounting to 95% of the previous UI benefit. About 70% of all formal employee jobs are of indefinite duration, but those with limited duration account for 60% of the annual labour turnover (ENCLA 2006). Fixed-term and other non-standard contracts are most common in construction and agriculture, followed by large parts of the service sector, while being less common in industrial enterprises and the transport sector (Table 2.1). It must be noted that these figures concern enterprises relations with their own employees, not those with sub-contractors and TWAs. (Some sub-contracting firms are probably counted as employers under the category business services, etc., which has a significant proportion of employees with task-related contracts.) All told, Chile s EPL is flexible in many respects, but the perception that some of its provisions might be too rigid, especially severance pay, is frequently mentioned not only as an extra cost for employers, but as a factor fostering disrespect of the law. Poor enforcement evidently reduces the value of such regulations for workers. But the legal uncertainty can nevertheless contribute to employers reluctance to hire workers with indefinite-duration contracts if these involve a risk of high costs or expensive litigation.

68 CHAPTER 2. TOWARDS MORE EQUAL JOB OPPORTUNITIES 71 Table 2.1. Incidence of fixed-term and other temporary labour contracts in 2006 Fixed-term Task or service Fees, etc. Total 1 Total: 1 thousands of Percentage of all employees with labour contracts persons Construction Agriculture, forestry Education Finance Business services etc Other services Commerce Mining Manufacturing, non-metal Health, social services Hotels, restaurants Transport, comm Fishing Utilities Manufacturing, metal Total Total = all employees whose contracts are not of indefinite duration. Source: ENCLA 2006 (Encuesta Laboral), covering firms with five or more workers. Sub-contracting and temporary-work agencies (TWAs) As a further option, big enterprises have achieved much flexibility over the past two decades by using sub-contractors and temporary-work agencies (TWAs). Both options were unregulated until 2007, when they became legally defined in a new section of the Labour Law. In OECD countries, by comparison, labour legislation usually regulates TWAs but not sub-contracting, which is more often treated as part of business law. But in Chile, a crucial function of these new rules is to distinguish the two phenomena from each other, a distinction that was hardly possible until the implications for labour law had been clarified. From now on, sub-contracting should only concern separate work processes and not merely dispatching or lending of employees, which is the typical function of TWAs. Although sub-contracting is common in OECD countries, its apparent role in Chile stands out by the extent to which it has raised suspicions intrinsically difficult to verify or reject that its real purpose might have been to relieve client firms of employer responsibilities. Such objectives are not necessarily incompatible with good labour standards. Nevertheless, it is pertinent to consider the possible role of sub-contracting in the context of Chile s generally high inequality and its often opaque labour market arrangements.

69 72 CHAPTER 2. TOWARDS MORE EQUAL JOB OPPORTUNITIES The policy decision to regulate sub-contracting, long resisted in business circles, responded to widespread claims that big enterprises were taking advantage of the lack of rules to hire and fire workers repeatedly through intermediaries, whose compliance with labour law was not the client firms responsibility. During the OECD review team s visit to Chilean authorities, the outcome was frequently described as a fragmentation of enterprises: the so-called phenomenon of 100 firms within a firm. 16 For client firms, this gave flexibility and lower labour costs and it prevented many workers from organising and bargaining together. For the workers concerned, typically low-skilled, the result was precarious jobs. According to the last ENCLA survey of firms with five or more workers conducted in 2006, thus before the reform over 40% of employers used sub-contractors. Among those who did, about one-third said it concerned work that was part of their principal activity. No figures were given about the numbers of workers concerned, but the practice was found in all economic sectors, with the highest incidence in big firms (Direccion del Trabajo, 2007). The sub-contractors were typically small or medium-sized enterprises, while 32% of them were individuals or families. Among all surveyed enterprises (with five or more employees), 12% had themselves done sub-contracting jobs during the past 12 months. Some 5% of all surveyed firms performed over 75% of their work as sub-contractors. Agreements between client firms and sub-contractors were most often oral (ibid.). It is not known how many of the workers concerned had labour contracts. However, sub-contracting was relatively common both in agriculture, where informality is widespread, and in big industry and service-sector firms whose own employees typically have formal contracts. In mining, for example, 95% of the firms own employees have labour contracts and earn relatively high wages. Under such circumstances, sub-contracting probably permits substantial labour-cost reductions even if the workers have formal labour contracts with the sub-contractors. As mentioned, the new law allows sub-contracting only for jobs that are separate from the client s own work process; in addition the client firm must give certain guarantees, e.g. for the event of a sub-contractor s failure to pay wages. These restrictions do not apply to the transitory workers dispatched from TWAs, but this solution is henceforth allowed only in certain situations of urgency and for no longer than three or six months, depending on the nature of the urgency. In effect, due to these tighter 16. One enterprise visited by the OECD employed about workers of whom a little more than 400 were its own employees, the others being sub-contractors or employees of sub-contracting firms. Sub-contracting contracts had been signed for five years in about 80% of the cases, while no information was available about the types of labour contract used.

70 CHAPTER 2. TOWARDS MORE EQUAL JOB OPPORTUNITIES 73 restrictions, many firms that previously used sub-contractors or TWAs are now expected to hire the workers or ensure that their jobs are separate from the main work process, in which case sub-contracting is still allowed. Moreover, the client firms shared responsibility for wages is putting pressure on them to act as labour inspectors and verify contract workers employment conditions. A survey conducted in industry during 2007 suggested little change since 2006 in the total use of external personnel including fee contracts, TWAs and sub-contracting (SOFOFA, 2007). 17 Such personnel was thus reported to hold 43.0% of available jobs in 2007 compared with 43.5% in Comparable statistics are not available from other sectors, but some firms have reportedly hired previous sub-contract workers e.g. cashiers in banks, whose jobs were considered as part of the main work process. According to officials interviewed in the Ministry of Labour and Social Security, the Labour Inspectorate has experienced many difficulties in implementing the new rules. The Inspectorate began to check who was acting as employer on many worksites, a type of investigation that is frequently obstructed. Big firms have challenged the powers of the Inspectorate in complex court cases, and the Supreme Court has partly sustained such complaints. In the mining sector, an attempt to enforce the new rules faced such strong legal and political opposition that the government accepted to suspend it, at least temporarily. 18 It is not known how the new rules may affect the future development of TWAs, but the immediate effect can be assumed to involve a significant reduction. At the time of the 2006 ENCLA survey, 38% of the enterprises used dispatched workers (trabajadores suministrados) and these were estimated to represent about 5% of employment in all surveyed firms. But over half of these jobs were expected to last over six months, implying that they would not be legal as TWA jobs today The employer association, SOFOFA, conducts annual remunerations surveys of industrial companies. The cited results are percentages of all workers in these companies who have contracts of one kind or the other. 18. CODELCO, a state-owned mining firm with some employees and many thousands of sub-contract workers, opposed the 2007 law change and its implementation. In late 2007, the Labour Ministry ordered CODELCO to hire over sub-contract workers. But the Supreme Court reversed this decision in April 2008, provoking a nationwide three-week strike in mines. 19. The 2003 CASEN survey asked employees and domestic workers to define their labour relations, and about 3% chose the reply option temporary services. However, it is not known to what extent the respondents might have understood this phrase as a reference to TWA jobs.

71 74 CHAPTER 2. TOWARDS MORE EQUAL JOB OPPORTUNITIES Working time For about half of the employed workforce including most employees, the usual working time is equal to the legal maximum, which was reduced in 2005 from 48 to 45 hours per week (Table 2.2). 20 A common complaint from enterprises including those visited by the OECD review team concerns the limited possibilities to reallocate the standard working time, notwithstanding some small changes towards more liberal rules that have recently been adopted. Table 2.2. Employed persons by effective working time Hours per week Percentage distribution Men , second half Women , second half <15 1% 5% 3% 8% % 6% 7% 11% % 4% 5% 5% % 12% 13% 13% % 50% 11% 43% 48 57% 5% 44% 3% % 7% 7% 6% % 2% 1% 1% % 10% 9% 9% Total 100% 100% 100% 100% Average working hours Source: Instituto Nacional de Estadisticas (INE). Part-time work has increased since 2001, contributing to the growth of female employment. However, the incidence of part-time in total female employment is still lower than in many OECD countries. By 2006, 19% of the employed women and 11% of the men worked less than 30 hours per week, compared with OECD averages of 26% and 8%, respectively. For employees in firms covered by the 2006 ENCLA survey, the proportion was only 14% for women and 3% for men, with the highest incidence in small firms. Overtime is allowed for up to two hours per day at a 50% wage premium, but only if this is agreed upon in writing in order to meet the needs of the enterprise. The working week must be either five or six days and it cannot include Sundays or public holidays, apart from certain sectors. 20. As in most countries, own-account workers are more likely than employees to deviate from the standard working time. See OECD (2005), OECD Economic Surveys: Chile, Table 5.2.

72 CHAPTER 2. TOWARDS MORE EQUAL JOB OPPORTUNITIES 75 According to ENCLA 2006, employees worked on average 0.7 hours of overtime per day, in this case with the highest incidence in large enterprises. The limited use of working weeks shorter than 45 hours is probably compatible with many households preference for higher incomes. But some rules in the law can unnecessarily discourage enterprises from introducing shorter working time. For example, if an enterprise introduces a shorter standard working week, the workers right to overtime pay will kick in from the reduced standard, while individual part-time labour contracts are not recognised as such if the agreed working week is over 30 hours, a fact that can also lead to higher overtime pay. Employers have repeatedly demanded a possibility to average working time over periods longer than a week, which is allowed only in certain areas such as transports. After heated debate in 2003, the authorities introduced a few minor changes in this direction, while rejecting a discussed proposal for more radical change called the adaptability law. 21 Had this proposed law been adopted, it would have become allowed to average working hours over a year via collective agreements in enterprises an idea still subject to controversial debate, as reflected in the Work and Equity Commission s report in Judging from OECD experience, such a reform would indeed be justified. But in addition, in Chile as in OECD countries, it must be recognised that the flexibility required in business is different from the type of flexibility that would facilitate employment of youths and women. As discussed further below, reforms to make working time regulations more flexible must therefore be accompanied by further measures to help parents combine work with family duties, such as child-care support and parental leaves. 3. Industrial relations and collective bargaining Chile s weak industrial-relations institutions are probably both a cause and an effect of the fragmented state of the labour market. Trade unions, being largely absent in small firms, have seldom been in a good position to promote formal employment and their role in other matters of law enforcement has also been limited. Industrial action has recently taken place on some important institutional issues, notably the application of the new sub-contracting law, but it is not clear how this may influence the chances of a fruitful social dialogue. 21. Cf. OECD (2003), Chapter IV. The limited reform adopted in 2003 permits such rescheduling that workers can stay at home on a one-day bridge between a Sunday and a public holiday. The list of sectors allowed to work on Sundays was also extended.

73 76 CHAPTER 2. TOWARDS MORE EQUAL JOB OPPORTUNITIES Certain groups are legally excluded from collective bargaining rights. This especially concerns civil servants, along with the employees of a limited number of enterprises that depend on state financing, as well as management staff and workers hired exclusively for temporary tasks. 22 The exception for temporary tasks could potentially concern many workers cf. Table 2.1 but this does not appear to be the case in practice, due to a restrictive administrative interpretation. 23 Trade union membership has been stable at about 11% of total employment since the late 1990s, which currently corresponds to 13% of all employees and 15% of those with a right to organise (Table 2.3). Higher membership is found in big enterprises, especially in sectors with significant public ownership such as mining, transportation and communications, and utilities (electricity, gas and water). But even in the group of enterprises with over 200 workers, a majority of the firms have no trade unions, a proportion that rises to 78% for firms with workers and 96-99% for the smallest size categories (ENCLA 2006). Table 2.3. Trade union membership in 2007 Agriculture, forestry, fishing Mining Utilities Manufacturing Construction Commerce Transport, comm. Finance, business services Social and personal services Number of union members Members as a percentage of employment 9% 42% 14% 20% 11% 10% 22% 7% 8% 11% The percentages refer to total employment in each sector. Source: Direccion del Trabajo (2008); Instituto Nacional de Estadisticas (INE). The Labour Law sets practically no limits to the scope or coverage of collective bargaining. But in practice, any bargaining that occurs tends to take place within enterprises and to focus on wages only. Worker representatives receive some outside assistance in a little more than half of the cases. Newly signed collective agreements have recently concerned Total 22. Articles of the Labour Law exclude the following groups from collective bargaining rights: civil servants including police and defence; employees of state-owned enterprises if they depend on the Ministry of Defence or have been financed at over 50% by the state in the past two years; apprentices; employees hired exclusively for temporary tasks; employees with general managerial powers and those authorised to hire or fire workers or to decide about production processes or marketing. 23. According to explanations provided by the Ministry of Labour and Social Security, the exception applies only if the employer notifies the Labour Inspectorate in each case and if the Inspectorate then decides that the worker can be legally excluded from collective bargaining.

74 CHAPTER 2. TOWARDS MORE EQUAL JOB OPPORTUNITIES 77 around workers per year (about 5% of employees), of whom about three-fourths were represented by trade unions, the others by ad hoc groups of bargaining workers. There is often more than one union or bargaining group in an enterprise, each of which is bargaining for its members. Collective agreements usually last two years, so around 10% of the employees are likely to be covered. However, only about two-thirds of the valid agreements are of the formal type (contratos colectivos) that must be bargained according to certain procedures. The remaining third were of a less-regulated type called convenios colectivos. 24 Judging from data submitted by the Ministry of Labour and Social Security, the greatest bargaining efforts aim to achieve scheduled wage adjustments to inflation, often about twice a year. Between 2000 and 2006, when the total real-wage growth was around 3% per year, the negotiated initial real-wage increases never reached 1% on average (Table 2.4). Bargained real-wage changes do not appear sensitive to cyclical changes. Almost half of the collective agreements envisage adjustments to productivity; but in practice, the partners have seldom been able to agree on how to achieve this. Table 2.4. Wage adjustments resulting from collective agreements Percentage of wage increase by type of instrument and type of worker representation Average initial increase (%) Trade union Other worker grouping Scheduled readjustments as a % of CPI Trade union Other worker grouping Readjustment periods (months) Trade union Other worker grouping Data for 2006 are preliminary Source: Direccion del Trabajo (DT). 24. Contratos colectivos can be negotiated by ad hoc bargaining groups as well as by trade unions, provided that they have been set up according to certain rules in the law and that certain bargaining procedures are followed. If the latter procedures are not followed as happens most often when ad hoc groups are concerned the parties can only sign convenios colectivos.

75 78 CHAPTER 2. TOWARDS MORE EQUAL JOB OPPORTUNITIES During 2008, the government has announced its intention to propose legal changes that would promote a wider use of collective bargaining, which it regards as the best way to promote social dialogue based on trust and co-operation. The aim would be to simplify bargaining in enterprises, make it richer in content and strengthen the representativeness of enterprise unions by involving more actors. To promote bargaining in matters other than wages, a procedure would be established for declaring unions representative in such questions as working hours, overtime, shift work and social benefits. The government has also offered to support training of trade union representatives in enterprises. This initiative was motivated by the policy makers perception that opinions about required changes were frequently not shared by employers and unions and that the spaces available for spontaneous dialogue had not been sufficient to create useful precedents. The suggestions had been designed to promote a more encompassing dialogue at enterprise level, but did not envisage any change with respect to the possible co-ordination between enterprises an idea that has been strongly promoted by unions but resisted by employers. Some international studies have suggested that sector-level wage bargaining leads to lower employment than either enterprise or national-level bargaining (OECD, 2006). Nevertheless, it would be useful in Chile if the social dialogue in large enterprises could be extended to cover some of the issues about sub-contractors and their employees. The minimum wage is relatively high The OECD (2003, Chapter IV) found evidence that a substantial increase in the minimum wage since the mid-1990s had contributed to a rise in unemployment. This result is still relevant because the minimum wage has remained in the range of 42-45% of the average wage for most of the period since The minimum wage was raised in mid-2008 to pesos per month, which then corresponded to a little over 45% of the average wage a higher proportion than in most OECD countries (Figure 2.2). Given Chile s relatively uneven wage distribution, the mentioned amount probably corresponds to over 50% of the median wage. Changes in the minimum wage are decided by Congress after discussions between the government and trade unions. It is unclear why business representatives do not to participate in these discussions, but the political decision makers should in any case take account of the full implications for the economy and the labour market. A high minimum wage can discourage job creation in the formal sector. For the smallest firms, where informality is still relatively widespread, it complicates the policy efforts to promote formality and to extend the

76 CHAPTER 2. TOWARDS MORE EQUAL JOB OPPORTUNITIES 79 pension and healthcare coverage. Significant groups earn less than the minimum wage; the 2006 CASEN survey reported that a little over 10% of all full-time employees aged excluding domestic services earned less than 0.95% of the minimum wage. 25 There is nevertheless some evidence that the high minimum wage is putting upward pressure on informal as well as formal wages, implying that it can discourage hiring in both sectors (Cunningham, 2007). Perhaps most preoccupying are the likely effects for youths, whose entry into formal jobs should be encouraged. The Reassessed OECD Jobs Strategy recommended reduced minimum wages for problematic groups, especially youths. Chile s minimum wage is currently reduced to 75% of the standard amount for workers who are younger than 18 or older than 65, and to 83% for domestic workers of any age. However, available labour market data leave no doubt that the special difficulties faced by youths concern young age groups in general and not only those under 18. In consequence, a reduced minimum wage rate should be applicable at least up to the age of % Figure 2.2. Minimum wage as a percentage of the average wage in OECD countries and Chile 50% 40% 30% 20% 10% 0% Note: Full-time minimum and average wages are compared before tax. Source: OECD (2006), Taxing Wages 2005/2006, Paris. 25. In contrast to many OECD countries wage statistics, the survey-based figures in CASEN 2006 reveal no significant concentration of wages near the minimum level, suggesting weak enforcement. Thus, only 2.2% of employees reported between 0.95% and 1.05% of the minimum wage. This does not exclude that a high minimum wage can have a positive impact on informal as well as formal wages via market competition.

77 80 CHAPTER 2. TOWARDS MORE EQUAL JOB OPPORTUNITIES 4. Labour taxation and informal employment Chile s labour taxation is relatively low compared with OECD countries, a fact that should facilitate a rise in formal employment. 26 The top marginal tax rate is 40%, but most workers pay only the mandatory contributions to social insurance (see Box 2.2). These are charged at a total rate of about 26% of the wage for employees, of which about 20 percentage points are deducted from their wages and the rest is paid by the employers. Box 2.2. Labour taxation and mandatory contributions Income tax is charged only on incomes higher than about 1.4 times the average wage. The marginal tax rate begins at 5% and rises in six steps up to a maximum 40%, which currently applies to wages of over about 15 times the average wage. This tax scale follows consumer prices. Social contributions: Employees must contribute to pension, healthcare and unemployment insurance. For the self-employed, only healthcare and work-injury insurance can be mandatory. The total contribution rate, usually about 26%, includes the following items: 12.5% for pensions, paid by employees and the self-employed. 7% for healthcare, also paid by employees and the self-employed % for work-injury and occupational diseases, paid by employers and the self-employed. These contributions are differentiated by sector according to occupational risks. Unemployment: Indefinite-duration contracts: 0.6% for employees and 2.4% for employers. Fixed-term contracts: 3% for employers. Social contributions are charged on incomes that exceed the minimum wage for pensions, while a lower floor applies in the other programmes. Pension contributions are currently charged on incomes up to about 3.5 average wages, while the ceiling applicable to unemployment insurance is about five average wages. These ceilings follow the price index. According to OECD experience, an efficient collection of taxes and social contributions is usually the most important instrument for enforcing formal employment regulations, but Labour Inspectorates can be important as a complement, especially in countries with large informal economies 26. In the average OECD country, a 10-percentage-points reduction in taxes and social contributions can be expected to reduce unemployment by 2.8 percentage points and increase employment by 3.7% (OECD, 2006). The impact on employment is even greater if effects on working time are taken into account.

78 CHAPTER 2. TOWARDS MORE EQUAL JOB OPPORTUNITIES 81 (OECD, 2004, Chapter 5). OECD (2007) found that Chile s tax administration is relatively efficient by Latin American standards, a fact that contributes to a lower incidence of informality than elsewhere in the region (see Chapter 1, Figure 1.8). However, less than half of all employed persons have earnings at the level where income tax is charged. Social insurance contributions are collected by the respective insurance administrators. While the incidence of informal employment is relatively low in Chile compared with neighbouring countries, it is still widespread by OECD standards. Chapter 1 above found that about one-third of the employed do not contribute regularly to the social security programmes, a situation that concerns not only most of the self-employed (who are not obliged to contribute) but also a little more than one-fifth of the employees. The employees who do not contribute often lack labour contracts as well, implying that few labour regulations are likely to be enforced. Several recent reforms and reform proposals have been designed to promote the formalisation of informal jobs. The 2008 pension reform (see Chapter 4) introduced subsidies to pension insurance for young workers, and it will gradually make pension insurance mandatory for the self-employed (assuming that they are known to the tax authorities). The Work and Equity Commission proposed several further measures of potential relevance, including a pre-formalisation programme that would permit small firms to become formal in a gradual manner over three years, and an in-work subsidy for poor persons who take up formal jobs. As discussed in Chapter 1, informality is often a result of tax evasion and/or a desire to avoid complying with various regulations. But in the present situation in Chile, a significant part of the remaining pockets of informality represents a fragile segment of the labour market, including many small firms often characterised by below-average levels of education, productivity and income as well as precarious job conditions. Enforcement measures that target micro-firms may therefore have to be accompanied by other targeted measures for them, as envisaged in the just-mentioned proposal for a pre-formalisation programme. Moreover, both taxes, social contributions and various regulations should be regularly reassessed with a view to enforcement issues. For example, as the government is now facing the challenge of implementing regulations of sub-contracting, it also needs to consider the risk that some businesses that previously used sub-contractors might now respond by turning to informal arrangements.

79 82 CHAPTER 2. TOWARDS MORE EQUAL JOB OPPORTUNITIES 5. The Labour Inspectorate (Direccion del Trabajo, DT) and labour courts Much of the existing segmentation and inequality in the labour market must be attributed to an uneven implementation of existing rules in the Labour Law. Considerable progress is under way, notably in terms of various legal procedures described in the following. The labour inspectorate henceforth called DT is the main enforcement body. As in most countries, it cannot monitor all enterprises, so an important function is to deal with disputes and complaints. The DT has employees, of whom are professional staff members, including the central administration, 15 regional offices and 80 provincial or municipal Labour Inspection offices. It is responsible for enforcing the Labour Law and regulations of hygiene and safety at work, and it can intermediate between workers and employers and it plays a role in informing them about the regulations. There is no mandatory registration of labour contracts, but the DT can check during inspections that workers have contracts and that social contributions are paid in proportion to declared wages. Over inspections were carried out in 2005, concerning 1.7 million workers or over 40% of all employees (Table 2.5, Panel A). About two-thirds of these inspections responded to complaints by workers, the others being conducted ex officio (Panel B). Most complaints concerned wages, followed by working time and lack of contract. Table 2.5. Labour inspections A. Number per year Inspections ex officio by complaint Workers covered B. Reasons for workers complaints Length of working day Wages Sanitary conditions Labour contract Contributions Source: Direccion del Trabajo (DT).

80 CHAPTER 2. TOWARDS MORE EQUAL JOB OPPORTUNITIES 83 In 2005, the DT imposed fines on employers in cases, of which about concerned labour contracts. 27 But the employer is usually given some time to correct the problem, and if this is done, the fine is reduced or suspended. If the employer chooses to pay the fine, the DT can seldom do more unless the worker brings the case to a labour court, a step that must be initiated by the person concerned. It therefore falls on the employee to check that the employer pays social insurance contributions. Apart from inspections, neither the DT nor any other public institution is likely to intervene if this is not done. For the employee, it can be difficult to detect non-payments and take the necessary legal steps: pension, healthcare and unemployment insurance are separate institutions using partly different procedures. 28 Pension contributions are delayed most often because the pension fund agencies (AFPs) generally allow this, requiring only that the employer acknowledges the debt. The tax authority is stricter, imposing monthly provisional tax payments and annual consolidations, with substantial penalties if the provisional payments prove too low. The DT can also seek to identify and inspect unregistered enterprises that are unknown to the tax authority, but this is not done to a very large extent. The tax authority does not intervene in such cases. As mentioned, however, the government plans to transfer such responsibilities from AFPs to the tax authority in a near future. The government has recognised that the court system for labour cases is cumbersome, bureaucratic and slow (often about three years before a judgement). Employers can frequently obstruct the process, e.g. by not providing the necessary proof that the complainant has worked in the firm. Significant policy efforts are therefore in progress to increase the capacity to handle labour cases and to improve the procedures. Specialised courts for labour and social insurance were set up in the most populated regions in 2006, but elsewhere the ordinary courts must also handle labour cases. The number of specialised labour judges will increase gradually from 19 to 84. The number of labour cases resolved increased from in 2004 to in 2007, of which under the specialised courts. A new court procedure for labour cases, introduced in 2008 in two small regions, will replace the present slow procedure in the whole country by Key elements of this reform, supported by the main employer and worker associations, are that 1) all procedures are oral and the judge hears 27. The usual penalty for lack of contract is 5 UTM (unidades tributarias mensuales, an inflation-adjusted unit for monthly taxation) or about USD The payment of contributions to healthcare insurance may be most easily checked by applying for a healthcare service, which will be refused if the insurance is not valid.

81 84 CHAPTER 2. TOWARDS MORE EQUAL JOB OPPORTUNITIES both the employer and the worker at the same time; 2) the judge can act ex officio to prevent that proceedings are delayed or paralysed by deliberate actions by either party; 3) proceedings are free of charge and special lawyers are available to defend workers; and 4) the possibilities of appeal are restricted. Where the new system is operational, it takes on average 90 days to resolve a case. In addition, the DT has begun to offer an optional fast-track or monitoring option for disputes of limited economic consequence (up to ten minimum monthly wages). If the worker chooses this option, which takes about 30 days, the DT first organises a conciliatory meeting; if no agreement is reached, the worker can still turn to the court, which can often decide quickly based on the minutes of the conciliatory meeting. 29 Another fast-track procedure has been created for cases that concern fundamental rights, including trade union rights. 6. An atypical unemployment insurance programme (UI) Chile s UI, described in Box 2.3, consists mainly of individual accounts, from which the accumulated contributions are paid out on job separation for any reason, most often as a lump-sum. A subsidised Solidarity Fund, little used until now, can give complementary support to claimants who are unemployed and dismissed for economic reasons, provided that their previous earnings were modest. Those claiming support from the Solidarity Fund are compensated according to a defined-benefit model, based on previous income and fixed replacement rates (up to 50%, with a modest maximum amount), paid in up to five monthly benefits. But the Solidarity Fund itself will contribute only when the individual account is empty. The use of individual accounts has almost no equivalents among UI schemes in OECD countries, though a few other Latin American countries have programmes that share some of its characteristics. Austria implements a similar individual-account programme see Box 2.1 above but this was introduced to replace employers severance payments and not as a complement to these. Austria also administers a UI scheme of an internationally common model as well as an unemployment assistance programme. 29. The worker can ask the court to make a judgement based on the minutes from the conciliatory meeting and other documents. The judge can then make a judgement or call the parties to another conciliatory meeting within fifteen days. In the first case, either party can ask the judge to come and explain the sentence in a conciliatory meeting, or appeal to a higher court.

82 CHAPTER 2. TOWARDS MORE EQUAL JOB OPPORTUNITIES 85 When the Chilean programme was introduced in 2001, workers ownership of individual funds was seen as a means to make participation attractive and to reduce moral hazards. It was deemed desirable to reduce the need to target the unemployed, which would be difficult in the absence of a fully-fledged employment service. But large parts of the accumulated funds must then be paid out for reasons other than unemployment, reducing the resources available for this particular contingency. The relatively strict rules for Solidarity Fund benefits were initially motivated by concerns with the risk of abuse (moral hazard). However, officials interviewed by the OECD review team acknowledged that these fears had been exaggerated, contributing to the limited use of the Solidarity Fund. At present, only some 10% of all claimants are eligible for benefits from this fund, and perhaps only half of them find it worthwhile to apply. In the Bill of August 2008, mentioned in Box 2.3, the government proposed to make the Solidarity Fund s guaranteed benefit scale somewhat more generous and to apply it in all cases of dismissal for economic reasons. In all such cases, the claimants would have to register as unemployed. As a further novelty, Solidarity Fund benefits would also become available to workers who complete fixed-term jobs, but with much less generous benefit amounts. As seen in Table 2.6, those terminating fixed-term contracts are the biggest group of UI claimants under the present (unreformed) system. Box 2.3. Chile s unemployment insurance (UI, seguro de cesantía) Financing and main structure Chile s UI, introduced in 2001, consists of individual accounts with a small Solidarity Fund as a complement. It is financed by contributions from 1) employers, who pay 2.4% of the wage for indefinite-duration contracts and 3% for fixed-term contracts, 2) employees with indefinite-duration contracts (0.6%); and 3) the state budget. Contributions are not charged on wages in excess of UF 90 (over five average wages). The state subsidy goes to the Solidarity Fund along with 0.8 percentage points of the employer contribution for indefinite-duration contracts. All remaining contributions are allocated to the workers individual accounts. Administration has been entrusted, after competitive bidding, to special bodies connected with the pension-fund administration. An indefinite-duration worker s individual account will thus receive 2.2% of the wage for up to 11 years in a job, after which no more contributions are charged. The worker will thus accumulate 26% of a monthly wage per contribution year, to be adjusted for financial returns, administrative costs and changing wages (see Table A) Similarly, a fixed-term worker would accumulate 36% of a monthly wage for every 12 months of work. By comparison, the severance benefit that employers must pay on dismissal for economic reasons amounts to a full monthly wage per year of service, with a ceiling at 11 years. The employer s contributions to a dismissed worker s UI account are deductible from the mandatory severance benefit.

83 86 CHAPTER 2. TOWARDS MORE EQUAL JOB OPPORTUNITIES Table A. Expected amounts of benefits from individual accounts compared with severance pay All amounts are expressed as numbers of monthly wages Number of years in a job Accumulated capital on the individual account Hypothetical benefits from the individual account Total Of which: from Total Month 1 Month 2 Month 3 Month 4 Month 5 Total employer contributions 1 Severance pay by the employer 1 After deducting the employer's contributions Note: These calculations are only a hypothetical illustration. They disregard administrative costs, financial returns and changing wages. Severance pay applies only when indefinite-duration workers are dismissed for "enterprise needs" or force majeure. 1. Assuming an indefinite-duration contract. Source: Ley del eguro de cesantia. Benefit rules Once a worker has been employed for 12 months (six months for fixed-term contracts), the individual account is available on job separation for any reason. Withdrawals of funds must be made in as many monthly payments as the number of years of service, up to a maximum of five. These payments come at declining amounts, which add up to the account balance if they are all paid. But to receive more than one benefit, the claimant must also register as unemployed at the municipal employment service (OMIL) from the second month. Once a claimant finds a job, he or she can collect the benefit for only one more month, while the remaining funds will stay on the account for possible future use. If the account balance is too low to permit a certain level of compensation, claimants who are dismissed from indefinite-duration jobs for economic reasons ( enterprise needs ; the Labour Code s Art. 161, or force majeure) and become unemployed can apply for an additional benefit from the Solidarity Fund. This can be done twice in a five-year period. When the Solidarity Fund is involved, it guarantees that the total benefit in the first month is 50% of the last wage up to a ceiling, in the second month 45%, and so on down to 30% in the fifth month (all depending on the number of benefit months that apply in each case). The benefit ceiling follows consumer prices, so it has declined since 2001 relative to typical wages. It is currently around 90% of the minimum wage in the first month of unemployment, and subsequently lower. In effect, a complement from the Solidarity Fund can be relevant mainly to workers with moderate job tenures (up to about seven years) and modest wages (cf. Table A).

84 CHAPTER 2. TOWARDS MORE EQUAL JOB OPPORTUNITIES 87 How many are concerned and what they receive Affiliation to UI has increased gradually and now concerns some 5.6 million workers, of whom about 3 million (and/or their employers) contributed in June 2008 (cf. Table B). Those with fixed-term or similar contracts accounted for over half of the insured, but contributed less regularly and with low average amounts. A large majority of accounts had accumulated less than one minimum wage (then CLP ; cf. Table C). Monthly benefits were paid to workers a figure that corresponds to about one-third of the survey-based unemployment, though it is not known how many of the recipients were unemployed (Table 2.6). The average benefit payment is about 30% of the average wage (Table 2.6, Panel A). Over half of the benefits concern temporary workers, who all receive lump-sums. Claimants who previously had indefinite-duration contracts collect on average two monthly payments. Indefinite-duration workers represent almost all of those considered as dismissed, but they also include a significant proportion of workers who quit their jobs voluntarily. The Solidarity Fund is involved in only about 6% of all benefit cases. Officials interviewed by the OECD team thought that a similar number of clients may well be eligible for Solidarity benefits but never apply for them, contenting themselves with the right to withdraw individual funds without having to register as unemployed. Table B. Individual UI accounts in June 2008 Type of contract, gender Thousands of accounts Average balance (in CLP) Thousands of contributions made in the month Average monthly contribution (in CLP) All insured workers Men Women Unknown Indefinite contracts, total Men Women Unknown Fixed-term or other temporary contract Men Women Unknown Source: Calculations based on SAFP data (

85 88 CHAPTER 2. TOWARDS MORE EQUAL JOB OPPORTUNITIES Category of account holders Table C. Persons affiliated to UI in June 2008 by contribution status and account balance Affiliated persons Number Percentage Percentage of all of the accounts contributors Up to Average account balance (in CLP) or more Contributors by taxable income (in CLP): Up to % 10% % 30% % 22% % 19% % 12% % 5% More than % 2% All contributors % 100% Percentage distribution 100% % 10% 9% 3% 2% Minimum wage earners ( CLP) % Non-contributors % All insured persons % Note: The monthly contribution to these accounts is 2.2% or 3% of the taxable income. Source: Calculations based on SAFP data ( The Government s Bill of August 2008 The proposed reform has the following main elements: 1. Benefits to the unemployed would as a rule be calculated according to the more generous five-month benefit scale now used only when the Solidarity Fund is involved. This Fund would be used whenever necessary to cover the expenses. 2. Access to the Solidarity Fund for the unemployed who previously had fixed-term or temporary contracts. To minimise negative effects on work incentives, this group would receive only two monthly benefits, replacing previous earnings at the rates of 35% and 30%. 3. Relaxed contribution requirement for benefits from the Solidarity Fund (still 12 months, but not necessarily continuous). 4. A possibility for the government to pay two additional monthly benefits after indefinite-duration jobs, applicable in situations of high unemployment. The two benefits will be worth 25% and 20% of the previous wage, respectively. After fixed-contracts, the income replacement rate would instead be increased. 5. Reallocating 0.2 percentage points of employer contributions for fixed-term workers to the Solidarity Fund.

86 CHAPTER 2. TOWARDS MORE EQUAL JOB OPPORTUNITIES 89 Table 2.6. UI benefits in June 2008 A. Benefits paid: number of cases and average amounts by source of finance Number of cases and average amounts, expressed as percentage of the average wage for both genders (CLP ) Total Individual account* Solidarity Fund Mixed Cases Average amount (% of average Cases Average benefit (% of average Cases Average (% of average wage) Cases Average (% of average wage) Total % % % % Men % % % % Women % % % % * Including persons whose benefits were paid out according to the rules about Solidarity Fund benefits (Art. 25). B. Benefits paid and approved benefit claims by reason and type of job contract Type of contract by gender Benefit payments New benefit claims approved: total and percentage distribution by reason Total Voluntary quit End of contract Individual (Art. 160) Economic (Art. 161)* Total % 20% 58% 5% 17% Men % 20% 58% 5% 17% Women % 22% 58% 4% 16% Indefinite contracts % 36% 5% 8% 50% Men % 34% 6% 8% 52% Women % 42% 4% 7% 47% Fixed-term contracts or similar % 14% 78% 4% 4% Men % 14% 78% 4% 4% Women % 14% 81% 3% 3% * Including force majeure (Art. 159, point 6). Source: SAFP ( Several factors make it difficult to predict the effects on labour market behaviour. Assuming that individuals have varying preferences with respect to risk and consumption, it is perhaps inevitable that some will be eager to withdraw their individual funds. 30 The proposal to extend the use of the Solidarity Fund s benefit scale would increase the proportion of benefit claimants who must register as unemployed, and their benefits would be interrupted if they find work. This can encourage excessive mobility in some 30. UI funds have a limited range of investment options, primarily fixed-interest instruments. Gross returns were 1.22% in real terms over the period August 2007-July 2008 (SAFP), but an administrative fee of 0.6% of the capital is deducted once a year, subject to some cyclical adjustments.

87 90 CHAPTER 2. TOWARDS MORE EQUAL JOB OPPORTUNITIES groups: after 12 months in a job, a worker would still get access to the individual funds as a lump-sum if he or she quits in agreement with the employer. As mentioned already, more consideration should be given to the possibility of a flexicurity package of reforms. If more generous UI benefits are envisaged for the unemployed, they should be combined not only with an effort to improve the public employment service as the government has envisaged, see below but also with the elimination of employers duty to pay severance benefits. 7. The public employment service An important lesson OECD countries have drawn from experience is that reforms to make the unemployment compensation more generous will increase the need to monitor and support individual efforts to find jobs. This was also stressed in the Chilean Government s Bill to Congress in August 2008, which envisaged a new information system for the employment service and improvements of the electronic labour exchange Bolsa Nacional de Empleo (see below). The SENCE offices that administer training and hiring-subsidy programmes for registered job seekers would also be strengthened. Employment services are provided by 233 Municipal Employment Intermediation Offices (Oficinas Municipales de Intermediación Laboral, OMIL), covering 277 of the 345 municipalities. They had altogether 654 staff members in 2007, with numerous small municipalities having only one employee for this function often in the town halls, not necessarily as a distinct office while bigger cities may be somewhat better equipped. 31 Though belonging to the municipal administrations, OMILs depend essentially on state funding. Most municipalities have too low revenues of their own to contribute much to the employment service, implying that few of them can develop policies of their own. 32 Key OMIL tasks are to certify unemployment on behalf of the Solidarity Fund, which is done in co-operation with the UI administration, and to keep a similar register of Chile Solidario clients. Both registered groups are presumed to seek jobs and demand the assistance they may need. However, 31. The biggest city areas, Santiago and Concepción, consist of several municipalities which have their own OMIL offices. 32. Local taxation targets a limited number of items such as alcohol, car registration, tourism and property development. The OMIL office in a prominent tourist resort, whose leader met the OECD team, was exceptionally well-equipped by Chilean standards.

88 CHAPTER 2. TOWARDS MORE EQUAL JOB OPPORTUNITIES 91 according to what the OECD team was told during visits in two regions, many clients do not come back to the office after the initial registration, and only a few of the most well-equipped OMILs are likely to contact them unless there is a suitable job offer. But OMILs also serve various other clients who come to the office spontaneously, including young first-job seekers. According to data submitted by the Chilean authorities, the annual flows of registered job seekers, new vacancies and filled vacancies are modest compared with the size of the labour force, and they have declined since In 2006 there were reportedly newly registered job seekers, new vacancies and filled vacancies. 33 Clients are encouraged to use the Bolsa Nacional de Empleo, a nationwide search engine available free of charge on the Internet, which the UI administration created in 2006 in order to help job seekers and employers to find each other. The Bolsa also helps OMILs to check that Solidarity Fund beneficiaries fulfil the requirements for certification. Still under development, it has registered fewer workers and vacancies per month than the OMILs, but the ratio of filled vacancies appears more favourable, e.g. during August 2008 it registered vacant jobs while 837 vacancies were filled ( Active labour market programmes (ALMPs) are relatively insignificant The OMIL network is supervised and complemented by the National Training and Employment Service (Servicio Nacional de Capacitación y Empleo, SENCE) part of the Ministry of Labour and Social Security which is represented in all regions and has 532 staff members. SENCE provides OMILs with technical and financial support and administers training and job-subsidy programmes for OMILs job seekers, notably the beneficiaries of the Solidarity Fund and Chile Solidario. Clients may also be referred to more broadly-targeted adult training via Chilecalifica, a programme managed jointly by the Ministries of Economy, Education and Labour. In 2007, subsidies were paid for recruitment of about unemployed clients, while under received training, to which must be added a similar number of Chile Solidario clients. (These measures are separate from SENCE s larger training scheme for employed workers see below but the same training centres can be used.) By OECD standards, all these job-subsidy and training programmes are small relative to the number of potential participants registered at the OMILs. Where they have been implemented, they have often appeared to be 33. Available figures for 2007 are much lower, but this is possibly a result of slow reporting.

89 92 CHAPTER 2. TOWARDS MORE EQUAL JOB OPPORTUNITIES successful and welcomed by employers, as for example an apprenticeship programme and a similar scheme under Chile Solidario (Jovenes), which together are expected to enrol around young clients in However, OECD experience of ALMPs indicates that such positive results would be difficult to sustain if they were to be implemented on a much larger scale. If more large-scale training schemes are found necessary for disadvantaged youth, e.g. school drop-outs, OECD experience indicates that the objectives are more likely to be achieved if targeted measures are implemented within the general education system. The most cost-effective type of active labour market policy is usually the basic job-counselling and placement. When unemployment benefit claimants are concerned, such services should be combined with effective controls of availability for work. To the extent that more expensive programmes such as training and job subsidies are targeted on the unemployed, job counsellors are usually best place to select the individuals whose job chances are likely to improve as a result of a particular measure. For this reason, an expansion of the more expensive active programmes should not be a high priority as long as the budgetary and administrative capacity is not sufficient even for job counselling. The municipal employment offices are probably so weak in large parts of the country that they are unlikely to fulfil their expected functions in relation to the unemployed. The chief priority must be to strengthen the job-counselling function and to monitor individual job-search efforts. This problem is set to become more urgent if UI and other benefit schemes become more generous. The additional services now provided by SENCE and the Bolsa Nacional de Empleo are evidently needed as a complement. Various options for organisational reform merit consideration, including that of replacing OMILs by a nationwide office network under the national employment service agency, as found in most OECD countries. 8. Job-related training and lifelong learning Extended initial education has already made large parts of the young generation better prepared for entering the labour market and for lifelong learning on and off the job. An important long-term effect is that demands for vocational training are addressed increasingly to higher education institutions, both at a young age and as recurrent courses for adults. In line with recommendations in OECD (2005), Chilean authorities decided in 2006 to 1) give more financial support of higher-education for students from low and middle-income households, 2) step up checks on quality through the

90 CHAPTER 2. TOWARDS MORE EQUAL JOB OPPORTUNITIES 93 accreditation of higher-education providers, and 3) pursue reforms that make academic training more attuned to market demand. 34 But education and training policy must still cater for numerous adults not eligible for higher education, a group that includes many of those at risk of unemployment. Moreover, lifelong learning usually depends on a complex interaction between work and learning that is difficult to achieve unless employers and workers share a mutual interest in skill enhancement. According to OECD experience, low-educated workers are usually among the least likely to attend job-related adult training, a problem that is accentuated when they are employed under precarious conditions. Secondary education, where enrolment is now relatively high, includes a significant element of vocational as well as general options. A large expansion of the system in the 1980s and 1990s was accompanied by policy efforts to develop vocational courses, targeting a range of occupations selected in co-operation with employers. A dual apprenticeship programme was also introduced, inspired by the system in Germany. But in Chile as in many OECD countries, more recent reforms have reduced the differentiation of secondary education, which now generally includes two years of common education for all followed by two final years with general and vocational streams. The OECD (2004a) argued that while the previous emphasis on vocational education may have been appropriate at the time, putting too much emphasis on vocational education at the secondary level may be undesirable if it reduces the chances that the pupils will subsequently be able to participate in tertiary education. Employee training in Chile is financed predominantly by tax credits, deductible from a 1% payroll tax, and to a lesser extent by grants financed by this tax. About 20% of the dependent employees reportedly participate each year (about 1 million persons in 2008, up from in 2000). As observed in OECD (2005 and 2007), the administrator, SENCE, has limited control because it must generally approve tax relief when the training is purchased from an authorised provider, and also, with few exceptions, if it is provided in-house. Employers are free to choose training content and to select trainees. A slightly higher tax rebate is available if the training is decided by a bipartite training committee at enterprise level, but very few enterprises use this option. In 2005, the tax rebate was replaced by training 34. The number of higher-education grants to students with modest incomes increased by about 20% in 2007, with additional support for students from the two lowest income quintiles and a new type of student loans. A new Higher-Education Standards Law (Ley de Aseguramiento de la Calidad de la Educación Superior) led to the creation of a national accreditation commission (CNA). Other changes concerned curricula, equipment and management (the so-called MECESUP2 Programme, second phase).

91 94 CHAPTER 2. TOWARDS MORE EQUAL JOB OPPORTUNITIES grants for the smallest micro-firms with up to one employee, where the frequent presence of informal and family workers had made it difficult to assess the wage costs. Policy makers have been concerned for a considerable time about the quality of such training. The relatively lax regulations combined with insufficient information especially in small firms have tended to make the programme supply-driven as employers are sensitive to training centres marketing. The government and SENCE have gradually tightened the requirements for accreditation of training centres (OTECs), with important changes notably in 2006, after which the number of approved training centres was reduced by almost one-third to about SENCE also seeks to provide more advice and information to employers about the training. A national certification system for vocational skills has been under development for several years. Regulated in law since mid-2008, it has been designed to cover the whole labour market although standards have not yet been developed for more than a limited part of the job market (e.g. tourism, food and beverage production, construction, technical and mechanical installations). The effort has been managed by SENCE until now, but the new law envisages the creation of a tripartite commission with powers to adopt sector-specific skill requirements. This will be co-ordinated jointly by the Ministries of Economy, Education and Labour and Social Protection under the Chilecalifica programme for lifelong learning, which covers vocational guidance and measures for educational up-skilling as well as vocational and technical training. Some tax rebates will also be available for skill evaluation and certification. Strengthen SENCE s controls and speed up the development of skill certifications As noted in OECD (2007), SENCE s decision powers need to be further strengthened if it is to use the tax relief effectively as lever to enforce quality. This is required especially when tax relief is accepted for in-house training in enterprises. Many OECD countries have similar experience of programmes to subsidise enterprise training, and for example Belgium, Canada, France, Ireland, Spain and the United Kingdom have from time to time used levy/grant or train-or-pay schemes resembling the present one in Chile (OECD, 2005, Box 5.3). Some positive effects have usually been reported, but international experience suggests that such programmes often have a high deadweight effect and they can lead to distortions of the training effort. In any case, subsidies and tax relief are less important than the institutional developments that are necessary in order to make the many

92 CHAPTER 2. TOWARDS MORE EQUAL JOB OPPORTUNITIES 95 different forms of general and vocational education and training compatible with each other and with the labour market. OECD experience suggests that the existence of widely recognised skill certifications is important, although the role of government in setting such standards is variable; it can often be reduced once an effective set of non-government training bodies are in place. Some countries such as Germany and Korea have long traditions of apprenticeships co-managed by employers and business associations, which have contributed to high vocational skills and relatively low youth unemployment in these countries. Given Chile s heterogeneous labour market, efforts to promote human capital accumulation must be designed to target a broader cross-section of the labour market, including small entrepreneurs and their formal or informal employees. This could justify a greater element of direct subsidies, if this can be afforded, for example by extending the just-mentioned training grant for micro-firms to somewhat bigger firms, e.g. those with up to five or ten formal or informal employees. Other programmes under the Chilecalifica also merit expansion, for example those designed to promote ICT training in SMEs. The up-skilling and labour training programme (Nivelación de Estudios y Competencias Laborales) financed by FOSIS (the Solidarity Fund related to Chile Solidario, see Chapter 3) could be extended to target a broader client base than youths, and it should not necessarily exclude informal workers. 9. Special issues concerning employment of women As seen in Chapter 1, female labour force participation and employment rates are lower than in almost all OECD countries, even though they have increased substantially. The situation is most preoccupying for young women, whose employment rates remain remarkably low both by international standards and compared with young men in Chile (Figure 1.4). Women s entry into the labour market is often obstructed by a combination of family responsibilities and difficulties faced in making the transition from school to work. As mentioned in Chapter 1, this problem is more widespread among young women in Chile compared with most OECD countries because Chilean mothers tend to have their first child at a lower average age (less than 24 years), with a significant incidence of teenage pregnancies among those with relatively low education and low household incomes. The policy challenge is thus to help parents combine work and family responsibilities, a goal that will require more flexibility in enterprises and more child care and related measures. As discussed above, labour demand needs to be encouraged by a relaxation of EPL, which currently favours insiders too much at the expense of

93 96 CHAPTER 2. TOWARDS MORE EQUAL JOB OPPORTUNITIES outsiders, as well as a reduced minimum wage and more flexible regulations about working time and part-time work. Concerning working time, however, OECD experience shows that reforms to increase flexibility can have undesirable effects for parents unless they have access to affordable child care (OECD, 2006). The overall effect of flexible rules can still be positive because it stimulates higher production and higher labour demand. But to make it possible for parents to accept the additional jobs, such reforms should be accompanied by further measures to help them combine work with family duties, such as childcare support and parental leaves. Child care is often the most crucial condition for female labour force participation, especially for parents with small children. As seen above, the problem in Chile is greatest for low-skilled women, who often seem to regard their potential work incomes as insufficient to justify the cost of private childcare. Their participation in the labour force can nevertheless be advantageous in the long term, considering the potential for on-the-job learning and other personal development. Mothers with low education also show the highest incidence of part-time work in Chile, indicating that affordable child care could help many low-income mothers to overcome poverty by working longer hours. The following childcare provisions are currently found in Chile: For children aged less than 2, Chilean employers are obliged by law to provide childcare if they have 20 or more female employees. Firms often do this in their own premises or by paying for care elsewhere, but many enterprises probably seek to keep the number of employed women below the mentioned limit. About half of the total child-care provisions were reportedly private in 2003, including 17 percentage points representing facilities managed by employers (Politeia, 2007). These provisions were used predominantly by households in the two highest income quintiles. Public child care generally targets low-income groups. The Junta Nacional de Jardines Infantiles (JUNJI), under the Ministry of Education, expects to increase its day-care provision for from about places in 2006 to places in Its day-care centres target children aged 3 months to 4 years. JUNJI also sponsors some preschool facilities for 5-year-olds with part-time working mothers, mainly in rural areas, and to help integrate children from ethnic minorities ( An additional programme, Integra, targets families in extreme poverty. Municipal preschool education covers mainly 5- and 6-year-olds. The government has announced plans to expand the enrolment of

94 CHAPTER 2. TOWARDS MORE EQUAL JOB OPPORTUNITIES Conclusions 4-year-olds, which until now has concerned under 15% of the age class. In 2006, the number of pupils was of which some were aged 4 or less. This chapter has discussed a broad range of policy issues in the labour market area. The overall goal should be to make the labour market more dynamic and flexible and at the same to offer more equal opportunities. Chile s labour market institutions already permit a considerable degree of flexibility for business, but this has been achieved too much at the price of an excessive segmentation of the labour market, which contributes to a high level of inequality. Such market segmentation is neither fair nor economically efficient as it hampers competition for jobs and reduces incentives for human capital investment in disadvantaged groups. Chile therefore has good reason to pursue its efforts to improve the social dialogue and to develop a policy package that would combine more flexibility for business with better social protection of workers. There is also reason to eliminate many labour market rules that discriminate too much between workers with different contract arrangements. Recent proposals to encourage enterprise-level collective bargaining about flexible working conditions and to support training of the participants in such bargaining are also welcome. Much remains to be done in improving the government s capacity to implement effective labour market policies. Not only is there a case for higher budgetary allocations to labour market policy, insofar as this can be afforded; there is also a need for more investment in administrative networks and their human capital, especially at local level. This notably includes the following: On-going reforms to strengthen the Labour Inspectorate and labour courts should continue along with further efforts to eliminate the remaining pockets of informality and non-coverage of social insurance. The new legislation about sub-contracting should be enforced and its underlying principles should be clarified, if necessary by further legislation. Unemployment compensation should be extended to the groups that face high risk of unemployment, and its different treatment of various labour market groups should be reduced.

95 98 CHAPTER 2. TOWARDS MORE EQUAL JOB OPPORTUNITIES The public employment service needs to be enhanced as a complement to higher spending on unemployment insurance. The possibility of merging it with the SENCE should be considered. The policy efforts to develop the education and training system for lifelong learning must have high priority. The quality of enterprise training needs to be ensured, and further efforts are needed to provide training to the workers of the smallest firms. Public support of childcare merits further development in view of the large numbers of parents for whom it is a necessary condition for labour force participation.

96 CHAPTER 2. TOWARDS MORE EQUAL JOB OPPORTUNITIES 99 Annex 2.A1. The Presidential Advisory Commission Work and Equity This Advisory Commission, chaired by Prof. Patricio Meller Bock, delivered its report Towards a Fairer Chile: Work, Pay, Competitiveness and Social Equity in April Key proposals of interest to this OECD policy review include the following: 1. In-work benefits, child benefits and job-related training for vulnerable groups. 2. More generous unemployment insurance and better employment services. 3. Measures to develop industrial relations and collective bargaining. 4. Stipends to the best pupils in secondary education. 5. A strategy to formalise subsistence-level enterprises. Some potentially crucial issues proved so controversial that the Commission it could not make unanimous recommendations. Different groups of commissioners therefore gave separate opinions, This notably concerned a number of institutional questions, about which the Chilean society has been divided for a considerable time: The governance of public programmes for job-related training. Reducing the severance payments which employers must pay to dismissed workers. If adopted, this would come as a counterpart to the abovementioned proposal to improve to unemployment insurance (which would require higher employer contributions and state subsidies). Who should set the minimum wage. How to determine which trade unions are representative; at what levels and about what they should bargain. Whether or not to create a Council for Social and Economic Dialogue. As mentioned in the main text, the government presented a bill to Congress on 18 August 2008, which proposed a partial reform of the unemployment insurance (UI) along the lines of the Commission s proposal.

97 100 CHAPTER 2. TOWARDS MORE EQUAL JOB OPPORTUNITIES Annex 2.A2. Measuring employment protection legislation (EPL) by the OECD scoring method Scores 0-6 from lowest to highest strictness a) Procedural inconveniences Delay to start notice Criteria for justified dismissal 1. Termination of indefinite contracts b) Difficulty of dismissal c) Notice and severance pay for no-fault individual dismissal Trial period before a job is protected Procedures Compensation for unjustified dismissal after 20y Possibility of renstatement Notice period after Severance pay after 9m 4y 20y 9m 4y 20y Average for 1) United States United Kingdom Canada New Zealand Ireland Australia Switzerland Hungary Japan Denmark Czech Rep Korea Slovakia Chile Finland Poland Austria Netherlands Italy Germany Belgium Norway Sweden France Greece Spain Mexico Portugal Turkey Source: OECD (2004), OECD Employment Outlook, Chapter 2, Annex 2.A; Códogo del trabajo as amended until March 2008; Ministerio del Trabajo.

98 CHAPTER 2. TOWARDS MORE EQUAL JOB OPPORTUNITIES Regulation of temporary employment a) Fixed-term contracts b) Temporary-work agencies (TWA) Valid cases other than the usual objective reasons Maximum number of succes-sive contracts Maximum cumulated duration Types of work for which TWAs are allowed Restrictions on number of renewals Maximum cumulated duration of temporary contracts Average for 2) 3. Special rules for collective dismissal Total: Average for all listed forms of EPL* United States United Kingdom Canada New Zealand Ireland Australia Switzerland Hungary Japan Denmark Czech Rep Korea Slovakia Chile Finland Poland Austria Netherlands Italy Germany Belgium Norway Sweden France Greece Spain Mexico Portugal Turkey * Weighted average. Source: OECD (2004), OECD Employment Outlook, Chapter 2, Annex 2.A; Ministerio del Trabajo.

99 102 CHAPTER 2. TOWARDS MORE EQUAL JOB OPPORTUNITIES Bibliography Acevedo, G., P. Eskenazi and C. Pagés (2006), Unemployment Insurance in Chile: A New Model of Income Support, SP Discussion Paper No. 0612, World Bank, Washington D.C. Anderson, P.M. and B.D. Meyer (1998), Using a Natural Experiment to Estimate the Effects of the Unemployment Insurance Payroll Tax on Wages, Employment, Claims and Denials, NBER Working Paper No. 6808, Washington D.C. Congreso Nacional (2001), Ley del seguro de cesantía, Santiago (quoted in August 2008 from Congreso Nacional (2008), Código del trabajo (first adopted in 2001, with changes until March 2008), Santiago. Consejo Asesor Presidencial Trabajo y Equidad (2008), Hacia un Chile Más Justo: Trabajo, Salario, Competitividad y Equidad Social (two documents: Informe Final and Informe Ejecutivo), April, Santiago. Cowan, K., A. Micco, A. Mizala, C. Pagés and P. Romaguera (2003), Un Diagnóstico del Desempleo en Chile, IADB and Ministry of Finance, Santiago ( Cunningham, W.V. (2007), Minimum Wages and Social Policy: Lessons from Developing Countries, World Bank, Washington D.C. Dirección del Trabajo (2007), ENCLA 2006: Resultados de la Quinta Encuesta Laboral, Santiago. Dirección del Trabajo (2008), Compendio de Series Estadísticas, Santiago. Escobar, L.E. (forthcoming), El seguro de cesantía en Chile: diagnóstico y propuestas para fortalecerlo, Chile 21, Santiago. INE Instituto Nacional de Estadísticas (2006), Fecundidad en Chile: Situación reciente, Santiago. Larrañaga, O. (2009), Inequality, poverty and social policy. Recent Trends in Chile, OECD, Paris. Mideplan (2006), CASEN 2006: Enquesta de Caracterización Socioeconómica Nacional, Ministerio de Planificación, Santiago.

100 CHAPTER 2. TOWARDS MORE EQUAL JOB OPPORTUNITIES 103 Ministerio del Trabajo (2007), ENCLA 2006: Resultados de la Quinta Encuesta Laboral, Dirección del Trabajo, Santiago. Ministerio del Trabajo y Previsión Social (2006), Encuesta de Protección Social 2006: Presentacion General y Principales Resultados (EPS 2006), Santiago. OECD (2003, 2005, 2007), OECD Economic Surveys: Chile, OCDE, Paris. OECD (2004, 2006, 2008), OECD Employment Outlook, OECD, Paris. OECD (2004a), OECD Reviews of National Policies for Education: Chile, OECD, Paris. Politeia (2003), Salas Cuna y Jardines Infantiles para Hijos e Hijas de Mujeres Trabajadoras, Primer Informe: Diagnóstico Inicial y Base Metodológica del Estudio, Santiago. Puentes, E. (2007), Occupational Choice, Wage Distribution and Technological Changes in a General Equilibrium Model. Evidence from Chile: , University of Chicago. SOFOFA (2008), Informe de remuneraciones y materias laborales de la industria, Diciembre 2007, Santiago ( accessed 31 January 2009). World Bank (2007), Doing Business 2008: Chile, Washington D.C.

101

102 CHAPTER 3. REDUCING POVERTY IN THE WORKING-AGE POPULATION 105 Chapter 3. Reducing Poverty in the Working-age Population Strong growth in economic prosperity has led to increased demands for a more inclusive society, with a more equal distribution of income and opportunities. So far, Chilean social policies for the working-age population have focused on developing health, housing and education policies (pension policies are discussed elsewhere). Using a combination of home visits and questionnaires, Chilean policy has been able to identify many of those in need and target support to low-income households. However, gaps in coverage remain, especially in rural areas. Other challenges remain, not least extending the local capacity to deliver services. Concerns about the quality of services have also been raised. This chapter gives a summary of education, health and housing policies, and analyses social policies, in view of the public resources put into them, their redistributive power, and their ability to identify clients and successfully target policies at those who are most urgently in need of support. The chapter looks at how the holistic approach of social service delivery in the Chile Solidario package helps clients to achieve better outcomes. The chapter concludes by looking ahead and discussing future Chilean social policy development, in particular the emerging issue of how to design and introduce a system of financial transfers to working-age families without weakening financial incentives to work.

103 106 CHAPTER 3. REDUCING POVERTY IN THE WORKING-AGE POPULATION 1. Introduction Reducing poverty has been one of the main priorities of the Chilean government since When considered against the absolute poverty threshold of 50% of the median income in 1996 as adjusted for inflation, poverty reduction has been impressive since At just over 10% in 2006, the poverty rate is only one-third of what it was in 1990 (Table 3.1). However, when measured against the relative poverty line of 50% of median income in respective years, poverty rates remained stable during the 1990s and only decreased between the years 2000 and Table 3.1. Poverty and its severity has fallen since 1990 Poverty indicators, Percentage of poverty Absolute line Relative line Poverty gap as a % of GDP, absolute line ,3 20,5 4, ,7 20,7 2, ,9 20,9 1, ,6 16,4 0,9 For detail on poverty definitions, see Chapter 1. Source: Larrañaga (2009). The trend in absolute poverty better reflects the decline in the severity of financial hardship in Chile (see also Chapter 1). One indicator on the depth of poverty is the percentage of GDP that would need to be spent on poor households for them to reach the absolute poverty line. Abstracting from costs of targeting and adverse financial incentive effects, it is estimated that providing all poor households with sufficient cash transfers to reach the poverty line would have cost 4.6% of GDP in 1990, while only 0.9% of GDP would have been needed in 2006 (Larrañaga, 2009). Strong economic growth in recent years has reduced absolute poverty rates and increased average incomes of lower and middle income groups, but as incomes of the highest income groups increased at almost the same pace, there has been only a small decline in income disparities which remain large compared with most OECD countries (Chapter 1). In recent years, demands have been growing to develop policies which make growth more inclusive and which reduce both poverty and income inequality. Within a strict budgetary environment with little redistribution of resources among income groups (Box 3.1), the social policy response has been to first develop key services such as housing, health and education, but is now also focussing on increasing employment opportunities and providing financial support to low-income groups of working age.

104 CHAPTER 3. REDUCING POVERTY IN THE WORKING-AGE POPULATION 107 Box 3.1. Public financing of social support in Chile: budgetary prudence and little redistribution in taxation Chilean fiscal policy in recent years has been prudent and transparent. In 2001, the structural budget surplus rule was introduced which stipulated a budgetary surplus of 1% of GDP net of the effects of cyclical copper-price fluctuations. In 2006, the structural budget rule was enshrined in the Fiscal Responsibility Law which also introduced rules on capitalisation of the Chilean Central Bank, the provision of public job-creation programmes when regional unemployment rates are in excess of 10%, and a Pension Reserve Fund. This fund is financed by between 0.2% to 0.5% of the previous year s GDP. Fund assets can only be used to finance (means-tested) pension benefits from 2016 onwards (OECD, 2007a). Any surplus beyond these three institutions will be placed in a Fund for Economic and Social Stabilisation. In view of the recent decline in public indebtedness and savings for future pension liabilities, the structural budget target was relaxed from 1% of GDP to 0.5% for 2008 which facilitated an increase in public spending, including on education (OECD, 2007a). The tax burden is not heavy in Chile; at about 17% of GDP it is well below the OECD average of around 37% (Figure below, Panel A). However, employees in Chile make mandatory contributions towards their retirement, disability and survivors pensions to private funds to the tune of about 12.6% of earnings (Chapter 4). In comparison, mandatory contributions to Superannuation pension funds in Australia concern 9% of earnings. These mandatory contributions are not counted as social security contributions in the OECD Revenue Statistics (OECD, 2007b) and to that extent the Chilean tax-to- GDP ratio is artificially low. In 2005, the value of mandatory pension contributions amounted to 5% of GDP in Chile (Chapter 4); by comparison social security receipts amount to 9.2% of GDP on average across the OECD in the same year (OECD, 2007b). The Chilean tax burden is not heavy 1 and VAT and excises make up half of central government tax revenue Panel A. Tax-to-GDP ratios in Chile and OECD countries, Without social security contributions With social security contributions

105 108 CHAPTER 3. REDUCING POVERTY IN THE WORKING-AGE POPULATION Panel B. Chilean central government tax revenue as a percentage of GDP, Other taxes Excises VAT Income tax 1. Local tax revenue is not included, and the tax burden does not include mandatory contributions to private (pension) funds. Source: OECD Revenue Statistics and Servicio de Impuestos Internos 2008 ( Another important feature of the Chilean tax system is its reliance on value-added tax (VAT) and other taxes on goods and services (including excises) in the tax mix. Because of the high copper revenues, the proportion of indirect taxes in central government tax revenue has declined from 60% during the 1990s to just over 50% in 2006 (Figure above, Panel B). This was much higher than the 32% OECD average; only Mexico and Turkey are similarly reliant on indirect taxation (OECD, 2007b). Personal income tax amounted to 30% of tax receipts in Chile in 2006, compared with an OECD average of 25%. The income tax system has 6 rates from 0 to 40% (Cantallops et al., 2007). Since the corporate tax rate is only 17%, there are incentives for high-income taxpayers to incorporate themselves and opt for (partial) own-account employment. The relative simplicity of operating value-added taxation (a VAT rate of 19% with few exceptions) can help compliance with tax collection (see Keen and Smith, 2007 for an overview of VAT-evasion issues). Registered taxpayers can obtain a credit for purchasing immediate goods and services if acquired from another registered taxpayer, so the system provides financial incentives to comply with the rules. With administration costs less than 1% of net revenue, the efficiency of the Chilean tax administration compares well with that in most of the OECD countries (OECD, 2007b). While proceedings can be further streamlined and enforcement improved it is estimated that the combination of a relatively simple system and increased enforcement has brought down the estimated proportion of non-reported VAT from about 20% in 2000 to 11% in 2005 (OECD, 2007b). Compared with the estimated VAT-evasion rate of around 5 percentage points or below in Denmark, Germany, the Netherlands and the United Kingdom, Chile can still make considerable improvements in tax collection, but it already compares favorably with Greece, Spain and Italy where estimated VAT evasion rates are in excess of 20% (Nam et al., 2003). A drawback of reliance on VAT is that it does not redistribute spending power. Indeed, comparisons of income distributions before and after taxation show that the overall effect of the tax system on the income distribution in Chile is, in fact, slightly regressive (Engle et al., 1998; Cantallops et al., 2007). Moreover, simulations of tax reform indicate that a steeper progressivity would have limited effect on the income redistribution, reflecting the relatively small overall income tax revenue. Taxing top-income groups more effectively and focusing benefits more on low-income groups of the working-age population has greater potential to generate a more equal distribution of resources and opportunities in Chile.

106 CHAPTER 3. REDUCING POVERTY IN THE WORKING-AGE POPULATION Social policy Overall spending and redistributive effect In the Chilean context, social policy consists of four key areas: education, health, pensions, and housing, with targeted income and family support for low-income households as an overarching objective cutting across these themes. Education, health and to a lesser extent housing policies are not among the core themes of this labour market and social policy review. Nevertheless, any discussion of Chilean social policy would be incomplete without some information about policy developments and concerns in these areas. Private arrangements for pensions and other social protection will be discussed in the next chapter. Regardless of whether or not education is counted, Chile ranks low in an OECD comparison of public social spending relative to GDP (Figure 3.1, Panel A). With public social expenditure at around 8.7% in 2007, Chile only spends more than Korea and Mexico among the OECD countries. Including spending on pension benefits and annuities that derive from mandatory pension contributions, publicly-mandated social spending (including benefits derived from mandatory pension contributions) amounted to just over 10% in 2005, about half the OECD average. Public social spending and spending on education in Chile amounted to almost 12% of GDP in 2007 (Figure 3.1, Panel B); almost two-thirds of total public spending (18.7% of GDP in 2007; MdH, 2007). Education and health accounted for around 3% of GDP each, but despite the development of a mandatory private pension system since the early 1980s, at 4.1% of GDP in 2007 public spending on pensions is the largest item of social spending. Such expenditure includes spending on benefit-entitlements established prior to the pension reform in the early 1980s, pensions for the armed forces and minimum pensions (Chapter 4). Compared with spending on education, health and retirement support, public spending on employment support, family support, and housing support is limited. Public spending on education and health has increased as a percentage of GDP since the mid-1990s, while spending on retirement support has increased only gradually and declined in proportional terms (Figure 3.1, Panel B). On the whole, public spending on social policy and education has been increasing in real terms (see Panel C), but it has not kept pace with GDP growth in recent years (except in the health area with public spending being 2.8% of GDP both in 2000 and 2006). The recent change in the structural budget rule (Box 3.1) allowed for further increases in public spending in 2008, including on education, which in 2006 amounted to 3.1% of GDP (by comparison the OECD average was 5% of GDP in 2004, OECD, 2007d).

107 110 CHAPTER 3. REDUCING POVERTY IN THE WORKING-AGE POPULATION 30.0 Figure 3.1. Public social and education spending has increased in Chile, but remains low in international comparison Panel A. Publicly-mandated social spending to GDP ratios in Chile and OECD countries, Panel B. Central government spending on social purposes and education as a percentage of GDP, Other Housing Family support Public pensions Education Health

108 CHAPTER 3. REDUCING POVERTY IN THE WORKING-AGE POPULATION 111 Figure 3.1. Public social and education spending has increased in Chile, but remains low in international comparison (cont d) Panel C. Public spending on social policy and education in real terms, (1990 = 100) Education Health Total social and education spending Non-health (mainly retirement support) 1. Publicly-mandated social spending includes both public social spending (which includes spending on health, but not on education) and mandatory private social spending (for a comprehensive overview of definitions, see OECD, 2007c). For Chile, such spending includes expenditures on pension benefits and outlays by pension funds on annuities for clients (worth about 1% of GDP in 2005), see Chapter 5. If only current outlays on pension benefits and annuity payouts (in a given year) were considered, mandatory spending would be about 0.1% of GDP. Source: OECD (2008), Social Expenditure database, , Larrañaga (2009), and Ministerio de Hacienda (2008), Estadisticas de las Finanzas Publicas; Spending on family supports includes services for (orphaned) youth, income-tested family benefits and allowances, and maternity payments. 35 Family allowances are paid almost completely out of general revenue by different social protection institutes. Payment rates are low, at about CLP per month at maximum (about USD 12), and phased out on income of CLP , where the payment is CLP These benefits towards dependents are nevertheless very important in the Chilean context as they are paid to about 2.7 million dependents (mostly children, but also 35. Maternity pay is available to formal-sector employees (about 70% of female employees have a signed employment contract and are paying social security contributions) who have contributed continuously for three months (or the equivalent over a six-month period) and is paid at 100% of the average of their last three months earnings up to a maximum of UF 60 (different rules pertain to the self-employed). Maternity allowance can be payable for a period from six weeks prior to childbirth and 84 days after childbirth. The associated allowance for a sick child can be paid up to the first birthday of the child.

109 112 CHAPTER 3. REDUCING POVERTY IN THE WORKING-AGE POPULATION other dependent family members, including widowed mothers, disabled spouses, students and/or grandchildren). The Chilean benefit system is relatively good at targeting the poor, with almost 75% of public social expenditure and public spending on education going to households from the four lowest income deciles (Table 3.2, Panels A and B). Because of the value of education and health support (and to a lesser extent, the minimum pension payment), total disposable income of households in the lowest income decile is almost three times the income they generated autonomously in Table 3.2 also shows that while the value of individual social benefits such as family payments, 36 the Chile Solidario participation payment, and drinking water subsidies may be low, they constitute a significant part of income in poor households. It appears that social policies in a broad sense achieve a significant redistribution of resources. For example, in 2006, net market income (see note 1 to Table 3.2) of the 10% of richest households was 31 times higher than for the 10% of poorest households, but only 12 times higher if their consumption of publicly-financed social services, healthcare and education is taken into account (Table 3.2). The redistributive power of benefit systems not only depends on the extent to which spending is targeted at low-income groups (rather than on those who contributed most, as is the case in systems based on insurance principles). It also depends on the overall size of public spending on social benefits. As mentioned earlier, public social spending in Chile is low in international comparison, but, with the planned introduction of cash transfers supporting working-age households and families (see below section Further developing supports to the working-age population ) the redistributive power of the Chilean benefit system will be strengthened significantly The family allowance (INP) is paid to workers in the formal sector, and as the poorest are most likely to be in informal employment, this payment is seemingly less important to households in the lowest income decile than the family allowance paid under the Chile Solidario regime (see Section on An integrated approach towards supporting the poor: Chile Solidario ). 37. Similarly, the degree of targeting public family benefits on poor families in the United States is the strongest across the OECD. However, since overall outlays on such programmes are small in international comparison, the overall redistributive impact is limited (Whiteford and Adema, 2007).

110 CHAPTER 3. REDUCING POVERTY IN THE WORKING-AGE POPULATION 113 Table 3.2. Public spending on education and health is the most important in redistributing resources towards the poor in Chile Panel A. Monthly value of household income and subsidies by deciles, 2006 (Chilean pesos, Nov. 2006) Income decile Income and support items I II III IV V VI VII VIII IX X Total Net market income Education subsidies Health subsidies Total social benefits Of which Minimum pension (PASIS) Family allowance (SUF) Chile Solidario bonus Drinking water subsidy Family allowance (INP) Unemployment compensation Total disposable income Total disposable income without social benefits Source: Adapted from Mideplan (2007). Panel B. Monthly value of household income and subsidies by deciles, 2006 (proportion of income and subsidy items) Income decile Income and support items I II III IV V VI VII VIII IX X Total Net market income Education subsidies Health subsidies Total social benefits Of which Minimum pension (PASIS) Family allowance (SUF) Chile Solidario bonus Drinking water subsidy Family allowance (INP) Unemployment compensation Total disposable income Total disposable incomewithout social benefits Net market income refers to household income including labour income, capital income (rents and interests), benefit payments from contributory pensions, and other private income (as, for example, transfers from relatives who do not live in the household). Income is measured after direct taxes and social security contributions (not including mandatory contributions to private pension funds), but before indirect taxes. 2. Education subsidies include the following subsidies: supports towards school supplies, text books, meals, (dental) health, as well as contributions to national education services (as, for example, the national council of school assistance and scholarships, or the pre-school supervising agency JUNJI). The monthly education subsidies are imputed from the values reported by households in the CASEN survey and the information provided by the Ministry of Education and its agencies. Health subsidies include all health benefits delivered or financed by the public sector. These subsidies are net of co-payments made by the beneficiary and contributions made by workers. The monthly health subsidies are imputed using information from three different sources: CASEN survey, and the financial statistics of the Ministry of Health as well as the Ministry of Finance. Source: Adapted from Mideplan (2007).

111 114 CHAPTER 3. REDUCING POVERTY IN THE WORKING-AGE POPULATION Education, health and housing policies Chilean social policy development since the mid-1990s includes two key elements. First, since the 1990s, policy intensified in the areas of education, health and housing policy, while prevailing budget constraints led to much prioritisation and targeting on the poor. Second, Chilean policies generally leave room for private-sector involvement in the provision of many social services, reflecting a generally strong belief in the powers of market competition to generate efficiency. The internationally most well-known example of the latter is pension policy (Chapter 4), but housing, health and education are also affected. Education As discussed in Chapter 1, Chile scores well below the OECD average in terms of educational attainment of the population and literacy rates for young people. Public spending on education is well below the OECD average (Figure 3.2). Recent Chilean governments have recognised the need for more human capital investment (and increased spending on education, Box 3.1). There is now nearly 100% coverage of the population and the school infrastructure has been improved. 38 After eight years of compulsory primary schooling (and about two-thirds of 4- and 5-year olds attending non-compulsory pre-school), about 87% of the year age group participates in secondary education. The proportion of youths aged 20 who are in higher education is about 34%, comparable with participation numbers often seen in OECD countries (see the detailed review of tertiary education in Chile available as of April 2009). Despite the significant investment, the Chilean education system suffers from a number of quality issues. Quality issues include a lack of technical/vocational focus in education. Strengthening linkages between education and enterprises (e.g. through apprenticeships and extending vocational secondary education in industrial fields), which improve the school-to-work transition and raise the very low youth employment rate in Chile are key policy concerns (OECD, 2004). The Chile Califica programme aims to develop a lifelong learning and training system with participation of the private sector, as there is too little capacity for catch-up education of adults with low general schooling. In general and across levels of education, the improvement in quality is lagging behind the recent increase in capacity. 38. Full-day schooling was introduced in 1997 for municipal and subsidised private schools. Until then, many classrooms were shared between morning and afternoon classes.

112 CHAPTER 3. REDUCING POVERTY IN THE WORKING-AGE POPULATION 115 Figure 3.2. Public investment in health and education in Chile is limited compared with OECD countries Panel A. Public spending on educational institutions as a percentage of GDP, 2004/05 Other educational institutions, including pre-schools Tertiary education Primary and secondary education Source: OECD (2007d), Education at a Glance. Panel B. Public expenditure on health as a percentage of GDP, Source: OECD Health Data 2008 and Ministerio de Hacienda (2007), Estadisticas de las Finanzas Publicas (

113 116 CHAPTER 3. REDUCING POVERTY IN THE WORKING-AGE POPULATION The Chilean education system was decentralised in the 1980s. Local authorities are responsible for staffing and maintaining school infrastructure, while the Ministry of Education has regulatory, supervisory and co-ordinating functions, including the setting of guidelines and curricula. There are three types of schools: municipal, subsidised private and fully fee-financed private (for an overview see OECD, 2004 and 2007a). Both municipal schools (which can receive additional municipal subsidies) and subsidised private schools are financed by an implicit voucher system. Central government makes a transfer to local authorities which then pay each school in their areas a certain amount per child depending on the level of education and daily participation (full-time or otherwise) so long as the parental co-payment to the school is within a set limit. Parents choose the school their children attend, thereby determining to which school the local authorities should make the payments. The voucher system helped to expand capacity within a short period of time, but otherwise, the effect of competition on school performance has been limited and there are considerable equity issues. In theory, funding follows children, but the system allows local authorities to cross-subsidise across schools in their jurisdiction. This provides an additional incentive to select students for (private) schools that wish to improve their performance. Selection is in theory illegal, but in practice it is widespread. The fee-based private schools select most obviously, though they must use 15% of their fee receipts to finance scholarships for low-income students. As from 2008, the Chilean authorities will be rolling out a differentiated voucher scheme in pre-school, primary and lower secondary education, which pays more to schools for children from low-income families (for identification purpose, the authorities use the Social Protection Index based on the Ficha de Protección Social, see the discussion of Chile Solidario below and information provided by health authorities). Given the different student populations, municipal schools are set to make the biggest financial gains. In order to improve quality, the Chilean system includes a Subvencion a los Establecemientos de Major Desempeno, a subsidy to generate excellence in performance. The National Education Evaluation Service (SNED) considers school performance in 6 dimensions: efficiency; improvement over time (both these factors are co-determined by pupils national SIMCE test scores); initiative (new projects); employment conditions; participation of vulnerable groups; and, participation by teachers and parents. The programme has operated since 1995 and in 2008/09 will be extended to assistant teachers. Awards are made to the group of teachers at a given school. Since payments are small (USD 6.5 per child for a total budget of 36 million pesos or USD for 2008), they are likely to have had only a limited effect on quality improvement.

114 CHAPTER 3. REDUCING POVERTY IN THE WORKING-AGE POPULATION 117 Table 3.2 illustrated that public support for education (and health) has a significant effect on the income distribution. Education also affects the (future) income distribution through human capital accumulation, and public health and education investment since the 1960s has contributed, if indeed it was not a precondition (Larrañaga, 2009), to economic growth following the liberalisation which took place after At the same time, access conditions to education and training opportunities for adults remain different across income groups. Higher income groups in Chile acquire the best quality and most expensive education in the private market; middleincome groups use public services, but may be able to get a better quality service in exchange for a co-payment ; lower income groups attend free public education (Larrañaga, 2009). 39 In this sense, the segmentation of the education market limits equality of opportunities and future reductions in income inequality. Health Some health outcomes in Chile, such as, for example, life expectancy (78 years for men and 81 for women) are quite close to the OECD average. Infant mortality rates are somewhat worse than the OECD average (Chapter 1). The 2008 Chile Crece Contigo programme (see below) is geared towards further improving child health outcomes. In view of the relatively low public health spending effort in Chile compared with most OECD countries (Figure 3.2, Panel B), health outcomes in Chile hold up relatively well. Nevertheless, much remains to be done to reduce socio-economic, territorial and ethnic health inequalities. The Chilean health system is a mix of public and private insurance, as well as care provisions. The National Health Fund (Fondo Nacional de Salud, FONASA) provides free insurance coverage for low-income households. Employees pay mandatory health contributions of 7% of earnings to FONASA or to a private health provider (Instituciones de Salud Previsional, ISAPRE), which may charge additional contributions depending on age, sex, family size and the type of plan. There are about plans with different prices and services. Many ISAPRA clients contribute about 9-10% of earnings. 39. In primary and secondary education, high-income students (10% of total students) attend private fee-paying schools, and get the best education as indicated by SIMCE test scores; 40% of students (mostly from the middle-class) attend private subsidised schools and get intermediate educational results; meanwhile the remaining 50% of students (mostly from low-income households) attend free public schools and get low results in the same school performance tests.

115 118 CHAPTER 3. REDUCING POVERTY IN THE WORKING-AGE POPULATION In 2007, FONASA clients accounted for just over 70% of the population and ISAPRE affiliates for almost 17%. The other 13% were covered by not-for-profit mutual plans or had no coverage (about 6-7% of the population, OECD, 2007a). With three-quarters of FONASA clients being treated in public facilities, and ISAPRE clients mostly using private healthcare facilities, the Chilean healthcare market is segmented. The current government may wish to merge the two systems into a single system with more redistribution of resources among income groups, but so far this has proved politically impossible. Private insurance plans do not cover all pathologies, but the number of covered conditions has tended to increase. They cover a broader range of care services and a wider choice of providers, and are perceived to offer better quality. The privately insured can use public institutions, but most of them opt for private facilities where co-payments are required, depending on each insurance plan and healthcare facility. Apart from price strategies, ISAPREs can reject applications through a selection mechanism. Hence, it is no surprise that ISAPREs tend to cover well-to-do groups, often of working age with limited health problems. Since 2002, AUGE (Acceso Universal con Garantías Explícitas en Salud) supports household expenditure on health care, especially for low-income households. AUGE provides guaranteed care for a list of about 70 selected diseases and is set to cover about 80 pathologies by For each of these, AUGE establishes an entitlement to treatment for FONASA and ISAPRE clients, for which a co-payment of up to 20% can be charged, though not for low-income individuals. If patients are not treated in a prescribed period of time, public authorities have to pay for private treatment. AUGE also intends to improve the quality of public services (there are considerable waiting lists and absenteeism of medical personnel in public hospitals and clinics curtails productivity) through a certification mechanism for medical facilities. Comprehensive evaluations on its effect on efficiency are not yet available, but at least in theory AUGE provides a link between supply and recipients and increases accountability of public health providers. AUGE could also contribute to increasing the role of disease-based budgeting, rather than retrospective budgeting on basis of past pay-rolls and fees paid for medical inputs and materials. The efficiency of the Chilean health system could be further improved. Risk is unduly concentrated in the public insurance system since ISAPREs can select individuals through price discrimination, restricted coverage and by simply refusing to accept a client. One option would be to extend access to the Solidarity Compensation Fund (currently only available to private insurers) to FONASA, thereby increasing risk-pooling. Another avenue to more intensive competition (and efficiency) is to equalise the level of

116 CHAPTER 3. REDUCING POVERTY IN THE WORKING-AGE POPULATION 119 co-payments for both ISAPRE and FONASA beneficiaries for homogeneous services provided by public and private providers (OECD, 2007a). ISAPREs are free to change the cost and coverage of the plans annually, though since 2005 price increases have been capped. To further improve quality, a system of health workforce certification is being established, including a registry of health professionals (at present it is not known how many certified doctors and nurses there are in Chile). Public and private hospitals will also be regularly evaluated, with authorisation of operations subject to review and results will be made available to the public. Housing Chilean housing policy has been rather effective in reducing capacity problems. Through direct construction and relocations governments in the 1990s made a concerted effort (Chile Barrios) to do away with shantytowns (campamentos). This has had concrete results: of the 950 shantytowns identified in 1990, almost all have disappeared, although some smaller relatively new ones have appeared in which about families live. Different housing programmes financed the construction of new housing. There is still a housing deficit, but with a housing stock of about 4.5 million dwellings (up from 2.5 million in 1990) and 0.5 million households in need of better homes (down from over 1 million in 1990), progress has been substantial (Navarro, 2005). At a national urban level, 71.2% of housing is owner-occupied, and just under half of those dwellings have been acquired through public housing programmes. Targeting exists but is not particularly tight: the share of owner-occupied housing in urban areas that used public subsidies for purchase is over one-third from the first to the fourth income quintiles, and one-sixth for the highest quintile (Larrañaga, 2009). As in other policy areas, the focus in policy development has moved towards quality issues. Past housing policies have led to the construction of a very large number of houses that were very cheap. The quality problems regard materials, designs and neighbourhood conditions. Current policy initiatives shift focus from houses to neighbourhoods to ensure they can support sufficient public and private service provision. Policy also aims to stimulate client involvement (frequently from the planning stage onwards), and engage people who receive housing support in the upkeep of their dwelling; low-value houses bought with financial support from the authorities can now also be sold within the initial five years of acquisition, giving owners an incentive to engage in maintenance while contributing to the development of a low-income housing market. Current housing programmes do not involve the direct construction or the provision of mortgages, but instead provide subsidies to different groups

117 120 CHAPTER 3. REDUCING POVERTY IN THE WORKING-AGE POPULATION of clients. The system is not intended to support those who cannot save, as otherwise clients will not be able to pay off private sector housing loans. Most of the support towards building new homes concerns households with incomes somewhere between poor households and those with average incomes. In terms of public support which has been around 0.8% of GDP over the past ten years (Figure 3.1, Panel B), the most important housing supports in 2007 were (MINVU, 2008a and 2008b): At almost 60% of all public housing outlays in 2007, the Fondo Solidario de Vivienda (Solidarity Housing Fund) which makes grants and loans available is the largest public housing programme. Clients have to have completed the Social Protection Record (see below), 40 and the minimum deposit by clients is UF 10 (about USD 400), 41 and the maximum subsidy amount is UF 470 for construction of new dwellings and UF 420 for the adjustment of existing houses (amounts vary across regions). The subsidy covers the cost of land, infrastructure and a unit of 38 m 2 including a bathroom, kitchen, bedroom and at least one other room. The innovative aspect of the policy lies in its collective nature: families design their own projects and must apply in organised groups of at least ten and at most 300 households, and with the (technical) support of a managing organisation who also manages the finances [a local authority, community group (or other NGO), or a specialised firm registered with the ministry]. The ministry adjudicates proposals in view of social and urban development objectives, as well as design features. The Sistema de Subsidio Habitacional (housing subsidies worth about 15% of public housing outlays) towards buying or building a new house: the minimum deposit is UF 50, the maximum house-price UF (or USD ) and the maximum subsidy (which depends on the house price), ranges from UF 90 to Applications are also assessed in view of the precariousness of current housing arrangements, the space available per person, water access and type of flooring. Excluded are those who own or whose spouse owns a house. 41. The Unidad de Fomento (UF) is the main measure for determining the cost of construction, values of housing and secured loans: it is a unit of account which adjusts the value of the Chilean peso on a daily basis ( Here we use the average value for June 2008: UF 1 was CLP or USD Depending on the characteristics of the house (e.g. if it located in an urban development zone or if it concerns a historical building), the accepted house price can double to

118 CHAPTER 3. REDUCING POVERTY IN THE WORKING-AGE POPULATION 121 The Subsidio Habitad Rural (rural housing subsidies) constitutes just below 10% of public housing outlays. It is targeted at poor families requiring a minimum deposit of UF 10 with a maximum subsidy ranging from UF 220 to 300. Chief quality criteria are location, access to day care, playgrounds and community services, and one way of addressing such issues is to build additional houses in existing neighbourhoods that already have good public and private services. Social housing should not be extended at the margins of cities, but as an integral part of urban development. Policy aspires to accomplish a greater social mix with subsidies towards housing projects providing a better mix in neighbourhoods. An integrated approach towards supporting the poor: Chile Solidario The introduction to this chapter showed that the decline in poverty came to a temporary halt in the second part of the 1990s. In response, Chile Solidario was introduced in 2002 to promote equity and enhance opportunities. It is a targeted anti-poverty instrument, within the overall fabric of education, health and pension supports. In contrast to past efforts, Chile Solidario is a regular social protection programme granting entitlements to the poor (rather than support being dependent and rationed on the basis of budget constraints). Its key client groups are low-earning couples and/or sole-parent families with children. Chile Solidario organises new and old programmes for the poor as a co-ordinated package, including a range of possible services such as healthcare, childcare, education, income and other family- support services. Thus far, cash benefits have played only a limited role. The key to Chile Solidario is the provision of sustained and regular psycho-social support, whereby social workers help clients to assert their rights to a range of potentially available services. From the government s perspective, this approach towards poverty requires correct identification of clients, sufficient service capacity and co-ordination of the different agencies involved. This latter role is assigned to the Planning and Co-operation Ministry Mideplan. Programme-related transfers to agencies and local administrations can be made dependent on their co-operation (Box 3.2). UF 2 000, with maximum subsidies paid through the Sistema de Subsidio Habitacional adjusted accordingly.

119 122 CHAPTER 3. REDUCING POVERTY IN THE WORKING-AGE POPULATION Box 3.2. Anti-poverty policy development and coherence in Chile There are many different ministries involved in Social Policy in Chile: Planning (Mideplan), Labour (which supervises insurance-based arrangements for formal sector employees), Housing, Education, Health, Culture, and the National Youth Service (SENAME). Raczynski and Serrano (2005) identified about 400 support programmes that were created by successive governments since 1990 and which involved almost 80 agencies, and included: the Chile Joven programme aimed at youth without jobs; Chile Barrios targeted at shantytowns (see the housing section); the Origenes programme targeted at the aboriginal population; the Mujeres Jefes de Hogar programme aimed at female heads of households; the local funding programmes of the Fondo de Solidaridad e Inversión Social (FOSIS, see below); and, the Chile Solidario initiative. The Mideplan Minister chairs the social cabinet committee with ministers responsible for these different agencies develops new measures and implements some social protection schemes, including Chile Solidario and Chile Crece Contigo. Mideplan will also take the lead in formulating new antipoverty measures in late 2008, following-up on the findings of the Presidential Advisory Council on Work and Equity. This practice is becoming a trend in Chilean policy development: the current pension reform (Chapter 4) was developed on the basis of findings by an earlier report of a publiclyappointed group of pension experts. Mideplan is also responsible for the national socio-economic categorisation survey (Encuesta de Caracterizacíon Socioeconómica Nacional CASEN) which has been held every three years since 2000 (bi-annual from 1990 to 2000); there also is a panel study which has had three waves so far (1996, 2001, and 2006). CASEN is used to identify demographic, health, education, housing, earnings and income characteristics of the population, and evaluate the impact of government policy on the social-economic situation of the population. CASEN has a large sample size (about households in 2006) and plays a key role in identifying poverty concentrations, to help identify regions/population groups where policy support is needed most. Under the aegis of Mideplan, the Fund for Solidarity and Social Investment (FOSIS) was established in 1990; it aims to fight poverty by strengthening the productive capacity of the poor. FOSIS has about 800 staffers; it runs about 13 programmes serving about clients each year. Broadly speaking FOSIS has two broad groups of programmes: i) programmes which promote self-sufficiency; for example, through the promotion of entrepreneurship, including micro-enterprise employment; and ii) programmes to promote social participation including, the Puente or bridge programme (see main text), but also, for example, initiatives for early childhood support, support for families with children 9 to 14 years of age, support for families of prisoners, and programmes against domestic violence. The broad-based approach towards delivery of a wide range of services (health, education, etc.), as in the Puente programme, involves many different departments and public agencies, which requires intensive co-ordination at central and local levels of administration. As a result, Chile Solidario and FOSIS staff have to manage different actors, networks, and programmes against the backdrop of different institutional cultures and budget rules. They may hold some sway over participating agencies by linking payments to their co-operation, but this is more difficult when it concerns larger agencies with their own substantial budgets. The World Bank identified institutional modernisation and training of central and local government staff, as well as social workers, as crucial to both Mideplan being successful in carrying out its co-ordinating role, as well the successful implementation of Chile Solidario in practical terms (World Bank, 2008). Indeed, overall anti-poverty policy coherence much depends on the ability of Chile Solidario and FOSIS staff and other local social workers to develop and improve interfaces between the different agencies involved.

120 CHAPTER 3. REDUCING POVERTY IN THE WORKING-AGE POPULATION 123 From the client perspective, Chile Solidario acts as the gateway to a range of networks and services (and some financial support), and helps them access the needed services. Client s active participation is key to success, and the contract of participation is signed by family workers and clients committing themselves to relevant actions, as for example, participating in medical check-ups, vaccinations for their children, ensure that their children stay in education, etc. The principle of mutual obligations in social policy is not unknown in OECD countries. On the one hand, government provides income support and services (for example, education, (vocational) training, a work placement programme, or childcare subsidies) to people without a job. In return for making the income transfer to the client, the government requires him or her to make use of the support services, so as to increase the client s labour market opportunities and household self-sufficiency. For example, for young people to receive income support they are often required to be in education, or participate in training and/or on a workplacement programme or be actively engaged in looking for a job. Similarly, while providing childcare supports, most OECD countries require sole parents whose youngest child is of primary school-age to look for work or participate in training programmes to enhance their labour market chances and reduce the poverty risk for these families (OECD, 2007e). By contrast, in Chile, conditionality in benefit receipt does not involve large sums unlike the well-known Mexican Oportunidades programme (Box 3.3): the payment for participating in Chile Solidario ranges from USD 10 to USD 24 per month (see below). 43 The Chilean variant of mutual obligations in benefit receipt ties participation to access to services, not to receipt of financial transfers on a comprehensive basis. 43. As the participation-payment is so low, the effect of the conditional cash-transfers component of the Chile Solidario programme (and its predecessor) on reducing income inequality is very small: a 0.1% decline in the Gini coefficient over the period. By comparison, Oportunidades is estimated to have contributed to a decline of income inequality of 2.7 points in the Gini coefficient between 1996 and 2004 in Mexico (Soares et al., 2007).

121 124 CHAPTER 3. REDUCING POVERTY IN THE WORKING-AGE POPULATION Box 3.3. Conditional family support in Mexico and holistic service delivery in some other OECD countries Conditional Cash Transfers (CCT) are credited with a positive effects on schooling, health, child labour, and poverty (e.g. Skoufias, 2000; Rawlings, 2004; Handa and Davies, 2006). However, as most of these policies are relatively new, the full impact will only be observed in the medium and long term. The Mexican Oportunidades programme was introduced in 1997 as Progresa. It aims to reduce malnutrition, promote regular health checks and enhance participation in education ( Benefits to poor families are conditional on children attending school and undertaking regular health checks. In 2007, the programme covered about 5 million families, around 25% of the total national households with coverage about 75% among households in rural areas (INSP, 2006), with an annual budget of over 36 billion pesos in 2007 (about 0.4% of GDP). Cash subsidies are paid periodically (to mothers rather than fathers) and can reach 22% of the recipients total income; education grants are slightly higher for girls than for boys (about 10% in secondary school); by design Oportunidades cash transfers are to have increased participating households income by about 20% (INSP, 2006). The Oportunidades targeting system consists of three stages. First, it selects the most deprived communities using census data on human development and basic infrastructure (56.3% of households taking part in the Oportunidades live in the eight states with the lowest degree of human development). Second, it identifies eligible households using a proxy meanstest of each household in the selected localities. Third, in rural areas, a community assembly must approve the final list of beneficiaries. In terms of education, health and nutrition, showed that outcomes improved: Education: Oportunidades has contributed to reduced dropout rates in primary education; increased enrolment rates in rural secondary schooling and a marked increase in female participation in education; and, has also led to a significant decline in child labour (INSP, 2006; and Parker and Skoufias, 2000). Health: interventions under the Oportunidades programme are credited with a significant increase in the use of medical services and significant reductions in infant and maternal mortality. The evidence on the effects of Oportunidades on child malnutrition is mixed (Hernández and Hernández, 2003; Huerta, 2005; and Meneses et al., 2003). Other OECD countries also have integrated (health, education, and social) service delivery programmes to enhance child development in disadvantaged areas or among disadvantaged groups. For example, New Zealand recently started various initiatives involving early intervention support to vulnerable families (Family Start, Early Start). US policy has long developed a range of investment programmes targeted to reduce child poverty and disadvantage, of which Head Start is arguably the most important. This is a pre-school programme for disadvantaged children that serves over children in predominantly part-day programmes, at a budgetary cost of over USD 5 billion. Over time, Head Start has been evaluated to generate long-term benefits with benefits outweighing the costs (for a recent evaluation, see Ludwig and Phillips, 2007).

122 CHAPTER 3. REDUCING POVERTY IN THE WORKING-AGE POPULATION 125 In the United Kingdom, the Sure Start programme which was initially targeted 250 disadvantaged areas has evolved into the Children s Centres programme with widespread coverage, which is planned to rise to centres in The initiative offers support from pregnancy through to starting school by offering integrated day care and early learning, health, family and parenting support. The similarity with the psycho-social support aspects of the Chile Solidario programme and the Chile Crece Contigo programme is illustrated by the underlying policy of the Children s Centre initiative: i) increase the availability and sustainability of affordable childcare places for children, especially those who are disadvantaged; ii) provide integrated services for health, education and emotional development of young children; and iii) provide services to parents to support them as parents and to help them become job-ready. However, a preliminary evaluation of the Sure Start and Children s Centre initiative found mixed evidence on the positive effect of the programme on child well-being (Belsky et al., 2006). Less socially-deprived parents seemed to benefit most and the most disadvantaged children and families the least. This shows the difficulty of targeting and of getting the most disadvantaged families to participate. Belsky et al. (2006) also found that programmes led by health services (through the existing network of health visitors) are more effective in serving the most disadvantaged than programmes led by other agencies. Identifying clients To increase access to public services efficiently, it is necessary to identify potential clients and their most urgent needs. Chile Solidario incorporated and superseded the Puente [Bridge (to services) programme], and the identification tool used by Chile Solidario originates from this initiative. The Puente programme was rolled-out over the 1990s, subject to federal (and local) budget constraints, first in regions with many low-income households (as identified by CASEN, see Box 3.2) where NGOs and local administrations were willing to participate in the programme. Once an area was selected, inhabitants were informed about the various projects, and social workers/local officials selected families to be invited to participate in the programme. Family support officers would visit the family at home and during this home visit collect information on the household needs in a structured manner. This so-called CAS-2 form collated information on family needs with respect to 53 minimum conditions along seven main categories: residency/location; family situation; health; education; employment status; income; and, housing and living conditions (note the similarity to the CASEN categories, Box 3.2). This information would then lead to an assessment of family needs and those classified amongst the most vulnerable in the population (often sole parent families) were invited to sign up to the Puente programme.

123 126 CHAPTER 3. REDUCING POVERTY IN THE WORKING-AGE POPULATION The CAS-2 form has evolved into the Ficha de Proteccion Social (Social Protection Record SPR). This is a similar tool (with the same seven categories) for the identification of needs and tailoring support, but now for a broader group of clients. Questions within categories have changed, for example, the CAS-2 form focuses on earnings of primary earners, while the SPR form asks about earnings for all household members age 12 and above. The SPR establishes the household or family score by adding up individual scores while applying equivalent scales to account for household composition. In order to correctly identify the degree of long-term vulnerability, the SPR aims to measure income generating capacity, of a family/household, not its actual earnings As a result, declared earned income only accounts for 10% in the overall SPR-index calculation. Key characteristics of the SPR include: Clients no longer have to wait to be invited to complete the relevant form but can do so at their own request. People have a strong incentive to complete the form, as SPR scores are now used for the assessment of needs in a range of supports, e.g. housing, education, health benefits. By 2008 about 2.5 million households with 8.3 million people (half the population), had completed the SPR; an impressive feat which underlies a successful implementation of targeted social policy. 44 Each of the individual (health, education or otherwise) support programmes has its own SPR score thresholds and/or additional conditions. For example, housing authorities combine the SPR score with other considerations such as current living conditions in order to determine eligibility for support. But while in the past benefits were allocated on the basis of the available funds, at present eligibility (and the number of beneficiaries) is defined by a specific rule. The SPR is valid for two years (in theory scoring rules can be revisited each year). Depending on the type of support, households may have to be evaluated regularly. For example, the child benefit payment to poor households (see below) is revised every two years in view of the SPR score. Some other benefits, e.g. disability support and housing support, may be re-evaluated less frequently, if at all. Clients must inform local authorities of relevant household situation changes (moving house, additional children, or change in the head of household of the family), but there is no requirement to 44. A related feature is that scores and client profiles are different than in the past, for example, the education profile of clients is now much more complex.

124 CHAPTER 3. REDUCING POVERTY IN THE WORKING-AGE POPULATION 127 report a change in jobs or income. It is unknown how often households report change in the intervening period between assessments, and it is unclear how local officials react to a re-assessment request if the household income situation takes a sudden turn for the worse. It is unclear just exactly how the answers to the SPR questionnaire affect the SPR score and thus potential benefit receipt. For example, it is not publicly known how the weights of the seven main categories (residency/location; family situation; health; education; employment status; income; and, housing and living conditions) affect the SPR score. This makes it more difficult for clients to adjust behaviour (shape questionnaire responses). It has been asserted that items that can be verified with relative ease (education level, housing, and household composition) carry a lot of weight in the SPR score while income (which would be much more difficult to verify) plays a limited role (with education being a good proxy in any case). SPR thresholds are discretionary, unless individual programme rules specify otherwise (e.g. linking generosity with a range of different scores). There is no phase out or gradual withdrawal of support. The SPR is doing a good job at identifying the targeted population. However, the SPR aims to measure income-generating capacity of households/families, and how individual responses affect SPR scores is not immediately obvious. However, as SPR scores are increasingly being used to establish entitlements of clients to services and (prospectively increasing) cash transfers, it will become increasingly important that household incomes are accurately reported and that the SPR allows for a gradual withdrawal of benefits, so as to avoid financial disincentives to work and continue to ensure that support is provided to those who need it most. Supports, coverage and public spending The Chile Solidario package includes the following support benefits in three broad groups: 1. Psycho-social support (in-kind support by family workers, including referral services), as under the Puente programme. Clients so selected sign a participation contract with the authorities to set out the actions they are to undertake; this could vary from getting an ID card to participating in networks, education, local projects, etc. The contract is also signed by family workers and commits authorities to provide relevant services. Families leave the programme after five years and cannot re-enter the stock: this is a transitory programme (there is little

125 128 CHAPTER 3. REDUCING POVERTY IN THE WORKING-AGE POPULATION point to engage in psycho-social-support twice). There is a small participation or protection payment which declines from a monthly starting payment of CLP (or USD 24 in June 2008) over a two-year period to (CLP 5 393; about USD 10). Upon exit of the plan after two years, the exit bonus is paid at the same low monthly rate as the Family allowance (CLP 5 393) for three years. 2. Five guaranteed minimum payments: a) Financial support to get an identity card; b) The Subsidio Unico Familiar or Family allowance worth CLP about USD 10; c) Subsidy to obtain drinkable water (subject to regional variation worth about CLP to CLP ); d) PASIS (the minimum pension, see Chapter 4); e) Subvencion Pro Retencion Escolar: Subsidy for schools to retain the child in question in school the following year. The education system in Chile consists of eight grades of primary school and four grades of secondary school, and the aim of the subsidy is to help low-income families and reduce the drop-out rate Preferential access to other supports that help achieve self-sufficiency. Given other conditions, e.g. the state of current dwellings, low-income families may receive priority selection for certain housing programmes and projects, but also receive health and education supports. Similarly, family workers may give priority in referring clients to local employment services (OMILs), but this priority is not necessarily shared by OMIL staff (see below). By design, Chile Solidario focuses on helping clients to access social services, rather than provide extensive financial support. As a result, public spending on benefits delivered under Chile Solidario is low at 0.15% of GDP (if spending on minimum disability and old-age pensions (PASIS, see the next chapter) were included spending would amount to about 0.5% of GDP). Between June 2002 and April 2008, Chile Solidario contacted over families, of which the vast majority (88.6%) completed or participated in the programme. Only 2.6% of the contacted families decided 45. The detail of payments under the Subvencion Pro Retencion Escolar were in June 2008 CLP (or USD 117) for children in 7 th and 8 th grade primary school); CLP for children in 1 st and 2 nd grades of secondary school; CLP for children in 3 rd and 4 th grades of secondary school and CLP for those in the last period before completion.

126 CHAPTER 3. REDUCING POVERTY IN THE WORKING-AGE POPULATION 129 not to participate; 2.3% could not be located at the address given; and, 6.5% were taken off the programme because of systematic non-compliance. In April 2008, there were just over families in the programme: families in the psycho-social support phase (this is scheduled to be increased to ), whereas families received the exit bonus. Benefits are paid to mothers rather than fathers as this is most likely to benefit the children (e.g. Woolley, 2004). Considering participation in 2008 (budget data), there are new clients to the psycho-social support phase, receiving an exit bonus, and using preferential access supports. In addition, there are special programmes for the elderly living alone (Programa Vinculos) covering elderly people (to be increased to ) and for the homeless (Programa Calle) covering persons to be increased to 6 000). In all, Chile Soilidario focuses on three key target groups: families; the sole elderly; and the jobless. Other family and child-related support programmes Chile Crece Contigo (Chile grows with you) is the most recent Chilean social protection initiative. The programme started in August 2007 (programme spending was around USD 6.2 million). It provides support to mothers and children from the fifth month of pregnancy onward until entry in pre-primary school (age 4). The plan is scheduled to be fully rolled out by 2009 with all mothers and children in the target group to have access to primary health care, additional health and nutrition support and psychosocial support. The programme provides a wide range, of support touching all relevant aspects of the family environment, health, cognitive development and education. The programme provides supports in a cascading form. Other than the 20% of children in higher income households who use private medical health facilities, all other children are seen by public health authorities in their first year. On the basis of SPR scores and health conditions, about 40% of children are regarded as vulnerable, and this category gets the most intensive form of support: home visits, eligibility to the family allowance (Chile Solidario), free access to childcare, preferential access to public support programmes (e.g. housing), and additional support for mental and/or physical handicapped children. An information guide has been developed for (prospective) parents with very young children. Monitoring is mainly by health authorities, with local authorities (who operate education and health systems) reviewing other development conditions. Childcare support has also emerged as a priority on the Chilean social policy agenda. To fight poverty, and promote child development, gender

127 130 CHAPTER 3. REDUCING POVERTY IN THE WORKING-AGE POPULATION equity and equal opportunities (OECD, 2007e), the government is expanding childcare capacity. One thousand childcare facilities were created from 1990 to 2005, but there are new facilities scheduled for the period. The 2008 budget covered the construction of 900 childcare facilities with places by local authorities under the aegis of the national childcare authority JUNJI (Junta Nacional de Jardines Infantiles) at a cost of USD 81 million. 46 A similar amount was available for the extension of kindergartens as supervised by the Ministry of Education. Spending on childcare and pre-school education increased from 0.08 in 2006 to 0.11% of GDP in 2007 despite the rapid increase of national income. This is still a very low amount: in OECD countries, such support often amounts to a quarter percent of GDP or more. Given the combined challenges of increasing female labour force participation and the quest for adaptability in the labour market, a further spending increase on childcare is a sound social investment, provided the quality of childcare provision is assured. Assessing Chile Solidario There are different aspects to the assessment of Chile Solidario, including targeting, labour market integration and poverty reduction, and capacity challenges: Targeting. There is general agreement that Chile Solidario supports reach the people who need it most. Soares et al. (2007) asserts that about 60% of the resources of both the Chile Solidario programme and the Mexican Oportunidades programme go to the poorest 20% of the population, which is somewhat surprising given the different target mechanisms of the two policies. Contreras et al. (2008) found that two-thirds of the Chile Solidario participants belong to the very poor population (households who have difficulty to meet basic nutritional requirements of household members), and another 15% to the poor population (households who have difficulty in meeting clothing, transport and housing needs, see Chapter 1 for more details on definitions of national poverty lines). As one-third of the very poor population is excluded, there are ongoing efforts to extend coverage, especially in rural areas. Labour market integration and poverty reduction. The Chile Solidario package seems to have had little effect in helping clients get back to the labour market and a small (and lower than expected) 46. JUNJI (part of the Ministry of Education) allocates childcare support to the poorest areas based on CASEN data. Local authorities are involved but NGOs often provide the services, making it easier to adjust capacity.

128 CHAPTER 3. REDUCING POVERTY IN THE WORKING-AGE POPULATION 131 effect on earnings and poverty reduction (Contreras et al., 2008). This is not surprising as public spending on employment support in the programme, and on labour market support more generally is extremely limited at around 0.1% of GDP. In addition, client characteristics are not favourable: clients are frequently not job-ready, face serious skill deficiencies, and are generally difficult to place. 47 In view of these client characteristics, local employment services (OMILs) find it difficult to provide supports that help clients become job-ready. In fact, referring clients who are not jobready is likely to be counterproductive as the referral will raise expectations among clients that subsequently prove beyond reach, causing frustration and loss of motivation. Childcare supports are another issue. Many Chile Solidario clients are mothers with small children, for whom access to affordable quality childcare arrangements is crucial to labour market participation. Although capacity constraints continue to exist, policy measures are being implemented to expand day-care capacity (see above). Capacity challenges. In addition to the coherence issues (see Box 3.2), the rapid development of new policies poses both quality and capacity challenges to service providers. There is widespread concern about whether local administrations have enough family workers to provide intensive family support, 48 and/or professional staff in employment offices to help match jobs and workers. Moreover, even if sufficient resources for staff are available, it is not certain that qualified staff can be found. OMILs already face capacity concerns as staff-to-client ratios of 1 to 300 clients do occur (Chapter 2), while a ratio of around 1 to 125 individuals (not families) is needed to be reasonably effective in helping clients (OECD, 1999). With half of the social/family workers without documented skills for the job, additional investment in the quality of family support and employment service workers will be key to raising the effectiveness of both services. 47. Training programmes for Mapuche (the largest ethnic minority in Chile) preparing them to make traditional handicraft for tourist sales, are likely to be too small-scale to be a structural solution. 48. Chile Solidario services (referrals, benefit payments) have to be delivered in a set time after a claim has been put in. There is no clarity on the sanctions in case of failure (by local authorities), as it has not yet happened. The rules on timeliness are a clear signal towards local authorities and providers that they have an obligation to deliver.

129 132 CHAPTER 3. REDUCING POVERTY IN THE WORKING-AGE POPULATION Further developing supports to the working-age population There are many poor families in Chile who have income from work, and in addition to health, education and/or childcare supports, there is growing pressure on the government to improve the financial position of low-income families. Increasing the currently low child allowances would be one option, but this does not strengthen the financial incentive structure, and it does not help public authorities pursue employment objectives. With employment rates among the poorest 10% of households at 30%, compared with 70% among the richest 10% of households, and with low female employment rates, there is ample scope for mobilising labour supply among groups of workers currently under-represented on the labour market. In-work benefits that increase the financial reward from being in work could be introduced to address employment opportunities of specific labour market groups (e.g. young workers or mothers with children) or workers with low skills and work productivity more generally. Such in-work policies are being used increasingly in OECD countries to help people into work and increase employment, and increase rewards to low-productivity employment (i.e. reduce working poverty through redistribution). In-work benefits can be either employment-conditional benefits and/or tax credits (similar objectives can also be pursued through employment subsidies and tax rebates given to employers). The Chilean authorities are considering proposals by the Presidential Advisory Council on Work and Equity which involve a general job-entry subsidy (with payments to both employers and employees) and targeted child benefits: A job-entry subsidy (Subsidio al ingreso del trabajo, SIT), the SIT would target the vulnerable and poor, and would be paid to a household member taking up work. The proposed subsidy rate is 30% for wages of up to UF 7.5 (equivalent to the minimum wage, CLP per month (or about USD 330), tapering off to zero at wages equivalent to CLP (USD 660). The worker would receive two-thirds of these amounts, while the employer would receive a subsidy for one-third of these amounts to stimulate hiring of workers under formal employment conditions (with health and pension contributions due to the tune of about 20% of earnings). Targeted child benefits: a conditional child allowance worth about UF 0.5 (or CLP or USD 22) per child per month. The benefit, to be paid to the mother, would be targeted at families in the two bottom income deciles, and be conditional on school-attendance and health check-ups. Clients would be identified through the SPR.

130 CHAPTER 3. REDUCING POVERTY IN THE WORKING-AGE POPULATION 133 This procedure was chosen because low-wage workers pay no income tax and because it maintains systemic coherence with other aspect of the Chile Solidario programme (including the family allowance). To avoid fostering dependency, it is proposed that the new child grant would be gradually reduced, to be eliminated after three years or when the child reaches a certain age (no maximum age for this child allowance was cited in the proposals). There are many uncertainties surrounding the proposals. For example, it is unclear, whether the in-work benefit will be based on household-earnings or individual earnings, or how the income testing is going to work if the SPR is only re-assessed once every two years and clients need not report changes in household income. In addition, with the growing number of income transfers that depend on SPR scores may start to affect household behaviour: clients will face increasing incentives to obtain low SPR scores. 49 The lack of public knowledge on how questionnaire responses affect SPR scores may mitigate this effect but it could also have unpredictable effects if clients make more or less erroneous assumptions concerning the way SPR works. In all, there seems to be considerable scope for unintended payments. It is also unknown to what extent introducing these plans will lead to changes in behaviour among workers and employers. For example, the proposed scheme is intended for job-starters. However, current workers and their employers may well change their behaviour, for example via a hike in lay-offs and new hires in order to benefit from the subsidies. Some workers will wish to formalise their employment relationship as SIT is worth up to 30% of wages at maximum, which is sufficient to cover the contributions that formal workers will have to pay for health and pension coverage (approximately 20% of wages). However, the wage subsidy will be considerably less than 30% on average across the earnings range (while keeping in mind that low-income households have free coverage of basic pensions and health care), so other workers may not have enough incentives to formalise employment contracts. Some evidence on in-work benefits across the OECD The government is expected to come up with response proposals by the end of 2008 for consideration by Congress in A detailed assessment is not yet possible, but there are some general points to consider on the basis of the experience with in-work benefits in many OECD countries. 49. At the extreme, clients who assume that low education gives a high score might well be less keen on improving their education if this is thought to deprive them of other desirable benefits.

131 134 CHAPTER 3. REDUCING POVERTY IN THE WORKING-AGE POPULATION The economic conditions and labour market conditions in Chile seem to favour the introduction of an in-work benefit scheme, without seriously deteriorating financial incentives for existing workers. In-work benefits are shown to be most cost-effective in situations where the distribution of earnings is wide, when earnings are low (and the wage floor is low) and when tax rates are low. Such conditions exist in the United States, where the Earned Income Tax Credit successfully makes work pay for low-wage workers, and are even more characteristic of the Chilean situation (for a full overview, see Immervoll and Pearson, 2009, forthcoming; and Pearson and Scarpetta, 2000). 50 The lower the wage floor, the lower will be the proportion of workers covered by a subsidy of a given size, thus reducing the costs of the scheme. Moreover, a low wage floor increases the scope for gradual withdrawal of benefit payments, keeping Marginal Effective Tax Rates 51 at a minimum, and financial incentives strong across the phase-out range. 52 Similarly, low tax rates and a wide income distribution also make it easier to gradually phase out benefits over a long earnings range without affecting large groups of middle-income workers, and increasing the cost of the scheme. To reduce the cost of the overall in-work benefit initiative, earning thresholds may be set somewhat lower to limit coverage, or be more stringently targeted at a specific group, for example women. Such targeting is effective in keeping overall costs down, permitting large subsidies to be focussed on those who do benefit from the programmes. Apart from a small tax credit for singles in the United States, the EITC is targeted at families with children. Across the OECD, authorities are reluctant to see children grow up in poverty. Targeted in-work benefits for families with children increase their incomes both directly and indirectly, by strengthening work incentives (OECD, 2007f). Research suggests that in-work benefits that are based on household earnings are most likely to increase labour supply of primary earners in couple families, and sole parents, but not of second 50. For example, for US evidence on the positive overall labour supply effects of the Earned Income Tax Credit, see Hotz and Scholz (2003) and Kneebone (2007); for evidence on the labour supply effects of the Working Families Tax Credit and subsequent associated reforms in the United Kingdom, see, for example, Brewer (2007). 51. The Marginal Effective Tax Rate (METR) is defined as that part of additional earnings which is taxed away through the effect of increasing tax and withdrawal of benefit income as earnings rise with increased hours of work. 52. If a good balance is to be achieved on poverty and work incentives outcomes, conflicting aims have to be addressed. An in-work benefit scheme will be most efficient in reducing poverty if it provides for exactly the difference between actual earnings and subsistence income. But such a design would imply a 100% METR on additional earnings and is thus inadequate from an employment perspective. Hence, the programme design has to embody a trade-off between equity, employment and fiscal cost objectives.

132 CHAPTER 3. REDUCING POVERTY IN THE WORKING-AGE POPULATION 135 earners (OECD, 2007f). 53 For couple families with no earners, an in-work benefit strengthens the financial incentive to have at least one adult in work. However, eligibility to the in-work payment also increases an income effect away from work; it becomes more affordable not to be in work, but e.g. care for children, and from a household perspective it is cheapest if the second earner (with lowest earnings per hour) reduces labour supply. Compared to schemes based on household earnings, in-work benefits based on individual earnings will be more expensive (they would simply cover more workers at a given income threshold). However, given the extremely low rates of female employment in Chile, it seems appropriate to strengthen financial incentives to work for women. Despite the extra cost, a system based on individual earnings makes more sense in the Chilean situation. Expectations of positive effects on formal employment should not be exaggerated. In-work benefits have not been tried on a wide scale in developing countries, but recent Mexican experience suggests that the effect on formalisation of labour could well be rather limited. Mexico s Primer Empleo programme subsidies social contributions for 12 months for businesses that comply with tax regulations and employ either youths in their first jobs or workers who have never been registered for social insurance. The Primer Empleo was introduced in 2007, and was expected to increase formal employment by jobs, but results were disappointing as in 2007 only about jobs were registered under the scheme (Revista Fortuna, 2007). Changes to the programme were implemented in early 2008, but on the basis of the 2007 experience it seems inappropriate to expect drastically improved results in formal job-creation for In all, in-work benefits should be seen in the context of overall reform to improve the functioning of the labour market, including changing the strictness of EPL, reform of severance pay and unemployment benefits and activation policies, SME development, and measures for training. 54 Furthermore, in-work benefits can reduce financial incentives for second earners. Hence, rather than trying to mitigate such effects and complicate benefit design, it is probably best to design straightforward benefit rules, 53. The design of an in-work benefit also affects the number of hours that recipients engage in paid work. For example, because of the benefit rules, the Working Families Tax Credit and the Working Tax Credit in the United Kingdom embodied strong financial incentives for sole parents to engage in work for 16 hours per week rather than a full-time working week (OECD, 2005). 54. An unfortunate effect of in-work benefits is that they reduce the financial incentives to invest in increasing one s own skills. Hence, in-work benefits should be complemented with policies that help the low-skilled to retain jobs once they find them and to promote skill acquisition to find better jobs.

133 136 CHAPTER 3. REDUCING POVERTY IN THE WORKING-AGE POPULATION accept that these are most likely to affect financial incentives for primary earners and adults in sole-parent households, and link benefit support with child-care policies that are much more effective in facilitating employment among second earners in couple families. 3. Conclusions With recent strong economic growth, the incidence of poverty has declined markedly, but at 10% (according to a national definition) poverty it is still high and the income distribution remains wide. This has triggered a search for policy options that could generate more inclusive growth. In recent history, public social policy in Chile has focused on removing shantytowns and helping low-income households to acquire inexpensive accommodation, extending primary and secondary education to all children, and making health care affordable to all. Policy has moved on from this stage, and now there is an increasing focus improving the quality of dwellings, neighbourhoods, healthcare, education, and social services. With limited spending on income support and little success in raising employment, Chile Solidario has not been very effective in reducing poverty until now. There has been much emphasis on developing networks which help people help themselves. This is an innovative approach that has borne dividends, but has its limits. The existing service infrastructure has been inadequate to help many of them into paid employment, which is a condition for any sharp reduction of poverty. Chile needs to continue its investment in the infrastructure of service provision (e.g. training of social workers and employment service staff). However, the Chilean system is good at identifying needy clients. The Social Protection Record covering about half the population appears to have provided authorities with a good profile of low-income households, which is increasingly used to target social, healthcare and education services. However, because the SPR scoring formula is not public knowledge, potential and actual clients cannot predict with certainty how their disposable incomes would change, for example, if a household member takes up or quits a job. Moreover, the SPR is intended to measure household earnings capacity, and does not focus on current income streams. Therefore, it is no surprise, that SPR thresholds are discretionary, and do not account for phasing out or gradually withdrawing income support. In the past, when support depended on availability of budgets this did not matter that much, but as SPR scores now entitle clients to services and increasingly cash transfers, this design features will come increasingly to the fore. Policy makers are therefore well-advised to either adapt the SPR so as to accurately measure current income changes, or design a supplementary tool alongside

134 CHAPTER 3. REDUCING POVERTY IN THE WORKING-AGE POPULATION 137 the SPR to measure short-term income changes. Policy has to ensure that financial support remains targeted on the most vulnerable while avoiding unintended effects on labour supply. In-work benefits (in the form of cash transfers) are being considered for introduction in 2009 as a way of supporting those in low paid jobs. Measures based on household earnings improve equity and have proven to be effective in raising labour supply among primary earners and adults in sole parent families, but the evidence also indicates that such policies can reduce the labour supply of second earners in couple families. Compared to schemes based on household earnings, in-work benefits based on individual earnings will be more expensive, but given the extremely low rates of female employment in Chile, it would be unfortunate if incentives for women to work were not strengthened. Despite the extra cost, a system based on individual earnings makes more sense in the Chilean situation. An alternative would be to integrate a household-earnings based in-work benefit for low-income workers with other measures such as, for example, a differentiated minimum wage and childcare supports: for example, by paying higher childcare support rates in case both adults are in employment for a specified number of hours per week. Chilean social policy is at a crossroads in that, for the first time, policy makers from all persuasions are debating how to deal with in-work poverty that affects large parts of the working-age population. If adopted, the recent proposal for in-work benefits could cost as much as half a percent of GDP three times the Chile Solidario spending (not counting the basic pension). Compared to an OECD average of just over 20% of GDP, Chile spends little on social protection (about 12% of GDP, not including education) and therefore the redistribution of resources within the Chilean tax/benefit system is limited. There appear to be several promising areas for investment with potentially high social returns: Chile Solidario, in-work benefits and childcare support. If there is fiscal leeway and if sufficient attention is being paid to programme design and implementation issues the Chilean authorities should invest more in an integrated package of these active social policies which promote labour force participation and self-sufficiency.

135 138 CHAPTER 3. REDUCING POVERTY IN THE WORKING-AGE POPULATION Bibliography Belsky, J., E. Melhuish, J. Barnes, A.H. Leyland, H. Romaniuk and the National Evaluation of Sure Start Research Team (2006), Effects of Sure Start Local Programmes on Children and Families: Early Findings from a Quasi-experimental, Cross Sectional Study, British Medical Journal, 16 June. Brewer, M. (2007), Welfare reform in the UK: , Working Paper No. 20/07, Institute for Fiscal Studies (IFS), London. Cantallopts, J., M. Jorratt, and D. Sherman (2007), Equidad Tributaruia en Chile: Un Nuevo Modelo para Evaluar Alternativas de Reforma, ECLAC/CEPAL Seminar ( 20Tributaria-Chile-Ver4.doc). Contreras, D., O. Larrañaga and J. Ruiz-Tagle (2008), Evaluación de Chile Solidario, Borrador, UNDP. Engel, E., A. Galetovic, and C. Raddatz (1998), Taxes and Income Distribution in Chile: Some Unpleasant Redistributive Arithmetic, Serie Economia, No. 41, Centro de Economia Aplicada, Universidad de Chile. Handa, S. and B. Davis (2006), The Experience of Conditional Cash Transfers in Latin America and the Caribbean, Food and Agriculture Organization, Agricultural and Development Economics Division, FAO- ESA, ESA Working Paper No ( Hernández, B. and M. Hernández (2005), Evaluación externa de impacto del Programa Oportunidades 2004, Tomo I, Educación, Instituto Nacional de Salud Pública, Cuernavaca, México. Hotz, V.J. and J.K. Scholz (2003), The Earned Income Tax Credit, in R.A. Moffit (ed.), Means-tested Transfer Programs in the United States, Chicago University Press, Chicago. Huerta, M.C. (2005), Child Health and Nutrition in Rural Mexico: Did Progress Improve the Life Chances of the Very Poor?, Ph.D thesis, University of London.

136 CHAPTER 3. REDUCING POVERTY IN THE WORKING-AGE POPULATION 139 Immervoll, H. and M. Pearson (2009), Make-Work-Pay Policies in OECD Countries: An Update, OECD, Social, Employment and Migration Working Papers, OECD, Paris, forthcoming. INSP Instituto Nacional de la Salud Publica (2006), Encuesta Nacional de la Salud y Nutricion ENSANUT (Results from the Health and Nutrition National Survey), Instituto Nacional de la Salud Publica, Mexico. Keen, M. and S. Smith (2007), VAT Fraud and Evasion: What do we Know and What can be Done?, IMF Working Paper No. 07/31, IMF, Washington D.C. Kneebone, E. (2007), A Local Ladder for Low-income Workers: Recent Trends in the Earned Income Tax Credit, The Brookings Institution, Earned Income Tax Series, April ( Larrañaga, O. (2009), Inequality Poverty and Social Policy, Recent Trends in Chile, OECD, Social, Employment and Migration Working Papers, OECD, Paris. Ludwig, J. and D.A. Phillips (2007), The Benefits and Costs of Head Start, NBER Working Paper No , National Bureau of Economic Research, Cambridge, MA, March ( MdH Ministerio de Hacienda (2007), Estadísticas de las Finanzas Públicas , Gobierno de Chile, Ministerio de Hacienda, Santiago de Chile. Meneses, F., B. Hernández, M.C. Baltazar, A. Camacho, M.B. Duarte, J.E. Urquieta, M.M. Téllez and M. Hernández (2003), Evaluación del cumplimiento de metas, costos unitarios y apego del Programa Oportunidades a las Reglas de Operación. Evaluación Externa del Programa de Desarrollo Humano Oportunidades, Ministry of Social Development (SEDESOL), Mexico. Mideplan (2007), Serie Análisis de Resultados de la Encuesta de Caracterización Socioeconómica Nacional (CASEN 2006), No. 2, Distribución del Ingreso e Impacto Distributivo del Gasto Social 2006, Santiago, June. MINVU Ministerio de Vivienda y Urbanismo (2008a), Subsidios ( MINVU (2008b), Estadisticas, Series Historicas, Viviendas Terminadas y Subsidios Pagados, Planilla Total Pais (

137 140 CHAPTER 3. REDUCING POVERTY IN THE WORKING-AGE POPULATION Nam, C.W., A. Gebauer and R. Parsche (2003), Is the Completion of EU Single Market Hindered by VAT Evasion?, CESifo Working Paper No. 974, CESifo, München, Germany. Navarro, M. (2005), Housing Finance Policy in Chile: The Last 30 Years, Land Lines, Vol. 17, No. 3, July. OECD (1999), The Battle against Exclusion, Vol. 3, Social Assistance in Canada and Switzerland, OECD, Paris. OECD (2003), OECD Economic Surveys: Chile, OECD, Paris. OECD (2004), OECD Reviews of National Policies for Education, Chile, OECD, Paris. OECD (2005), Babies and Bosses, Reconciling Work and Family Life, Vol. 4, Canada, Finland, Sweden and the United Kingdom, OECD, Paris. OECD (2007a), OECD Economic Surveys: Chile, OECD, Paris. OECD (2007b), OECD Revenue Statistics , OECD, Paris. OECD (2007c), $ocial xpenditure Interpretative guide to the OECD Social Expenditure Database SOCX, OECD, Paris ( OECD (2007d), Education at a Glance, OECD, Paris. OECD (2007e), Babies and Bosses, Reconciling Work and Family Life: A Synthesis of Findings for OECD Countries, OECD, Paris. OECD (2007f), Benefits and Wages, 2007 Edition, OECD, Paris. OECD (2008), OECD Social Expenditure Database , OECD, Paris. OECD (2009), OECD Reviews of National Policies for Education, Tertiary Education in Chile, OECD, Paris. Palma. J. and R. Urzua (2005), Anti-poverty Policies and Citizenry: the Chile Solidario Experience, MOST2 Policy Papers No. 12, UNESCO, Paris. Parker, S. and E. Skoufias (2000), The Impact of PROGRESA on Work, Leisure and Time Allocation, International Food Policy Research Institute, Washington, D.C., October. Pearson, M. and S. Scarpetta (2000), What Do we Know about Policies to Make Work Pay?, OECD Economic Studies, No. 31, OECD, Paris, pp

138 CHAPTER 3. REDUCING POVERTY IN THE WORKING-AGE POPULATION 141 Raczynski, D. and C. Serrano (2005), Las políticas y estrategias de desarrollo social: aportes de los años 90 y desafíos futuros, in.p Meller (ed.), La Paradoja Aparante. Equidad y Eficiencia: resolviendo el dilemma, Editorial Taurus, Santiago de Chile. Rawlings, L.B. (2004), A New Approach to Social Assistance: Latin America s Experience with Conditional Cash Transfer Programs, Social Protection Discussion Paper Series, World Bank, Washington D.C. Revista Fortuna (2007), Fracasa programa Primer Empleo ( July. Skoufias, E. (2000), Evaluation Report: Impact of the Education, Health, and Nutrition Program, International Food Policy Research Institute, Washington, D.C. Soares, S, R. Guerreiro Osorio, F. Vario Soares, M. Medeiros and E. Zepeda (2007), Conditional Cash Transfers in Brazil, Chile and Mexico: Impacts Upon Inequality, International Poverty Centre, United Nations Development Programme, Brasilia. Woolley, F. (2004), Why Pay Child Benefits to Mothers?, Canadian Public Policy, Vol. 30, No. 1., pp World Bank (2008), Social Protection Technical Assistance Project, ( pipk= &thesitepk=40941&menupk=228424&projectid=p )

139

140 CHAPTER 4. THE NORMALISATION OF CHILE S PENSIONS SYSTEM 143 Chapter 4. The Normalisation of Chile s Pension System Chile replaced its Pay-As-You-Go social insurance pension system in the early 1980s with a private pre-funded system of individual savings accounts which covers old-age, disability and survivor contingencies. Pension reform was successful in many ways: it restored public confidence in pension saving; pre-funding of pensions has contributed to the development of financial markets and economic growth; and projected future public pension outlays in Chile are lower than in many OECD countries (and far less susceptible to the dynamics of ageing populations). However, pension reform failed to achieve its coverage objectives, and many Chilean workers, including many women, youth and low-income workers with low levels of educational attainment, either did not make sufficient contributions to obtain an adequate contributory pension or have no coverage at all. The coverage problems contributed to a major pension reform in 2008, which introduced a basic solidarity pension as of right, and targeted measures to stimulate coverage among workers with no or little pension saving. Reform also aims to develop a third tier of voluntary private pension saving. In all, Chile s pension system is developing towards the three-pillar pension norm, which is often held as an international benchmark. This chapter discusses the social policy aspects of the reformed pension system and, in particular, the coverage issues which in many ways have proven to be the Achilles heel of Chilean private pension system.

141 144 CHAPTER 4. THE NORMALISATION OF CHILE S PENSIONS SYSTEM 1. Introduction The bad experience with collective (public) insurance schemes in Chile goes a long way towards explaining the choice of the current system of private-funded mandatory individual accounts. Chile was the first Latin American country to set up a social insurance system back in By the late 1970s, the system had evolved into a pay-as-you-go system (PAYG) with 35 schemes, which was deemed unfair and corrupt as generous (frequently early) retirement benefits were often paid out on the basis of political lobbying to special interest groups (Asher and Vasudevan, 2008; Ben Braham, 2007; and Iglesias, 2009). In 1981, a system of individual accounts was introduced to avoid such practices and establish clear links between contributory records and benefit entitlements. Other arguments in favour of private protection arrangements included making efficiency gains, increasing flexibility and individual choice and increasing pre-funding which deepens financial markets and limits the effect of ageing populations on annual financial flows (Pearson and Martin, 2005). The 1981 pension reform was successful in many respects, not least since it restored public confidence in pension saving. The investment returns made by Chilean pension fund operators since 1981 have exceeded expectations and the introduction of a capitalised pension system contributed to the development of financial markets and economic growth in Chile (Corbo and Schmidt-Hebbel, 2003). Moreover, future public pension liabilities in Chile are lower than in many OECD countries and far less susceptible to the fiscal pressures associated with an ageing society. However, the private contributory system has over the years failed to meet some other of its objectives, particularly in terms of coverage. Many low-income workers without formal contract are not at all covered by the private social protection system, while others have paid into the system, but their contributory record is insufficient to procure the minimum pension benefit, leaving public authorities to make up the shortfall and provide a basic pension. In a similar fashion, the arrangements for disability benefits that are also covered in the private pension system involve some cost-shifting by private insurers to public budgets. In all, the private pension system has relied on public authorities to pay for benefits where the private system fails to provide adequate provisions. Since 1 July 2008, pension reform came into operation to address issues such as coverage among the population, including lower income groups, women, young people and the self-employed. However, the reform addresses a range of issues, including, for example, financial incentives towards early withdrawal of pension wealth (OECD, 2006a), but also

142 CHAPTER 4. THE NORMALISATION OF CHILE S PENSIONS SYSTEM 145 measures to increase market efficiency and regulatory reform (see, for example, Iglesias, 2009). The Chilean pension reform of the early 1980s attracted a lot of attention and gave rise to a substantial literature (for example, Corsetti and Schmidt- Hebbel, 1995; Diamond, 1993; Edwards, 1996; Queisser, 1998; and World Bank, 1994). This is not repeated here. Instead, this chapter will focus on the social policy aspects of pension policy, and while it does not claim to be comprehensive in its treatment of pension market institutions and investment regulations, it will briefly discuss the key challenges of increasing competition investment regulations and financial risk management. This chapter takes the new pension system and its parameters as in place since 1 July 2008 as the basis for discussion, but the reader should be aware that, while some aspects of the reform have already been put in place, other measures are expected to be introduced over the next four years. 2. The Chilean pension system With the 1981 pension reform the existing contributory (PAYG) pension schemes were brought together under the Umbrella of the Instituto de Normalizacion Previsional (INP), while the individual pension accounts that were introduced at the time are operated by pension fund management companies (Administradora de Fondo de Pensiones AFP). Many clients of the old system changed to the new mandatory private pension system as facilitated by Recognition Bonds which acknowledge the value of the entitlement built up under the old system (Box 4.1). In addition, there was a publicly-supported Guaranteed Minimum Pension (Minima Pension Garantia MPG) for those who have contributed for at least 20 years (or 240 monthly contributions), but whose contributory record was nevertheless insufficient to generate a minimum pension (the MPG was an entitlement). Furthermore, those of retirement age without a pension (or a very small one) could apply for a means-tested social assistance pension (Pensiones Asistenciales PASIS), but this benefit was not an entitlement, and budget-rationed (if the budget was exhausted in a given year, no new claims would be awarded). Finally, there are separate pension programmes for the armed forces. Since reform started to be rolled out in July 2008, Chile s pension system involves: 1. The Basic Solidarity Pension (Pension Basica Solidaria PBS) and a Pension Solidarity Complement (Aporte Previsional Solidario APS) for low-income households without sufficient pension contributions. 2. The mandatory capitalised individual pension saving accounts. 3. Voluntary private pension saving in individual accounts.

143 146 CHAPTER 4. THE NORMALISATION OF CHILE S PENSIONS SYSTEM The role of private pension saving in Chile has been limited so far, but reform aims to develop a significant third tier of voluntary private pensions. Since March 2008, employer-sponsored voluntary pension programmes allow employers to make pension contributions on behalf of their employees which are considered as regular expenses for tax purposes (the conditions of such employer-sponsored voluntary personal pension savings schemes must be similar for all company-staff). In addition, through new tax incentives the government also tries to stimulate individual private pension saving. 55 With the development of a basic pension entitlement, mandatory second-tier pension saving and fiscally supported voluntary pension saving, Chile s pension system has moved towards the three-pillar pension norm, which is often held as an international benchmark. Figure 4.1 shows publicly-mandated pension spending across the OECD area (including public spending on pensions and benefits accruing from mandatory contributions to private funds). In Chile, publiclymandated spending amounted to 5.8% of GDP in 2005, which is just below the OECD average of 6%. Figure 4.1 also illustrates how important the public role is in the current pension set-up: public spending was 4.8% of GDP in 2005, which is high compared with the population age profile (Chapter 1), but this is largely of a transitory nature as related to the pension reform in 1981 (see Box 4.1). Of all publicly-mandated spending in Chile in 2005, about 1% of GDP concerned pensions paid through pension payments or Programmed Withdrawals (see below) and annuity payments derived from the mandatory private pension system (though pay-outs often financed by government contributions). 56 For Chile, the data in Figure 4.1 include spending on pension benefits and outlays by pension funds on annuities for 55. Since the 2008 reform the amount of pension savings which can be deducted annually from the income tax base is UF 50 per month (or USD as of June 2008). Workers can also choose to pay income tax over pension contributions now or when the pension is paid out. Workers who choose the former option and who engage in voluntary pension saving to top up the mandatory contributions can benefit from a subsidy of about 15% of the amount of voluntary savings up to a maximum of about USD 410, or ten times the amount of mandatory pension savings in a year, whichever is the lowest. 56. Of all old-age pensions paid in 2004 (not including early retirement pensions), 81% were paid by the INP. AFP paid out 19% of the benefits, which were mainly financed out of government contributions to make up the minimum pension payment (Fundacion Nacional de Superacion de la Pobreza, 2005).

144 CHAPTER 4. THE NORMALISATION OF CHILE S PENSIONS SYSTEM 147 clients which amounted to about 1% of GDP in 2005 (Table 4.A1.1). 57 If only current outlays on pension benefits and annuity payouts (in a given year) were considered, mandatory private pension spending would have been about 0.1% of GDP in Figure 4.1. Publicly-mandated pension spending in Chile is close to the OECD average Public and mandatory private pension expenditure in percentage of GDP, Public Mandatory private Source: OECD (2008), Social Expenditure database ; and Iglesias (2009). In 2007, almost 30% (or about 1.2% of GDP) of public spending on pensions was on benefits for the military (CAPRADENA) and police (DIPRECA). Average pension payments under these schemes are well above the average wage and much more generous than those paid by the old or new pension systems (OECD, 2003). The Armed Forces pension systems run a deficit which is almost wholly paid from the government budget. Reform of these pension arrangements is overdue and proposals will be considered after the 2009 presidential elections. 57. This recording practice is consistent with reporting in the OECD Social Expenditure database (OECD, 2008a) which includes similar data on spending on pension payments and lump-sums through the Superannuation programme in Australia.

145 148 CHAPTER 4. THE NORMALISATION OF CHILE S PENSIONS SYSTEM Box 4.1. Reforming the pre-1981 PAYG scheme into a private-funded pension system During the late 1960s and early 1970s, both the Frei and Allende governments, which were of different political persuasions, had come to the conclusion that piecemeal reform of Chile s then PAYG system would be insufficient: a new system had to be put in place. The PAYG system was widely considered as unfair, as it had been captured by special interest groups; it was also expensive to run and not targeted at those most in need of support. Widespread and profound dissatisfaction with how the PAYG system operated was the key driver of reform, not financial pressures arising from population ageing (Iglesias, 2009). Nevertheless, it was not until 1981 that pension reform was introduced. The introduction of the mandatory pensions system in 1981 only covered those who started to work in 1983 and contributors to the previous system were not compelled to sign up to the new scheme. Nevertheless, by the end of 1983, 77% of contributors had switched to the new AFP programme. To some extent, this was related to widespread dissatisfaction with the old system and a vigorous public information campaign carried out at the time. Two financial incentives also played a key role in convincing pension savers to switch regime: i) while employees had to cover employer contributions (which were relatively low compared with the contribution rates of the old PAYG programme), they were compensated by a more-than-proportional increase in wages; and ii) the government issued a state-guaranteed recognition bond for the contributions workers made to the old system (as adjusted for inflation and a real interest rate of 4% per annum). Switching did not always turn out to be an informed decision; decisions disadvantageous to individual savers largely concerned specific groups of workers in the public sector (Iglesias, 2009). Time-series data on public spending on civilian pensions illustrates how a large part of the cost of pension reform to the public was borne early on in the reform process. From 1981 to 1984, public budgetary allocation towards transitory spending commitments established under the old PAYG pension system increased from 1.9% of GDP in 1981 to 6.5% of GDP in 1984 because of the continued payment of pensions while contributions declined (exacerbated by the economic crisis of the time), and the payment of recognition bonds to those who transferred to the new mandatory pension system). By 2007, this had declined to 2.7% of GDP (Figure below). At the same time social pensions and minimum pensions (see below) are a permanent fixture of the government budget and these outlays increased from 0.2% of GDP in 1981 to just below 0.5% in The transitory part of public pension spending will naturally decline to a quarter of a percentage of GDP in 2030, while the minimum pension commitments to those without sufficient contributions were projected to amount to 1.1% of GDP in 2015 and 1.6% by 2030 prior to the 2008 pension reform (Marcel Commission, 2006). The real cost is likely to be higher in view of the 2008 measures (Iglesias, 2009). The pre-1981 pension system was not in deficit, but projected future cash flows foreshadowed significant future deficits. Available evidence on the implicit pension deficit (the present value of future pension payments) varies in magnitude (Iglesias, 2009), but all point to a significant reduction of this debt because of the 1981 reform, which has thus improved the long-term fiscal sustainability of the Chilean system. In any case, the fiscal effect of the pension reform does not seem to have been an important obstacle to reform. During the years immediately following reform, the pension deficit was largely financed through public savings (by cutting back on general government expenditure) and increased taxation (a temporary tax, 3% of salary paid by employers phased out by 1 percentage point per annum over the period). At the same time treasury bonds were sold to pension funds so that their investment increased from USD 2.2 million in 1981 to USD 864 million in 1986 (Iglesias, 2009).

146 CHAPTER 4. THE NORMALISATION OF CHILE S PENSIONS SYSTEM 149 After 1985, debt financing became less important, while privatisation (the sale of public assets) became an important source of funding: government assets worth 7% of GDP were sold off over the period twice the amount of privatisations in the United Kingdom (Niemitz, 2007). Pension reform contributed to an increase in national saving, but this was largely driven by government saving. The effect on private savings, though significant, has been relatively small (Iglesias, 2009). Public spending on civilian pensions will halve over the next 25 years Public spending on civilian pensions as a percentage of GDP (not accounting for the 2008 reform) 8.0 Transitory spending(1) Permanent spending(1) Total Transitory spending refers to spending on pension commitments derived from the pre-1981 PAYG system; permanent spending concerns public spending on minimum pension programmes as forecasted prior to the 2008 pension reform. Source: Based on data from Iglesias (2009) from the Marcel Commission (2006). As post-reform contributions were lower than pre-reform contributions, the tax on labour was reduced with reform, which contributed to a small increase in employment (e.g. Schmidt-Hebbel, 1998). The establishment of funded pension systems, pension funds and the associated life insurance industry has led to the deepening of local capital markets (in 2005 pension fund assets amounted to two-thirds of national Income; Ben Braham, 2007), while an improvement in regulations and reporting requirements generally enhanced capital market transparency and efficiency. Through its effects on, in descending order of importance, national saving, capital productivity and total factor productivity (a more efficient interaction between capital and labour), pension reform is estimated to have been responsible for 25% of economic growth during the 1980s and early 1990s (Schmidt-Hebbel, 1998).

147 150 CHAPTER 4. THE NORMALISATION OF CHILE S PENSIONS SYSTEM Poverty among elderly and redistribution of pension spending The increase in prosperity has contributed to the decline in poverty among the population in general, as well as amongst the elderly. Minimum pension payments may not be high, but they compare well against the very low transfer payments to the working-age population (Chapter 3). Households in receipt of a pension are unlikely to be extremely poor (only 1.2% of them), and the poverty rate among the elderly is 7.3% compared with 13.7% of the population in general (Figure 4.2). Figure 4.2. The income position of the elderly is relatively good in Chile Poverty rates for the general population and the elderly (65+), Extreme poverty - Elderly Extreme poverty - General population Poverty - Elderly Poverty - General population Source: Based on CASEN, various years. For definitions of poverty thresholds, see Chapter 1. In 2006, about 75% of men and 60% of women aged 65 and over received benefits accruing from contributory pensions (including the MPG), 58 and 70% of these contributory pensioners received less than CLP (USD 242, about 80% of the minimum wage, Larrañaga, 2009). The 2006 CASEN survey showed that about 14% of men and 17% of women over 65 years of age received the minimum pension, and that in all, public pension payments constituted around 6.2% of household income. However, their distributional impact is low. Public pension payments are still dominated by the entitlements established under the old system (Box 4.1) and relevant amounts are related to contributory records. Hence, 58. Two out of five women who received pension payments in 2006 were widows receiving survivor pensions.

148 CHAPTER 4. THE NORMALISATION OF CHILE S PENSIONS SYSTEM 151 the distribution of pension payments is relatively similar to the earnings distribution in terms of household income quintiles (Chapter 1). The Gini coefficient of the distribution of old-age and widows pensions paid by the state is 40.7%, only a few points below the Gini coefficient of the wage distribution, and the Gini coefficient of the household income distribution with and without pensions paid out by the state, is 55.9% and 58.1% respectively (Larrañaga, 2007). Preliminary estimates of the effect of the introduction of the basic solidarity pension suggest reform may lead to an increase of 11.5% in average income of the first household quintile and 9.5% in the first two quintiles, as well as a reduction of around 2 percentage points in the proportion of poor households. The basic solidarity pension Since 1 July 2008, the Chilean pension set-up includes the basic solidarity pension (PBS) and its supplementary payment (APS), which superseded the previous system of PASIS and MPG payments. The generosity of the basic pension system has increased significantly: 1. Unlike PASIS which was potentially subject to budgetary rationing, the new PBS system establishes individual entitlements. 2. Coverage of the first-pillar scheme will also increase because the supplementary APS-payment will also cover those who have not contributed for 20 years to the second-tier pension system (a precondition for receiving the MPG), although some women may lose out since there is no survivor PBS. 59 Currently, about one-third of the 65-year-olds do not have pension coverage, and nearly 70% live in households with a family income below the minimum threshold. By 2012, about 23% of all senior citizens will be eligible for the PBS, and about 70% of the clients will be women, in view of their limited access to contributory pensions (see below) and their longer average life expectancy. 3. In 2007 the MPG was about CLP (about USD 180) per month, about twice as high as the PASIS payment (see Figure 4.A1.1 in the annex for a historical overview). The PBS system pays the poorest elderly (65+ for men and women) a minimum pension of CLP Since the July 2008 reform survivors will only receive a temporary pension financed out of the funds accumulated in the account of the spouse prior to his/her decease. On depletion of these funds survivors will have to wait until age 65 before claiming an old age PBS.

149 152 CHAPTER 4. THE NORMALISATION OF CHILE S PENSIONS SYSTEM (or USD 122), which will increase to CLP (USD 152) in mid-2009 (Table 4.1). Disabled individuals between 18 and 65 years of age can receive a disability PBS payment of the same amount. 60 Table 4.1. The Basic Solidarity Pension (PBS): evolution of payment rates and expected coverage Date (1 July) PBS benefit Threshold up to which APSsupplement can be paid Estimated coverage of PBS/APS pensions among the elderly population 2008 CLP USD 122 CLP USD % 2009 CLP USD 152 CLP USD % 2010 CLP USD 152 CLP USD % 2011 CLP USD 152 CLP USD % 2012 CLP USD 152 CLP USD % Source: Based on data from the Chilean government; exchange rate as applicable in June 2008, USD 1 = CLP Eligibility to the basic pension is established with reference to the Social Protection Record (SPR); the 40% poorest elderly households correspond to a SPR score below points, while the 45% threshold corresponds to points. The system aims to cover the 60% poorest elderly households from 2012 onwards (Table 4.1). The 2008 reform has improved systemic coherence by establishing clear links between the basic pension and the income-dependent supplement (whereas PASIS and MPG payments were unrelated). Low-income individuals receiving AFP pensions (or a pension paid by the INP) can get a supplement. The amount of this old age APS (there is a similar supplement for disabled clients) depends on the Base Pension the individual receives from the other (second-tier) pension programme: Old age APS = PBS Adjustment Factor * (Base Pension) Thus, for a client with no income from another pension programme, the APS payment is equal to the PBS. At the other extreme, the Maximum Base Pension threshold has been set at CLP per month in 2012 (approximately USD 517, but it will be inflation-indexed), and above this 60. Disability PBS payments are made to poor disabled individuals (within 60% of lower income households from 2012 onwards), aged between 18 and 65, who do not receive a pension from another social security programme and who lived in Chile for at least 5 out of six years before being declared disabled.

150 CHAPTER 4. THE NORMALISATION OF CHILE S PENSIONS SYSTEM 153 income threshold no supplements can be received. 61 Table 4.1 shows that the maximum amount to which pensions will be topped up will increase significantly from 2008 to Also, in 2012 the value of the adjustment factor will be 0.294; in other words the implicit effective marginal tax rate on contributory pension will be 29.4% in 2012 (in 2008 it was as high as 85.7%). 62 In the old system, the MPG involved weak incentives to save for those who had established their 20-years of pension contributions and who had not saved enough to obtain a pension higher than the MPG. These clients faced an implicit tax rate of almost 100% on the contributions they paid towards the MPG threshold (such saving up to MPG level did not help to increase individual entitlements, but only reduced the cost of the MPG for the government). Under the new rules, making new pension contributions always increases the value of the future (second-tier) pension payments, while the implicit marginal effective tax rate on contributions is reduced to 29.4% in 2012 (low-income workers may well face a lower rate after introduction of an in-work benefit, Chapter 3). Nevertheless, the implicit tax rate may prove to be too high a hurdle for many workers to start making contributions to the mandatory pension system (OECD, 2007a). Until the recent reform, eligibility to means-tested benefits depended on household income as determined by all household members. The PBS is different in that eligibility is determined by the spouse and dependent children up to the age of 24 living in the same household; the authorities have assumed responsibility for senior citizens who do not generate enough income, independently of the resources of relatives other than the spouse, whether they reside in the same household or not. This change in conditions preserves the many extended families that exist in Chile. Often different generations of one family are included in one household and, if the significantly increased (minimum) pension payment were made contingent on the income of all other household members, it would not have been unlikely that many extended households would have split (or not be eligible for payment). The effect of the PBS on household formation is uncertain: the increase in family resources may lead to a break up in households (who now have enough income to do so), while in other cases families may invite grandparents into the household. 61. Design of the APS supplementary payment (see below) will also leave most current MPG recipients better off. However, under certain conditions some part-time workers, of which there are not many in Chile, with a contributory record of 20 years may be worse off (Iglesias, 2009). 62. In 2012, the PBS is scheduled to be CLP and the maximum income threshold is CLP ; the adjustment factor will then be (75 000/ = 0.294).

151 154 CHAPTER 4. THE NORMALISATION OF CHILE S PENSIONS SYSTEM The mandatory private pension system Key characteristics of the mandatory private pension scheme AFP include: 1. The system covers old-age, as well as general (not employment-specific) disability and survivorship contingencies (see below Section Disability and survivor coverage ). 2. Contributions are deposited (by employers) in personal (individual) accounts operated by the pension management fund (AFP) chosen by the employee. 3. The (monthly) contribution rate to old-age benefits is a proportion of the wage and was 10% in 2008 up the maximum monthly contribution of UF 60 per month (about USD 2 448). Pension contributions are deducted from the income tax base, and any pension received at retirement is added to it. The contribution toward the disability and survivor risk amounted to an average 0.95% of the wage in December 2007, with on average another 1.45% being paid in administrative fees (Iglesias, 2009). 4. The defined-contribution old-age benefit depends on: i) the balance accumulated in the account of each worker at retirement; ii) his/her life expectancy and that of his/her family members; and, iii) investment returns. By contrast, the disability and survivor pensions are defined benefits ; their amount is set in proportion to the historical average wage of the contributor. International comparison on pension replacement rates based on a full contributory record at average wages (World Bank, 2007), show that at a gross replacement rate of about 55% the value of pension benefits in Chile is about 15 percentage points below the OECD average (Figure 4.3).

152 CHAPTER 4. THE NORMALISATION OF CHILE S PENSIONS SYSTEM 155 Figure 4.3. Replacement rates in Chile are relatively close to the OECD average Gross and net (after tax) replacement rates for mandatory pension programmes, Men, percentage of average wages Gross replacement rate Net replacement rate Source: OECD (2007b), Pensions at a Glance; and World Bank (2007), Pensions Panorama. Pension market characteristics: lack of competition and individual choice in investment The state plays a regulatory and supervisory role in the private pension set-up and provides guarantees to fund members in case an AFP (or the life insurance company which pays annuities) goes bankrupt. Government regulations also affect market competition as well individual choice in investment options, and financial risk management. The 2008 reform also aims to increase competition by getting agents to compete for rather than in the market, and to allow the use of new forms of financial tools to further increase yields, even though in view of the financial crisis, current markets are not the ideal investment environment (Annex 4.A2 provides an overview of the Chilean pension market with a focus on competition and investment issues in Chile; see Dayoub and Lasagabaster, 2008, for an overview across Latin America). There is little competition among AFPs in terms of price (user charges) or yields (rates of return on investment). Competition was fierce during the

153 156 CHAPTER 4. THE NORMALISATION OF CHILE S PENSIONS SYSTEM mid-1990s when AFPs used (expensive) direct sales efforts and marketing campaigns to attract clients, but with a tightening of the rules on switching between funds in October 1997, the number of salesman fell from in 1997 to in Entry to the pension market is free except for commercial banks, mutual funds and insurance companies that cannot hold shares in AFPs, which helps shield pension-fund managers from competition by other financial intermediaries. Economies of scale effectively dictate an AFP needs about half a million clients to survive in the market (Iglesias, 2008), and the lack of competition has contributed to a concentration of AFPs in the market from 21 in 1993 to five in On the whole, the operation profit margin of AFPs was on average 18% from 1983 to 1998, but with the shake-out of salesmen operating profit margins have exceeded 25% (Figure 4.A2.2). Initially, the increased profits were at least partially passed on to clients, but since 2001 administrative charges have stopped falling (Figure 4.A2.1), while operating profits have been fluctuating around 30%: almost twice as high as during the first 15 years of the Chilean pensions system. This suggests there is room left to reduce user charges in the AFP system. The experience with financial management of PAYG funds prior to reform led to strict public regulation of AFPs in terms of portfolio selection, investment rules and market behaviour. However, over the years the Chilean authorities have gradually relaxed rules to increase both individual choice and increase yields. 63 Since 2002, each AFP can offer five different portfolios and risk profiles to their clients ranging from risky (up to 80% of capital in the individual account is invested in equity) to a risk-averse portfolio which mainly invests in fixed-income instruments. AFPs are legally required to guarantee a minimum return, which contributes to AFP pension-fund managers taking a relatively careful, with little difference in investment behaviour as smaller AFPs copy behaviour of larger ones. Since 1981, the annual average of real pension fund yield has been in excess of 10%. This result is comparable with pension fund returns in European and North American OECD countries, where the real rate of return (net of management fees) are typically around 7% (Antolin, 2008; and D Addio et al., 2009). 63. Berstein and Chumacero (2003) estimated that in 2002/03, total assets managed by AFPs could have been at least 10% larger in the absence of stringent regulation of investment limits.

154 CHAPTER 4. THE NORMALISATION OF CHILE S PENSIONS SYSTEM 157 Benefit payments: pay-out options and early withdrawal of pension wealth In a strictly regulated environment, AFP clients have basically three pay-out options: 1. Life annuities sold by life insurance companies involving a monthly income until decease of participants, with survivor payments to beneficiaries afterwards. When a worker buys a life annuity, he/she transfers the financial risks attached to both longevity and the rate of return on investment to the life insurance company, and the contract is irrevocable. Life insurance companies hardly existed before the 1981 pension reform, but they have quickly developed since: in 2007 two-thirds of the pension payouts deriving from mandatory saving concerned annuities (Table 4.2). 2. Programmed withdrawals (PW): the worker keeps the funds in his/her personal pension account and draws a regular pension subject to the AFP rules. Benefit rates are set for one year only; and new rates are set each year: about 25% of clients receive their pension benefit in this manner. 3. Temporary income with a deferred life annuity. This is a hybrid of the two previous payment methods: it involves an agreement with a life insurance company to pay the client a life annuity from some future specified date; until that time clients receive a monthly pension financed out of funds in his/her personal AFP account. This method covers less than 5% of all pay-outs. Table 4.2. Distribution of pensioners by payment methods Stock of number of pension-payments (percentage of total) paid out, December 2007 Programmed withdrawals Annuities Temporary PWs Total Old age Early withdrawals Total Source: Superintendencia de Pensiones. Programmed withdrawals (PWs) and annuities both have a gradual withdrawal profile, but they provide a very different time stream of benefits and risks. Programmed withdrawals have the advantage that they allow the retiree to: i) get his/her money out of the system more quickly than an annuity would, due to the required mortality and interest-rate assumptions; ii) choose and vary the AFP and investment portfolio, selecting a (risky) portfolio with a higher expected return than annuities; iii) leave a bequest to his/her heirs if he/she dies early; and iv) switch to an annuity later on, if

155 158 CHAPTER 4. THE NORMALISATION OF CHILE S PENSIONS SYSTEM desired, whereas the choice of an annuity is irreversible. These advantages might seem to make PWs attractive to retiring workers, especially those with high discount rates (short life expectancy), bequest motives and investment experience. However, with PWs there is no investment and longevity insurance, and annual payouts can vary with the rates of returns on investment. The annual income stream will diminish over time and will become very small if the worker lives long enough. To risk-averse workers the latter is a strong incentive to annuitise. The choice of payout method also depends on the value of accumulated capital in the individual account. In general, those with limited pension savings (generating a benefit just above the MPG and PBS) will choose PW as personal savings will be used relatively quickly while a minimum pension is guaranteed. However, the government also guaranteed 75% of the annuity value in excess of the minimum pension (and 75% of the annuity in excess of the PBS up to a maximum of UF 45 or USD in June 2008) to cover insurance company insolvency, and this public back-up of annuitisation affects the pay-out choice for workers with medium and large accumulations. In addition, there exist regulatory incentives and constraints, which lead insurance companies to compete (in contrast to AFPs who operate PWs) to offer a present value of annuity payments which is relatively high compared with the present value of future PWs. Moreover, for many workers it is easier to withdraw their pension savings through annuitisation as they are marketed by insurance company sales agents who prepare calculations, process the paperwork and sell annuities to clients (rather than guiding them to AFPs delivering PWs), often before the normal retirement age. The normal retirement age in Chile is 65 for men, 60 for women, but early withdrawal is allowed once a specified minimum amount of pension saving has been accumulated (the term early withdrawal seems more appropriate than early retirement, since the initial withdrawal of pension wealth often does not coincide with retirement from the labour market). Since 1988, workers have been permitted to stop contributing and start withdrawing once their replacement rate is 50% of their own wage and 110% of the MPG. With recognition bonds considered as part of the accumulated amount, and high rates of return to retirement accounts during the 1980s and 1990s (on average in excess of 10% per annum in real terms), a high proportion of workers met the early withdrawal conditions once they reached their 50s. For those who qualified, it is rational to withdraw early, stop contributing, and either consume or save in a more flexible form. In order to reduce the incidence of early withdrawal, the benefit formula is being tightened to make early withdrawal less attractive: the replacement rate was raised to 70% of the average wage and 150% of the MPG (and

156 CHAPTER 4. THE NORMALISATION OF CHILE S PENSIONS SYSTEM 159 since July 2008: 80% of the APS threshold, see Table 4.1), and the reference wage now refers to an average over the last ten years with a maximum to the number of months without contribution that can be included (Iglesias, 2009). Hence, it is no surprise that so many clients opt for annuities. Workers with small accumulations retire at the normal age and take PW pensions. But the majority of workers withdraw early and almost 90% of them purchase annuities (Table 4.2). This large percentage which is far greater than in other countries seems to be explained by incentives and constraints imposed by guarantees and regulations, as well as by the limited information available to AFP members. One of the aims of the 2008 reform is to improve the information stream to AFP members and help them make more informed choices on their withdrawal decision. 64 Low coverage: the Achilles heel of private protection arrangements Coverage of the AFP programme increased markedly during the period because of the many workers who switched from the PAYG system to the individual private AFP programme (Box 4.1), particularly in the period After this initial period, coverage in terms of the working-age population in employment increased from about 40% in 1987 to 60% in 2007 (Figure 4.4, Panel A). By OECD comparison, coverage is low: when measured against the labour force, pension coverage in Chile was about 54% in 2005 (Figure 4.4, Panel B), well below the OECD average of 83% (OECD, 2008b). Coverage of pension schemes is also linked with national income. In countries where significant parts of the working-age population are in low productivity (e.g. small-scale agriculture), low-wage (and/or informal) employment, coverage of pension systems is relatively low. In recent years, Chile s employment growth has mainly been in the low-wage employment sector (Chapter 1), and there is no reason to believe that such workers will be more able to finance their health and pension contributions than in the past. The 2008 pension reform increased generosity of minimum pensions is financed out of general government (tax) revenue, which is tantamount to a redistribution of wealth from stronger growth sectors (e.g. the mineral sector) to low-productivity ones. Financing pension reform out of general government revue seems to be the best of available options, but the strategy is not without risk. For example, the cost of the 2008 reform is projected to 64. The reform will establish a fund for pension education that provides information on all aspects of the Chilean pension system, including information on yields, returns, annuities, etc. across different AFPs. This includes projections on future pension payments based on current assets and different contributory profiles (e.g. a full record or none at all). A network of accredited pension advisers will also be developed.

157 160 CHAPTER 4. THE NORMALISATION OF CHILE S PENSIONS SYSTEM amount to 1.3% of GDP per annum in However, if plans to extend private pension coverage among a greater group of workers (see below) were not to have the desired effect in terms of mandatory individual pension saving, the total cost of the 2008 pension reform may well be higher than originally envisaged. Figure 4.4. Pension coverage is increasing but remains low in international comparison Panel A. AFP contributors as a proportion of the working-age population (15-64) in employment, Panel B. Coverage of mandatory pension schemes as a proportion of the labour force,

158 CHAPTER 4. THE NORMALISATION OF CHILE S PENSIONS SYSTEM 161 Figure 4.4. Pension coverage is increasing but remains low in international comparison (cont d) Panel C. Coverage of mandatory pension schemes related to national income per capita (log scale), Contributors refer to contributors in December of a given year who have not yet retired and who made the contribution in the month for the wages they earned that month. 2. Australia, Austria, Belgium, Canada, Finland, France, Germany, Ireland, the Netherlands, Sweden, the United Kingdom and the United States are the countries included in the box in the figure. Source: Panel A: Contributors, Superintendencia de Pensiones; Employment data, both sexes, (avg. November year x to January year x+1); INE; Panel B: OECD (2008b), Pensions at a Glance: Asia/Pacific; and, Panel C: World Bank Pensions database. There may be fewer female than male AFP members and contributors, but, ever since the introduction of the AFP programme, coverage among females as a percentage of female employment is higher than for men, by about 4 to 5 percentage points (Iglesias, 2009). However, men are more likely than women to have a full contributory record during their membership period (Figure 4.5). Almost half of the female AFP members and one-third of the male members made no pension contribution from January 2004 to September 2006 (Figure 4.5); over the same period half of the men who contributed to an AFP fund contributed in full (as opposed to one-third of the women, with no discernible gender difference in contributory behaviour of those with a partial contributory record over the period). Contribution densities are also considerably higher among groups of workers with higher

159 162 CHAPTER 4. THE NORMALISATION OF CHILE S PENSIONS SYSTEM income and education levels, and workers who are at least 45 years of age women often defer affiliation until middle age (Asher and Vasudevan, 2008); contribution densities among low-income workers and workers with low educational attainment are limited (see also Cox, 2006). Figure 4.5. The contribution density 1 among female workers is relatively low Distribution of members by gender along the intensity with which they have made pension contributions during the January 2004-September 2006 period Women Men % 20-40% 40-60% 60-80% % 100% 1. The contribution density is the number of months in which an individual has made pension contributions as a proportion of months of membership in the AFP scheme. Source: Based on data provided by the Government of Chile. Another key aspect of AFP coverage is the low participation by selfemployed workers. Coverage for this group of workers is not mandatory and is historically very low. In December 2007, only out of 1.8 million self employed made pension contributions, and self-employed contributors accounted for 1.6% of all AFP contributors, whilst self-employed workers represented 27% of total employed people in the country (Iglesias, 2009). Workers with low contribution densities are unlikely to achieve the replacement rates shown above, and (without public coverage) are particularly vulnerable to the poverty risk in old age.

160 CHAPTER 4. THE NORMALISATION OF CHILE S PENSIONS SYSTEM 163 Policy initiatives to increase coverage There are different reasons why different groups do not save (enough) for retirement. Often young people are simply not (yet) concerned about their retirement. Mothers who provide full-time care for their children have withdrawn from the labour force (and stopped making contributions). Other workers, including many of the self-employed, are unlikely to make voluntary pension contributions unless the relevant returns are considerably higher than returns to market saving. Otherwise, they will prefer liquidity, especially if it concerns cash-constrained, low-income households. 65 Recent reform therefore tries to expand coverage through a mix of compulsion and financial incentives to make pension savings more attractive for groups with a low contribution density: 1. The introduction of mandatory coverage of pension contributions for self-employed workers who pay income tax (this measure is scheduled for gradual implementation over the period). The Internal Revenue Service will take over the collection of pension contributions (paid monthly or yearly based on the individual income declaration). The contribution rate will be the same as for salaried workers (with the same maximum threshold) and determined over 80% of declared income. Mandatory coverage will not be extended to some other groups of self-employed workers who are not obliged to declare income, e.g. artisanal fishermen and taxi drivers. 2. To encourage pension saving among young workers and increase formal employment: employers of contributors aged with an income below 1.5 times the minimum wage (about USD 430), will receive a subsidy for their first 24 months of employment equal to 5% of the minimum wage (around USD 14). In addition, these participants will receive a similar bonus (USD 336 or 24 times USD 14) in their personal account from 2011 onwards. Some young workers may be covered in To encourage female pension saving, the state will pay contributions for mothers corresponding to 18 monthly MWs per child (about USD 518). 65. Evidence on non- or under-declaration of income for pension contribution purposes suggests this is not the major issue (again, the self-employed are not obliged to contribute): accumulated unpaid contributions, both declared and undeclared, are equivalent to less than 0.7% of the pension funds (Iglesias, 2009). The low non-compliance rate is also related to penalties for employers who do not pay contributions on behalf of their workers or who do so late. AFPs are legally obliged to recover outstanding contributions and have strong financial incentives to do so, as they can only charge management fees if contributions are registered in personal accounts (Iglesias, 2009).

161 164 CHAPTER 4. THE NORMALISATION OF CHILE S PENSIONS SYSTEM This bonus will then receive a rate of return equal to the one obtained by pension fund C from the date the child is born until the mother reaches 65 years of age (despite a standard pension age at 60 for women). Also, in case of divorce, women can receive up to 50% of spousal pension entitlements, depending on their employment history. However, it is unclear as to why reform did not increase the normal withdrawal age for women to 65 years of age in view of the life expectancy trends and the recommendations of the Marcel Commission. Thus far, the success of financial incentives to increase pension saving among the self-employed has not been very effective, and it remains to be seen to what extent these new initiatives will be more effective. Furthermore, the self-employed who are currently charged income tax on the basis of their VAT invoices may well change their behaviour faced with an increase of almost 20% of contributions (including towards health coverage) on their income. Because of this, reform may well lead to a decline in reported income for this group of workers, and much will depend on the effectiveness with which the tax authorities can implement the scheduled reform. Disability and survivor 66 coverage: reduce cost-shifting while maintaining efficiency In Chile, workers who fall ill, in the first instance can claim a sickness cash benefit at 100% of gross earnings up to a maximum of UF 60 (about USD 2 445) per annum. This benefit is financed by health insurers without specified maximum duration, so that health insurers are responsible for alerting clients to apply for a disability benefit, if they have not initiated such procedures themselves already. If incapacity is employment-related clients may have a choice to apply for a benefit deriving from occupational accident and diseases legislation. 67 This scheme involves an employer contribution of 0.9% of wages and an additional contribution which varies with occupation and is reassessed annually, but which cannot be higher than 3.4% of wages, for example, in glass and ammunition industries. Contribution rates concern all workers in a company regardless of their specific task. 66. Survivor pensions can be received by widows (and since 2008 reform widowers), children (legitimate and biological), the mother of the biological children, and if there are no such beneficiaries, entitlement passes to the parents of the deceased AFP member. 67. Someone who applies for an employment-related occupational accident and diseases benefit cannot simultaneously apply for a benefit under the AFP scheme; claimants can apply for the other benefit if the initial application to either scheme was not successful.

162 CHAPTER 4. THE NORMALISATION OF CHILE S PENSIONS SYSTEM 165 AFP contributors not yet of pensionable age are entitled to a disability benefit in case of a non-employment related accident or illness. The defined-benefit payment is financed out of the capital in the AFPs members account and (if necessary) a supplement ( the additional payment ) which is paid by the AFP and funded out of a group-policy the AFP takes out for its members with a life insurance company. 68 Compared with the average wage, the defined-benefit payment generates a gross replacement rate of around 60% (the benefit is 70% of the individual s average historical wage in case of total disability and 50% of that wage in case of partial disability). Figure 4.6 shows this is on par with gross replacement rates in other OECD countries; net (after-tax) replacement rates in OECD countries are often about 10 percentage points higher. Figure 4.6. Disability benefit replacement rates in Chile are on par with other OECD countries Gross replacement rates for disability benefit payments paid under the mandatory pension programmes, percentage of average wages, Source: Chile: Iglesias (2009); other countries: OECD, Breaking the Barriers - Sickness, Disability and Work (various issues). 68. When a worker covered by the contract becomes disabled or dies, the so-called additional payment is deposited into the individual s pension account. This additional payment is equal to the difference between the necessary capital to finance the legally prescribed level of pensions (in other words, the present value of estimated future pension payments) and the balance accumulated in his/her personal account at the time of disability or death.

163 166 CHAPTER 4. THE NORMALISATION OF CHILE S PENSIONS SYSTEM For a client to be recognised as fully disabled, he or she must have lost as least two-thirds of his/her earnings capacity (the system classified those with a loss of half to two-thirds of earnings capacity as partially disabled). Initial claims are assessed along defined disability criteria by 21 regional medical commissions of three doctors appointed by the Superintendencia of AFPs. Clients who were declared to be disabled were re-assed three years later. During this three year period, AFPs could not use the capital in personal accounts to fund the claim, but they had to pay do so out of their administrative fees (with which AFPs bought a group-insurance policy for their members). Because of this, AFPs had financial incentives to curtail the costs of disability by limiting the number of successful claims. Assessment procedures include participation by AFPs and insurance companies and both agencies pursue a strict application of the rules, fighting applications before they even get to the medical commission on technical grounds and appealing assessment outcomes where possible. Even when the system allows it, public officials may have limited direct financial incentives to appeal against successful disability claims, so that in many countries with public systems, appeals are only actively pursued by unsuccessful claimants. In the Chilean system, both AFPs and unsuccessful claimants have financial incentives to appeal and the number of successful appeals by both parties approximately cancels each other out, which contributes to a relatively low incidence of approved disability cases. James et al. (2008) show that the hazard rates of workers becoming disability pensioners in the AFP programme are 65-80% lower than under the previous public system, and that the reduction in disability hazard is related to systemic targeting of those with the most severe medical conditions. The low hazard rates in Chile s private disability programme suggest it has chosen to minimise the chance of incorrectly granting claims, at the risk of turning down many deserving clients. However, this has been done at the cost of public budgets, as unsuccessful applicants sought to receive a minimum pension (to be eligible to the MPG clients who have been declared disabled need ten years of contributory record as opposed to the 20 years for old-age claimants) or a means-tested PASIS benefit. 69 Abstracting from the cost to the public purse, the cost of the private insurance of disability and survivor risks is low in international comparison at less than 1% of wages, about a quarter of what it would typically be in many European OECD countries (e.g. Andrews, 1998). However, this 69. It is unclear to what the assertive approach of AFPs in disability-assessment procedures has had an upward effect on the number of claimants of the occupational accident and disease programme.

164 CHAPTER 4. THE NORMALISATION OF CHILE S PENSIONS SYSTEM 167 outcome should be put in context with the limitations in AFP coverage and the fact that disability is concentrated among poor families who are least likely to be covered by the AFP scheme. Changes to disability and survivor insurance The 2008 reform with respect to disability and survivor insurance included the following items: The disability and survivor insurance premium will be differentiated across genders: at present women are cross-subsidising men as the actual costs are about 45% lower than for men. 70 Mandatory re-assessment of certified totally disabled clients after three years was abolished as claims are not frequently reversed, so that in these cases the insurer can now fund the disability claim out of the personal account from the outset (partially disabled continue to be re-assessed). 71 In future, rather than competing against each other to select clients whose disability risk is relatively low, a consortium of all AFPs will use a public auction mechanism to buy a common disability and survivor insurance contract for two-year periods from a life insurance company (or a group of such companies). The resulting disability and insurance fee will then be applied to existing members regardless of the AFP to which they belong. Furthermore, the disability and survivor insurance fee will no longer be part of the administrative fee (see above), but will be separately identifiable, and from July 2011 onwards will be paid by employers (rather than workers). 72 It is too early to assess the overall effect of these reforms, but experience with reform of disability programmes in OECD countries holds some lessons. From a narrow assessment-procedure perspective, it is desirable to avoid the cost of a procedure which seems to have little effect on disability 70. In 2004, the average cost of disability and survivor insurance was 0.86% of wages; while the cost for women was 0.57%, the cost for men was 1.01%. Because of the reform, men and women will start paying the cost of the insurance for men, but the difference between this cost and the cost for women will be deposited in women s pension s accounts. 71. James et al. (2008), report that out of 100 initial applicants, 42 claimants were re-evaluated after three years, of which 40 were declared permanently disabled. 72. The notion of employer responsibility for paying disability and survivor insurance will be phased in gradually; the fee will be charged to employers with 100 or more contracted workers as from 1 July, 2009, to be extended to all employers from 1 July 2011 onwards.

165 168 CHAPTER 4. THE NORMALISATION OF CHILE S PENSIONS SYSTEM status. However, if re-assessment procedures were abolished to reduce costs for AFPs this could also have been achieved by allowing the AFP to charge back the costs of the initial three years temporary payment to the clients personal account, if the full disability status of the client was upheld after three years. The drawback of this measure is that it suggests that once fully disabled, always fully disabled, and as policy messages go, that is a bad one for different reasons. It provides incentives to potential applicants (and perhaps in future some AFPs) to apply for a steady stream of future income payments under the programme; and, it ignores the capacities of persons with disabilities. It is also unclear as to why employers are made responsible for disability contributions. Employers are likely to pass on the cost of this to their workers in future. Furthermore, experience in OECD countries shows that increasing the transparency of disability costs in this way has little effect on employer-behaviour. Experience-rating of disability premiums is likely to be more effective in terms of the preventive measures by employers, as the experience in the Netherlands shows (OECD, 2008c). In fact, the Chilean system involves sectoral experience-rating in its employment-related occupational accident and diseases scheme. Policy may consider to extend experience-rating to enterprises for both the employment-related and private disability schemes. The recent reform may well increase the costs of the private disability scheme significantly, perhaps by as much as 25% (Iglesias, 2009). In particular, this increase in cost would be related to the increased coverage of the disability and survivor insurance (more self-employed low-income workers with an elevated disability risk) as well as new institutional arrangements. In future, individual AFPs will not have incentives to actively follow disability assessment, as this no longer affects their operations. In fact, individual AFPs may well help their clients with the administrative procedures to obtain a permanent disability benefit at the expense of the insurance company (or a consortium of insurance companies) which won the tender. The insurance company will in any case try to pass on costs to employers, who in turn will pass on cost to workers. In any case, since the tender-period only concerns two years; the relevant insurance company has limited incentives to invest in the quality of their assessment procedures, and the ensuing higher incidence of disability payments is likely to increase premiums. To the extent that the rise in costs of the private disability programme reflects the expansion of worker coverage 73 (and reduce public sector 73. The coverage of some groups of workers may increase costs, while the tightening of early retirement regulations will increase coverage among older workers. On the other

166 CHAPTER 4. THE NORMALISATION OF CHILE S PENSIONS SYSTEM 169 coverage), this is understandable. However, Chile should avoid the experience of some OECD countries which found that cost of disability benefit schemes are hard to cut back once the inflow of new claimants into these schemes has increased significantly (OECD, 2004, 2006b, 2007c and 2008c). Recent reform involves a risk that less assertive application of assessment procedures may lead to a growth in the number of recipients. It is of course too early to assess the effect disability reform will have, but it needs to be closely monitored to ensure that the right people receive the assistance to which they are entitled. 3. Conclusions More than 25 years of experience with mandatory private pensions has restored confidence in pension saving in Chile. One can argue that individual pension schemes may involve relatively high administrative costs compared with collective voluntary occupation pension saving as in the Netherlands, or that a system of individual accounts involves little interpersonal redistribution (a sine qua non for being classified as a social programme). However, such caveats need to give due account to the fact that Chile, unlike most OECD countries, had a very bad experience with corrupt pension schemes. The pre-funded Chilean system has performed well over the years. It has deepened Chilean capital markets and helped economic development. Furthermore, it is not very susceptible to the dynamics of ageing populations, and in that way the Chilean system is ahead of many OECD countries which are still grappling with the issues of necessary pension reform in view of prospective increases in public spending commitments. The rates of return on investment have been higher than originally envisaged. However, competition is limited in the pension market, and recent reform aims to generate price competition by having AFPs compete for a share of the market (new entrants) rather than trying to stimulate competition within the market. At present, operational profits of AFPs seem to be at a structurally higher level than prior to the late 1990s, and this does not seem to be related to a change in cost structures. There seems to be room to reduce user charges beyond the changes generated by reform of disability insurance. The 1981 pension reform failed to achieve one of its key objectives: ensure adequacy and coverage for the vast majority of Chilean workers. As in most OECD countries, the density of full-time earners with a full contributory record over 30 years is diminishing and there are increasing hand, the projected increase of coverage among younger workers will off-set these effects to some extent.

167 170 CHAPTER 4. THE NORMALISATION OF CHILE S PENSIONS SYSTEM numbers of workers with patchy employment profiles. Contribution densities are particularly low among low-income workers with low levels of educational attainment, young workers, female workers, and the selfemployed. Recent reform targets increased coverage among these groups. Public intervention is needed if a predominantly private pension system wishes to be successful in term of adequacy and coverage, and the Chilean authorities deserve to be commended for their introduction of a basic solidarity pension as of right. In addition, new measures have been introduced to stimulate private pension saving, so that Chile is moving toward the international benchmark model of a basic public pensions, mandatory second-tier pension saving, topped up with third-tier pension savings for those who want to save more. The Chilean authorities started to improve benefit support to the elderly well before recent initiatives to support the working-age population unfolded (Chapter 3), and as a result the poverty rate among the elderly is relatively low. However, there is no room for complacency and the success of recent reform cannot be taken for granted in the future. Reform, rightly, aims to increase coverage of pension contributions to groups to low-income households and the self-employed. However, there are concerns that these initiatives may not be successful. For example, the implicit tax rate of up to 30% may be a deterrent for many low-income workers to save for retirement and some of the self-employed may stop reporting income to the tax authorities rather than pay contributions towards pension insurance. The private AFP scheme has managed to limit access to its disability provisions. This has been at the public expense, since unsuccessful applicants are likely to seek recourse to publicly-financed minimum income benefits. Recent reform is likely to increase cost of the private disability programme as coverage is extended to groups of the self-employed, and cost may rise further because of institutional changes with AFPs no longer having incentives to keep costs down. It is too early to assess the disability reform will have, but it needs to be closely monitored to ensure the right people receive the assistance to which they are entitled. Finally, it is not immediately obvious why employers have been made responsible for disability contributions, and experience in OECD countries has shown that increasing the transparency of disability costs in this way has little effect on employer-behaviour. Experience-rating of disability premiums is likely to be more effective in terms of the preventive measures by employers, and the Chilean authorities may wish to consider introducing this principle in the private disability scheme.

168 CHAPTER 4. THE NORMALISATION OF CHILE S PENSIONS SYSTEM 171 Annex 4.A1. Background data to Chile s private pension system Mandatory pension contributions Table 4.A1.1. Most outlays on pension payments are through annuities rather than programmed withdrawals Pension benefits receipts and outlays (billion of pesos, current currency) Pension fund outlays towards annuities 1 Regular pension payments Temporary payments 2 Annuity payments Other 3 Total Old age Invalidity and survivors Old age Invalidity and survivors Old age Invalidity and survivors Invalidity and survivors Old age Invalidity and survivors All The present value of annuities bought that year (equivalent to capital transferred to life insurance companies to buy annuities). 2. A pensioner can buy a deferred annuity, with pay-outs starting n years from the moment a person ceases to make contributions. From that moment until annuity payments start, the person can receive a temporary benefit financed out of the personal balance which was not used to buy the deferred annuity. 3. Other payment as covered by insurance, e.g. benefits arising from accidents, illness and death. Source: Estados Financieros AFP and SAFP ( Table 4.A1.2. AFP portfolio characteristics and yields and default age rules for clients, 2008 Portfolio Average Limit on investment in equities, as a Default age designation 2 yield 1 proportion of balance in account Minimum Maximum Men Women A 18.2% Not applicable Not applicable B 13.1% < 35 < 35 C 9.9% D 7.5% > 50 > E 4.3% Largely fixed income instruments Not applicable Not applicable 1. Annual average real rate of return by type of portfolio over the period ( download 12 September 2008). Each year each AFP must guarantee that the average real return in the last 36 months is not lower than the lesser of i) the average real return minus 4 percentage points for the funds A and B (with a higher equity exposure), and minus 2 percentage points for the funds C, D and E (with a higher proportion of fixed income securities) or ii) 50% of the average real return of all the funds. 2. Older clients (men over 55, women over 50) cannot choose portfolio 1 in any case, while beneficiaries of PWs in the same age group cannot choose portfolios A and B. Switching among portfolios is allowed, but only the first two changes within a single year are free of charge. Source: Estados Financieros AFP and SAFP (

169 172 CHAPTER 4. THE NORMALISATION OF CHILE S PENSIONS SYSTEM Trends in minimum payment rates The Minima Pension Garantia (MPG) was about CLP (about USD 180) per month in 2007, but this varied with age (Iglesias, 2009). In 2007, the PASIS payment amounted to about one-third of the minimum wage, while the MGP was equivalent to two-thirds of the minimum wage. Average wages of contributors to the mandatory pension system were about eight times as high as the PASIS payment (Figure 4.A1.1). Since 2004 minimum pension payments increased relative to (minimum) wages, and this trend continued with the July pension reform: the PBS payment in 2008 was almost 25% higher than the 2007 PASIS payment in nominal terms. Figure 4.A1.1. Evolution of the minimum wage, average wage of contributors to the mandatory pension system (AFP) and minimum pension payments In Chilean pesos of December Minimum wage Average wage AFP MPG(1) PASIS(2) Minimum pension guarantee payment for retirees less than 70 years of age. 2. PASIS payment for retirees from over 70 to less than 75 years of age. Source: Iglesias (2009).

170 CHAPTER 4. THE NORMALISATION OF CHILE S PENSIONS SYSTEM 173 Annex 4.A2. The Chilean pension market: competition, individual choice and financial risk management In 2008, there were five Administradora de Fondo de Pensiones (AFPs) Capital, Cuprum, Habitat, Planvital and Provida, which operate pensions saving accounts and in this process fulfil a record-keeping, investment and benefit-payment function. Since 1981, the importance of the pension market industry has grown spectacularly: from about 1.4 million clients and pensions assets worth about USD 0.33 billion to almost 8 million clients by the end of 2007 and assets amounting to USD 111 billion. Competition and the lack of it AFPs charge fees on the payment of contributions, payment of programmed withdrawals (PWs) usually a proportion of 1.25% of the pension payment, and on management of voluntary pension savings accounts (AFPs could charge fees on transfer of client s accounts to a different AFP, but this does not happen in practice). AFPs have to charge identical fees across their clients, and since there are no substantial differences in services, operational costs (or fees) between clients, high-income workers are financially most interesting to AFPs. In December 2007, administration fees amounted to about 2.4% of wages, which included a 0.95% fee on disability and survivor fees; two AFPs also charged a flat amount in addition to the variable fee (in future AFPs will no longer be allowed to charge flat fees to increase market transparency). Administrative charges as a proportion of the taxable wage were highest in the early 1980s when APFs charged high fees to recoup part of the their income losses due to low wages and employment, but from 1984 to 2005 administrative fees fell from 3.6 to 2.25% of the average taxable wage in 2001, to increase to 2.4% in 2007 (Figure 4.A2.1). International comparisons of administrative charge ratios are fraught with difficulties, because of systemic differences such as, for example, different types of charges and contributory periods. Gomez-Hernandez and Stewart (2008) consider 40-year charge ratios across 21 countries, and find that this charge ratio of the Chilean (and Israeli) systems are below average and below charge ratios for (in ascending order) Polish, Slovak, Hungarian,

171 174 CHAPTER 4. THE NORMALISATION OF CHILE S PENSIONS SYSTEM Czech and Turkish defined-contribution pension systems (the result has to be interpreted with care as the chosen methodology favours a system which charge fees on wages, rather than on assets). Administrative fees may not be overly high in international comparison, but that does not mean they could not be reduced in view of market and cost structures in Chile. 4.0 Figure 4.A2.1. Administrative charges declined until 2001 As a percentage of client earnings In 2007, two AFPs also charged a flat fee. Source: Iglesias (2009) and the Superintendencia de pensiones ( Different AFPs charge different fees, but differences are small. Many clients also find it difficult to obtain and understand all relevant information about pension options and their implications for future pension wealth. The mandatory nature of the system may also contribute to a certain degree of apathy among clients, as many of them seem to be unaware or uninterested in the relatively small differences in fees across AFPs (Asher and Vasudevan, 2008). The low price elasticity of demand limits price competition (e.g. Berstein and Cabrita, 2007; and Valdés, 2005). In the absence of price competition (and yield competition, see below), competition among AFPs mainly concerned (expensive) direct sales efforts and marketing campaigns. Indeed, in the mid-1990s, AFPs competed aggressively on the basis of marketing strategies, with sales agents luring in clients with more or less expensive presents and gifs. At peak, marketing

172 CHAPTER 4. THE NORMALISATION OF CHILE S PENSIONS SYSTEM 175 expenditure amounted to one-third of fee income in 1997 (up from about 10% in 1990; Iglesias, 2008), but with a tightening of the rules on switching between funds in October 1997, the number of salesmen fell from in 1997 to in 1998 (Asher and Vadusevan, 2008; and Iglesias, 2009). The increase in sales and marketing activities, reduced profit margins (which increased again since 1998) and contributed to a concentration of AFPs in the market from 21 in 1993 to 5 in Entry to the pension market is free (subject to fulfilling basic conditions), but economies of scale effectively dictate an AFP needs about half a million clients to survive in the market (Iglesias, 2008). There are few restrictions on AFP ownership except, significantly, for commercial banks, mutual funds and insurance companies that cannot hold shares in AFPs, which helps shield pension-fund managers from competition by other financial intermediaries. On the whole, the operation profit margin of AFPs was on average 18% from 1983 to 1998, but with the shake-out of salesmen operating profit margins have exceeded 25% (Figure 4.A2.2). Initially, the increased profits were at least partially passed on to clients, but since 2001 administrative charges have stopped falling (Figure 4.A2.1), while operating profits have been fluctuating around 30%: almost twice as high as during the first 15 years of the Chilean pensions system. This suggests that there is room left to reduce user charges in the AFP system. Figure 4.A2.2. AFP profits seem to be at a higher level than in the 1990s 50 AFP industry operating profit margin The operating profit margin is the difference of operating income and expenses as a proportion of operating income. Source: Anexo Estadistico of the Sistema Chileno de Pensiones as published by the Superintendencia de pensiones (

173 176 CHAPTER 4. THE NORMALISATION OF CHILE S PENSIONS SYSTEM Increasing individual choice and the rate of return on investment The experience with financial management of PAYG funds prior to reform led to strict public regulation of AFPs in terms of portfolio selection, investment rules and market behaviour. There is, however, a balance between regulation and individual choice as well as regulation and rates of return on investment, and over the years, the Chilean authorities have gradually relaxed rules to increase both individual choice and yields. In terms of portfolio selection, originally each AFP offered only one portfolio to clients. Since 2002, AFPs can offer five different portfolios and risk profiles to their clients. Portfolio A invests 40 to 80% of capital in the individual account in equities: portfolio B, 25 to 60%; portfolio C, 15 to 40%, portfolio D, 5 to 20%, while portfolio E mainly invests in fixedincome instruments. In order to avoid clients who are close to retirement or those who are already retired taking considerable risks, they are not allowed to choose the more risky portfolios depending on their age (Table 4.A1.2). Furthermore, if clients upon joining an AFP for the first time do not reveal a preference, they are allocated a portfolio according to their age (portfolio B up to age 35; portfolio C for middle-aged workers; and portfolio D for older workers, see Table 4.A1.2). By the end of 2007, almost 3 out of 8 million clients had chosen a portfolio other than the default one, and 75% of these clients had opted for portfolios A or B; younger clients are more likely to take more risky investments (Iglesias, 2009). In the end, portfolios B and C cover about 40% each of all AFP clients while the interest for both very risky and risk-averse portfolios is limited (Figure 4.A2.3). Figure 4.A2.3. Most AFP members have investment portfolios with intermediate risks Distribution of AFP members by investment portfolio, 30 June 2008 Portfolio D 8.4% Portfolio E 0.6% Portfolio A 13.9% Portfolio C 37.3% Portfolio B 39.8% Source: Superintendencia de pensiones (