Address for correspondence: Milko Matijascic, IPEA – Brasília, SBS, Quadra 1, Bloco J, Ed. BNDES Térreo, 70076-900 Brasilia-DF, Brazil; Email: email@example.com. Stephen J. Kay, Federal Reserve Bank of Atlanta, 1000 Peachtree St. N.E., Atlanta, GA 30309, United States; Email: firstname.lastname@example.org.
Federal Reserve Bank of Atlanta, Georgia, United States
Address for correspondence: Milko Matijascic, IPEA – Brasília, SBS, Quadra 1, Bloco J, Ed. BNDES Térreo, 70076-900 Brasilia-DF, Brazil; Email: email@example.com. Stephen J. Kay, Federal Reserve Bank of Atlanta, 1000 Peachtree St. N.E., Atlanta, GA 30309, United States; Email: firstname.lastname@example.org.
To achieve national goals defined by the 1988 Brazilian Federal Constitution, cash benefits alone are insufficient in the absence of more robust social services to reduce inequalities and improve social cohesion. The Constitution, albeit of national importance and international significance, has not addressed many institutional and administrative weaknesses in the design of the national pension system. Although coverage has been increased and inequality reduced, these measures are not sufficient. Brazil's ambitions to further develop social policies (and, indeed, to live up to its accorded international status as a social policy leader) may be constrained by an over-reliance on conditional cash transfers such as those provided under the Bolsa Família programme. Brazil faces a major political-economy challenge in addressing all these issues because the policy reform process is difficult, and, more importantly, because of the embedded role of vested interests. Moreover, Brazil must tackle these issues in the face of growing fiscal pressures, which could weaken the current political legitimacy of social policy and undermine important recent successes.
Le modèle de retraite brésilien: les défis à relever
Les programmes qui versent des prestations monétaires ne permettront pas, à eux seuls, d'atteindre les objectifs nationaux définis par la Constitution fédérale brésilienne de 1988 si des services sociaux plus performants ne sont pas mis en place pour réduire les inégalités et renforcer la cohésion sociale. Malgré son importance sur le plan national comme international, la Constitution ne remédie pas aux nombreuses faiblesses institutionnelles et administratives inhérentes au système national de retraite. L'extension de la couverture et la réduction des inégalités, pour réelles qu'elles soient, ne sont pas suffisantes. En outre, il est possible qu'une dépendance excessive à l'égard des dispositifs de transferts monétaires conditionnels tels que Bolsa Família empêche le pays de développer encore les politiques sociales comme il l'ambitionne (et de mériter ainsi sa réputation internationale de «fer de lance» en matière de politique sociale). Or, le Brésil est, à cet égard, confronté à un énorme défi en termes d'économie politique non seulement parce que réformer est un processus difficile, mais aussi et surtout en raison du rôle que jouent les intérêts catégoriels. De surcroît, le pays doit affronter ces problématiques sur fond de difficultés budgétaires croissantes, ce qui pourrait affaiblir la légitimité politique de la politique sociale et remettre en cause certains succès récents importants.
El modelo de pensiones del Brasil: programa pendiente
Para lograr los objetivos nacionales definidos por la Constitución Federal del Brasil de 1988 las prestaciones monetarias por sí solas son insuficientes si no se cuenta con servicios sociales más sólidos para reducir las desigualdades y mejorar la cohesión social. La Constitución, a pesar de su importancia en el plano nacional e internacional, no aborda muchas de las deficiencias institucionales y administrativas del régimen de pensión nacional. A pesar de que la cobertura se ha extendido y se han reducido las desigualdades, estas medidas no han sido suficientes. El objetivo del Brasil de elaborar nuevas políticas sociales (y, de hecho, de estar a la altura de la condición acordada internacionalmente de «líder de la política social») puede verse limitada por la excesiva dependencia de las transferencias monetarias condicionadas, como las que se conceden en el marco del programa Bolsa Família. El Brasil se enfrenta a un reto político y económico importante al tratar de abordar todas estas cuestiones, ya que el proceso de reforma de las políticas es difícil y, sobre todo, debido al papel intrínseco de los derechos adquiridos. Además, el Brasil debe afrontar estas cuestiones en un contexto de creciente presión fiscal, que podría debilitar la legitimidad política actual de la política social y socavar los importantes éxitos obtenidos recientemente.
Das brasilianische Rentenmodell: die anstehende Agenda
Die in der brasilianischen Verfassung von 1988 verankerten Ziele können nicht mit Geldleistungen allein erreicht werden, wenn nicht zugleich robustere Sozialleistungen zur Verringerung der Ungleichheit und zur Verbesserung des gesellschaftlichen Zusammenhalts eingeführt werden. Obwohl die Verfassung wichtig ist für das Land und auch international Bedeutung hat, bestehen nach wie vor viele institutionelle und administrative Mängel bei der Gestaltung des staatlichen Rentensystems. Die Deckung konnte zwar ausgeweitet und die Ungleichheit verringert werden, aber diese Maßnahmen sind nicht ausreichend. Die Anstrengungen Brasiliens, die Sozialpolitik weiterzuentwickeln (und auch den anerkannten internationalen Status als „Vorreiter der Sozialpolitik“ zu verteidigen) könnten dadurch beeinträchtigt werden, dass sich das Land zu stark auf bedingte Geldtransfers wie diejenigen des Programms Bolsa Família verlässt. Brasilien steht mit der Bewältigung dieser Fragen vor einer großen volkswirtschaftlichen Herausforderung, da politische Reformen schwierig umzusetzen sind und vor allem gut verankerte Interessengruppen eine Rolle spielen werden. Überdies muss Brasilien diese Fragen vor dem Hintergrund einer immer angespannteren Finanzlage angehen, welche die gegenwärtige Legitimierung der Sozialpolitik und die wichtigen neusten Erfolge gefährden könnte.
Бразилъская пенсионная моделъ: будущая повестка
для достижения изложенных в Конституции Бразилии 1988 года националъных задач по сокращению неравенства и укреплению социалъной сплоченности одних лишъ денежных пособий недостаточно – необходимы более эффективные социалъные услуги. Конституция, несмотря на свое националъное и международное значение, не решила многих институционалъных и административных проблем националъной пенсионной системы. хотя удалосъ расширитъ охват и снизитъ масштабы неравенства, данных мер явно недостаточно. усилия Бразилии по далънейшему развитию социалъной политики (и ее стремление соответствоватъ своему международному статусу “лидера социалъного обеспечения”) могут бытъ ограничены ее излишним доверием к программам условных денежных трансфертов, таким как Bolsa Família (“Семейный кошелек”). Перед Бразилией стоит сложная политэкономическая задача, посколъку процесс политических реформ всегда требует значителъных усилий, к тому же он часто сопряжен с конфликтом интересов. Кроме того, Бразилии приходится решатъ все эти вопросы в условиях серъезных финансовых ограничений, которые могут снизитъ легитимностъ социалъной политики и свести на нет ее недавние успехи.
لتحقيق الأهداف الوطنية المعرّفة في دستور الاتحاد البرازيلي (الفدرالية البرازيلية) لعام 1988، لا تُعتبر المنافع النقدية وحدها كافية في غياب الخدمات الاجتماعية الأكثر قوّة ونجاعة للتخفيف من الفوارق وتحسين التماسك الاجتماعي. على الرغم من الأهمية الوطنية والدلالة العالمية، لم يتناول الدستور العديد من نقاط الضعف المؤسسية والإدارية في تصميم نظام التقاعد الوطني. وما زالت هذه التدابير غير كافية وإن تمّت توسعة الشمول والتخفيف من الفوارق. ولعلّ اعتماد البرازيل المبالغ فيه على التحويلات النقدية المشروطة كتلك التي اشتمل عليها برنامج "بولصا فاميليا - Bolsa Família يقيّد طموحات البرازيل الساعية إلى تطوير السياسات الاجتماعية بشكل أكبر (وهي في حقيقة الأمر تسعى إلى الارتقاء إلى المنزلة العالمية التي عُرفت بها "كرائدة للسياسات الاجتماعية"). تواجه البرازيل حالياً تحدّياً اقتصادياً و سياسياً رئيسياً في معالجة هذه المسائل بسبب الصعوبة التي تكتنف عملية إصلاح السياسات وخصوصا بسبب الدور الضمني الذي تلعبه المصالح المكتسبة. بالإضافة إلى ذلك، يجب على البرازيل أن تعالج هذه المسائل والتصدّي لضغوطات المالية العامة المتنامية التي من الممكن أن تضعف الشرعية السياسية الحالية للسياسات الاجتماعية وتقلل من شأن النجاحات المهمّة التي حققتها البرازيل مؤخراً على هذا الصعيد.
El modelo brasileiro de seguridade social: as propostas pendentes
Para alcançar as metas nacionais definidas pela Constituição Federal do Brasil de 1988, as prestações pecuniárias apenas são insuficientes, na ausência de serviços sociais mais sólidos para reduzir as desigualdades e incrementar a coesão social. A Constituição, malgrado a importância nacional e a implicação internacional, não abordou muitos dos pontos fracos institucionais e administrativos no modelo do sistema nacional de seguridade social. Embora a cobertura tenha aumentado e a desigualdade tenha sido reduzida, tais medidas não são suficientes. As ambições do Brasil de avançar as políticas sociais (e, de fato, corresponder ao seu status internacional de que usufrui como “líder em política social”) podem sofrer limitações em razão da dependência excessiva das transferências de fundos condicionadas, como as que são dotadas para o programa Bolsa Família. O Brasil está diante de um importante desafio econômico-político, o de enfrentar todas essas questões, porque o processo de reforma política é difícil e, o que é ainda mais importante, em razão do papel vinculado dos interesses em jogo. Ademais, o Brasil deve tentar resolver essas questões diante de crescentes pressões fiscais, que poderiam enfraquecer a legitimidade política atual dos programas sociais e solapar os importantes sucessos recentes.
With respect to pension provisions, Brazil has been an outlier in Latin America. While most of the larger countries in the region implemented paradigmatic reforms that featured defined contribution (DC) individual accounts to either substitute or complement public pay-as-you-go (PAYG) systems, Brazil instituted parametric reforms that emphasized fiscal transfers for lower-income workers, combining social insurance with social assistance pensions, as determined by the 1988 Federal Constitution (Chamber of Deputies, 2010).
The contrast in policy across the region became apparent in the 1980s and 1990s when high administrative and transaction costs associated with individual accounts led to projections of lower income replacement rates. Despite the movement towards DC pension systems, the hoped for positive impacts of rapid GDP growth, higher levels of investment and job formalization did not meet the earlier upbeat projections in countries such as Argentina, Chile or Mexico (Kritzer, Kay and Sinha, 2011).
Under these circumstances, Brazil's social security system became something of a showcase, since fiscal transfers were generous in providing lifetime benefits. The combination in Brazil of a conventional defined benefit pension system and social assistance cash transfer benefits for the most vulnerable groups, along with high rates of coverage, received widespread popular support.
Nevertheless, the Brazilian model for old-age pensions and disability insurance, although successful in many respects, suffers from high levels of inefficiencies and requires careful critical analysis. In 2011, public expenditures on social policies amounted to 23 per cent of GDP, and pension benefits reached 14.5 per cent of GDP. The political debate over reform continues, but often lacks detailed analysis with respect to the distributional consequences and the extent to which the system favours the better off. Low levels of accountability and transparency with respect to public policy jeopardize what may be, in fact, an interim policy success story. That is to say, in the wake of challenges heightened after the 2007–08 global financial crisis and the rise of emerging economies that may not fully respect basic social rights as defined by International Labour Organization (ILO) Conventions and Recommendations on social security, there are new challenges.
This article focuses on pension policy within the context of Brazil's social security system, and the need for a new reform agenda that emphasizes social rights that are consistent with the objective of sustained economic growth. Reforms that would ensure equitable and efficient pensions need not depend upon constitutional amendments, but can also be achieved through less onerous means such as ordinary legislation and administrative and institutional reform measures. The article is structured as follows: after a discussion of pension coverage, the performance and effectiveness of Brazil's social security pension provisions are then examined. We then examine Brazilian social security reform and conclude with policy proposals.
Pensions in Brazil: The success of universal coverage
Retirement coverage in Brazil can be traced back to the colonial period, when in 1554 the Santa Casa de Misericórdia in Santos first provided pensions. In 1919, after the formation of the International Labour Organization, workplace injury insurance was established and guided by new international recommendations. Social security legislation since 1919 reflects increasing state intervention, centralization, and universalization of coverage for workers in more precarious occupations – all of which was foreseen as early as 1945, but not enacted upon until the early 1970s. Initial regulations were tightly linked to Brazil's labour code, but over time they have taken on their own juridical identity (Matijascic, 2002; Matijascic and Kay, 2008).
Current Brazilian social policy is based upon the Social Order section of the Constitution enacted in 1988 and regulated by Laws 8.212 and 8.213 from 1991 (workers in the private sector) and Law 8.112 from 1990 (federal civil servants) concerning pensions. The 1988 Constitution's principal innovation was the creation of the concept of “social security”, a term intended to unify health, social assistance, unemployment and pension policies. Furthermore, it created the social security budget – which segregated funds from other public budget lines – with the intention that social security budget resources could only fund social security. The most significant changes with respect to coverage in the 1988 Constitution were:
Utilizing the Brazilian minimum wage as the benefit floor;
Establishing equivalent rules for urban and rural benefits; and
Equalizing benefits for men and women.
The 1988 Constitution was, according to Werneck Vianna (1998), a reaction to the institutional model imposed after the 1964 military coup that deposed the President-elect, João Goulart. The range of reforms introduced between 1964 and 1968 sought to modernize the Brazilian State through public policies similar to those found in the leading Western economies. During that period, the military government instituted tax reform, created the Central Bank, and reformed social policies. The reforms sought to concentrate personal incomes in the upper income deciles to support stronger consumer spending, which was expected to lead to sustained economic growth, and thus higher per capita income that would progressively incorporate all of the population. Consumption was to be the driving force behind the Brazilian economy.
Social institutions were organized in order to provide protection for lower-income workers, with those seeking higher quality services expected to pay for private provision. This was the essence of the so-called conservative modernization of social policy. Since the cost of private provision was unaffordable for more than two-thirds of the labour force, the focus of the authors of the 1988 Constitution was to consolidate a system of universal protection through pensions, social assistance, health, and education with progressive improvement in quality (Werneck Vianna, 1998). It is important to note that this situation still persists, with only a minority of the Brazilian population obtaining health care, education, and pensions via market provisions.
The most salient result after the introduction of the new constitutional provisions was the expansion in the number of old-age benefits paid to rural workers and senior citizens living in households with income below a quarter of the minimum wage, an unofficial poverty line. For rural workers, the retirement age was 65, but after the Constitution of 1988 it was lowered to age 60 for men and age 55 for women. For the elderly in poverty, the age to receive benefits was lowered from age 70 to age 67 in 1996, and to age 65 in 2003 for men and women.
According to the 1988 Constitution, constitutional amendments1 require two voting rounds in the Chamber of Deputies and two voting rounds in the Senate, both of which require a minimum of 60 per cent of the votes cast. Any amendment introduced in the Senate requires two additional rounds of votes in the Chamber of Deputies.
Debate over pension reform has been centred upon parametric measures that focus on benefit rules and new forms of retrenchment. In 1998, a major reform included:
Substituting length of service for length of the contribution period, with the introduction of the fator previdenciário (retirement factor), with rules similar to those of notional accounts;
Fixing the minimum retirement age for new entrants to the public sector at age 60 for men and age 55 for women (ages 53 and 48, respectively, for those already working); and,
New rules for disclosure, accountability, vesting, and portability of pension funds to stimulate defined contribution funds.
In 2003, a new pension system was established that would institute defined contribution plans for civil servants earning over the equivalent of USD 2,200 per month. Despite the approval of the constitutional amendment, the proposal received a great deal of opposition from the parties of the Left, and the law applies to those who have entered the civil service after February 2013. The 2003 reform introduced other relevant measures to link the value of the benefit to contributions, including the requirement that workers pay contributions on benefits above the private-sector ceiling, which remains in effect despite judicial challenge (President Cardoso had been unsuccessful in his attempt to impose such a measure).
The 1998 and 2003 reforms were not trivial, raising requirements for accessing benefits and reducing replacement rates for workers in the public sector who have contributed for fewer years or who retire earlier than workers in the private sector. It is important to stress that the legislature refused to delink minimum benefits from the minimum wage or to remove the rules for National Institute of Social Security (Instituto Nacional do Seguro Social – INSS) benefits from the 1988 Constitution. The impact on coverage can be seen in Figure 1.
Figure 1 presents the most fundamental elements of social security in Brazil: the 1988 Constitution and the consolidation of democracy have resulted in improved access to benefits, but the number of contributing workers with guarantees of social rights is still low, having risen from 33 per cent in 1978 to 41 per cent in 2012. The number of workers without coverage has dropped from 48.2 per cent in 1978 to 31.2 per cent in 2012. In short, the 1988 Constitution guaranteed the rights of the more vulnerable sectors of the population and removed the stratification of the conservative modernization. As a result, there was a notable rise in coverage, but without a concomitant expansion of contributors: a mismatch that has had a significant impact on debates over reform.
The 1988 Constitution and its impact on the distribution of benefits and public policy performance
After being severely criticized for its levels of social inequality in the 1990s, as described by Matijascic (2002), Brazil has become a showcase for developing countries given its extension of coverage, including non-contributory benefits. The distribution of benefits illustrates the importance of providing old-age benefits according to the capacity to contribute (MPS, 2013; data from September 2013 that includes private-sector workers and workers in small-sized cities):
18.5 per cent were contributory benefits, with a contribution period of 30 to 35 years of contributions;
26.9 per cent are disability benefits;
9.5 per cent are benefits for attaining age 65 for male urban workers and age 60 for female urban workers, with a contribution period of 15 years;
15.6 per cent are for rural workers who have reached age 60 (men) and age 55 (women);
29.4 per cent of benefits are granted to workers older than age 65 or disabled persons with a family per capita income below one quarter of the national minimum wage.
Only 18.5 per cent of benefits are paid to persons that have contributed regularly. Those with partial contributions, that is to say, for persons receiving disability benefits or for urban workers granted old-age pensions, represent 36.4 per cent of benefits in payment. Those that contribute little or nothing, including rural workers and the poor, receive 45 per cent of benefits. To sum up, 81.5 per cent of Brazilians who receive a pension do not contribute regularly via payroll taxes. In this context of fragile labour markets and very high levels of precariousness, it is evident that tax transfers are essential to achieve universal coverage. The social security budget is responsible for this function.
Between 1978 and 2012 there was a substantial increase in the number of families with a member receiving a pension. In 1978, 8.3 per cent of families in the lowest income decile had a household member receiving a retirement pension compared to 60.1 per cent in 2012, while for the median and upper income deciles, the figures rose from 20.9 per cent and 20.7 per cent to 46.6 per cent and 46.8 per cent, respectively. The most substantial increase occurred among lower-income families. Improved access to non-contributory pensions, along with outreach efforts, were essential for increasing the number of families receiving old-age benefits. In order to measure the impact of old-age benefits on family income, Table 1 quantifies the role played in family income by social security cash benefits.
Table 1. Social security transfers as a percentage of Brazilian family income according to family income deciles
Source: PNAD microdata for selected years.
The increase in the number of people across all income deciles living in households who receive old-age benefits is matched by a considerable increase in retirement and social assistance pensions. The data presented in Table 1 shows the change after the introduction of the 1988 Constitution. One can observe that in 1978, there was no progressiveness; that is to say, the lower-income deciles did not receive benefits proportionally higher when compared to 1988. Although there was some backtracking in 1998, there was a strong surge in redistribution in 2008. The increase in access to old-age benefits that are not earnings-related has been fundamental for raising redistribution via social security.
The reduction in poverty after the payment of benefits is more robust and effective in 2010, after the consolidation of the 1988 Constitution, which set a floor for the minimum wage, and facilitated access to benefits for rural workers, the aged, and families with incomes below a quarter of the minimum wage. According to Tafner (2006), families with elderly persons have income levels 43 per cent above those that do not.
In short, the 1988 Constitution expanded benefit coverage for those with few regular contributions, linking benefits to the minimum wage floor that came fully into effect in 1991, and increased family income associated with old-age benefits as well as increasing the number of families covered. The improved financial status and reduction in poverty among the oldest cohorts and the poor is striking, as intended by the design of these policies.
Effectiveness of public policy: Controversy over inequality
Payment of old-age benefits raised income protection coverage for the oldest cohorts to patterns comparable to countries of Western Europe. Beyond that, it also generated an impact on household income, serving to reduce poverty (Matijascic and Kay, 2013).
One of the principal criticisms of public policy has been the level at which the minimum wage is set, and to which pension benefits are linked. As noted by Tafner (2006), a point of contention is that pension income for poor households serves as an important source of family income, and therefore is a disincentive to accepting low-wage employment, thereby distorting the labour market and making the country less competitive. The Bolsa Família programme has received similar criticism, albeit that Giambiagi and Tafner (2007, p. 167) have argued that Bolsa Família is less distortionary than the minimum wage.
However, that argument does not take into consideration studies, such as those summarized in Figures 2 and 3, which utilize the poverty line in Brazil. A later study by Giambiagi and Tafner (2010) also does not mention empirical studies of inequality.2 This approach differs from other studies, which focused on the impact of benefit payments (IPEA, 2010, p. 10).
With the new legal framework after the 1988 Constitution, the payment of old-age benefits reduced vulnerability in the older population with respect to access to income. Figures 2 and 3 show the results for families in extreme poverty, i.e. income below one quarter of the minimum wage per capita, before and after the payment of old-age benefits in 1988 (before the 1988 Constitution had an impact) and in 2012.
Inequality is also an ongoing concern in Brazil. The 30-year movement of the impact of old-age pensions on reducing inequality can be seen using the Gini coefficient, which was 2.6 per cent in 1978, 3.3 per cent in 1988, 5.5 per cent in 1998, and 10.3 per cent in 2008. This is noteworthy given that improvements were limited prior to the 1988 Constitution, with progress accelerating after its implementation. Significantly, inequality decreased more substantially between 1998 and 2008, especially after the purchasing power of the minimum wage increased in 2004.
As Matijascic and Kay (2010) argue, the contribution of Esping-Andersen and Myles (2008), based on Marical et al. (2006), provides a useful analytic framework. Initiatives like IPEA (2010) are not sufficient to provide a comparative international benchmark. Using the Gini coefficient, or reductions in inequalities, provides a clearer indicator when comparisons are made with other countries.
Esping-Andersen and Myles (2008) demonstrate that the welfare states of Western Europe and North America achieved reductions in inequalities that are more or less pronounced regarding the level of decommodification of each country with respect to social protection. In other words, the less the redistribution of income or access to social services depends on the market, the greater the potential reduction in inequality. Furthermore, based on Marical et al. (2006), the authors find that greater potential for reducing inequalities comes from the provision of social services, which accounts for why the Nordic countries achieve greater equality.
In applying the criteria of Marical et al. to Brazil, the reduction in inequality after income transfers or social services3 was higher than that of the Nordic countries. Thus, social policy in Brazil was positive in terms of the reduction in inequalities through social services or income transfers, respectively, as measured by the Gini coefficient. Brazil's results were better than other welfare state-typologies according to Esping-Andersen and Myles (2008),4 in contrast to Giambiagi and Tafner (2007, 2010). The comparative results for reductions in inequalities via the provision of social services and income transfers were:
Universal or Social Democrat typology: 37 per cent (social services) and 16 per cent (income transfers)
Liberal typology: 24 per cent (social services) and 4 per cent (income transfers);
Conservative typology: 24 per cent (social services) and 3 per cent (income transfers), and,
Brazil (data for 2008): 41.5 per cent (social services) and 18.5 per cent (income transfers).
Figure 4 demonstrates the impact of different social policies on inequality following Marical et al. (2006) and shows that health and education have an especially powerful impact on income distribution and reductions in inequality. It is important to note that income transfers have statistical composition effects and aggregate data cannot simply be composed from its constituent parts.
Matijascic and Kay's (2010) results presented in Figure 4 have limits. None of the countries that are compared to Brazil has similar per capita income levels or labour market structures with such ineffective regulation and such large informal sectors. It would be useful to compare and benchmark Brazil to neighbouring countries with similar per capita income levels. Brazil's high scores presented in Figure 4 are potentially misleading. Given that Brazilian income distribution is so regressive, not least because labour market regulation and tax policy does little to reduce income inequality, social policies can therefore potentially achieve major outcomes regardless of their efficacy and efficiency.
These results have important implications for public policy. The finding that social services, especially health care, have a bigger impact on reducing inequality is significant. That is to say, in a country with such high inequality, there are limits to what can be achieved by income transfers via old-age pensions, whereas expending additional resources, on education and health care especially, can be more effective in reducing inequality.
The consolidation of the social protection model focusing on offering support to families, but in a context where private institutions cater to those with higher incomes, is still dominant in Brazilian social policy. While the 1988 Constitution represents progress toward the goal of universal coverage and the Social Democratic model, the persistent deficiencies in the quality of public social services means that Brazil cannot be labelled a “welfare state”.
A possible counter-argument is that the lower levels of per capita income in Brazil and limited infrastructure are obstacles to forming a welfare state. The level of quality of health and education services, especially compared to other countries with similar per capita income levels as Brazil, is poor.
The quality of services provided through the social protection system presents challenges, especially with respect to health care and education.5 Matijascic and Kay (2013) showed that coverage for health care was very unequal, especially in the poorest regions or poorest metropolitan neighbourhoods. Inadequate access to coverage and assistance for health care reduces healthy life expectancy in Brazil, as defined by the World Health Organization (WHO, 2009), and results are poor when compared to neighbouring or the other BRICS countries (Russian Federation, India, People's Republic of China and South Africa). Although Brazilian healthy life expectancy is higher than in countries such as Bolivia, India, South Africa or the Russian Federation according to 2007 data, the total years of non-healthy life is higher in Brazil. This data suggests that disease prevention is insufficient, with the Unified Health System (Sistema Único de Saude – SUS) prioritizing curative treatment before prevention.
Moreover, Brazil's history of privileging cash transfers over social services is clearly at the crossroads. Public opinion polls show that dissatisfaction with health care and education is the most significant. According to the WHO (Estadão, 2013), public spending on health care is below the world average. This problem is particularly acute given that more than 70 per cent of the population cannot afford to pay for private health care. This lack of investment, along with ineffective budget expenditures, reflects years of lost healthy life that contributes to higher spending on curative medicine (which leads to higher costs than would be the case with investment made in effective preventive care). Last, but not least, the loss of healthy life may accelerate pension system expenditures and lower labour productivity. These social policy challenges have a direct impact on economic performance.
The distribution of funds between cash transfers to families and social services is overly biased towards cash benefits, which is problematic if the aim is to reduce inequalities and promote social cohesion. If we consider pension benefits, conditional cash transfers and unemployment insurance, Brazilian spending is equivalent to 17.1 per cent of GDP. Health care expenditures are equivalent to 3.6 per cent of GDP, much lower than the 4.1 per cent of GDP spent on private health plans. An additional 5.9 per cent of GDP was spent on education (2011). The share destined for social services such as education and health care is low given that more than 70 per cent of families depend on public health care exclusively, and about 86 per cent of families depend upon public education.
Conversely, spending on survivor benefits is equivalent to 3.2 per cent of GDP, a share that is much higher than in other countries (James, 2009). According to data from the ILO6 and OECD,7 total spending on survivor pensions in Chile is about 1.9 per cent of GDP, compared to 1.3 per cent in Argentina and 0.3 per cent in Mexico. In OECD countries, the average is about 0.9 per cent of GDP (with Italy spending 2.5 per cent of GDP and Belgium spending 2.1 per cent of GDP). Given that the importance of the male breadwinner model has declined, expenses for survivor benefits should decrease and in a few decades become marginal. If, hypothetically, Brazil's survivor benefit expenditures as a percentage of GDP were equivalent to those of Italy or Belgium, countries with a male breadwinner tradition and demographically-older populations, it could raise expenditures by 31 per cent (63 per cent when compared to the OECD median). In short, the fiscal impact of survivor benefit policies should not be underestimated.
It is important that the singular features of Brazil's social security system are fully considered – to do otherwise leads to analytical errors. Giambiagi and Tafner (2010) and Schwarzer and Santana (2013) present analyses based on labour markets characterized by stable salaried employment. This is not, nor has it ever been, the case in Brazil, as described above, where over 80 per cent of benefits paid to workers in the private sector were unfinanced or partially financed by contributions. These studies utilize examples that compare the collection of contributions as a percentage of salaries and benefit expenditures. This is problematic given that little over one third of Brazil's economically active population contributes while 95 per cent of the population aged 70 or older has access to benefits. To exclusively tax workers who make contributions to social security would result in exorbitant contributions or would result in restricted access or severe benefit cuts.
Article 202 of the Brazilian Constitution obliges the government to provide funds to pay social security expenditures and, if needed, introduce specific taxes to finance social security or to raise tax rates. Giambiagi and Tafner (2007, 2010) and Schwarzer and Santana (2013), among other experts, do not take this into consideration, and instead present calculations based on INSS or public service pension schemes' collections and expenditures. This omits value added taxes and taxes on profits, both of which are required by the Constitution to fund the social security budget. As a result, calculations based on constitutional mandates result in major surpluses rather than deficits. Omitting these other sources of revenue can be misleading, since policies that may be cost effective can still show deficits depending on how they are accounted for in the budget (Matijascic and Kay, 2008).
How Brazil's demographic profile is accounted for is also important. Giambiagi and Tafner (2010) and Schwarzer and Santana (2013) warn that Brazil is ageing rapidly and without constitutional reforms, the system will become unviable given the tight relationship between contributors and beneficiaries. However, in a society with disproportionately few contributors yet nearly universal coverage, this will require, as we have seen, the transfer of fiscal resources. Furthermore, the ratio of contributors to beneficiaries can be modified by:
Reducing informality in the labour force, which will increase social security contributions as has been the case in Brazil over the past decade;
Improving health conditions and epidemiological monitoring to reduce disability benefit rolls as well as health care expenditures.
In considering dependency ratios, age is only one consideration, and developing effective policy requires taking a broader perspective on the labour market.8
Living conditions remain precarious, given that the inequality of income in Brazil is still amongst the highest in the world and more than 46 per cent of the population live in houses with poor sanitation (IBGE, 2012). In other words, cash transfer policies are important, but they have limits, since their main objective is to transfer income between generations and not social groups. Social services, economic, social and environmental regulation, as well as taxation and public policy expenditure, have a major role to play, indeed a more prominent one if the goal is to reduce inequality and promote higher levels of development.
The neglected role of institutional and administrative reforms
Since the passage of the 1988 Constitution, there has been an ongoing debate over the need for market-friendly reforms that promote private investment and financial stability by reducing the tax burden. Giambiagi and Tafner (2010) advocate constitutional reform to achieve these objectives, arguing that rural and public assistance pensions have perverse effects, more so even than the distortions created by the pension schemes for public-sector employees.
The political Left in Brazil defends these benefits with ideological zeal, considering these legal protections to be untouchable. Fagnani (2011) and Cohn (2010) emphasize the reduction in inequalities, and defend without distinction the universal benefit floor and statutes that protect the aged and rural retired workers, and other less equity-enhancing rules concerning survivor benefits, the accumulation of multiple benefits, and rules allowing workers to receive benefits while still being employed.
None of these arguments advocate adopting major parametric reforms, most of which do not require constitutional amendments. Parametric reforms can be targeted at benefit plans and management practices, not to mention the adoption of benchmarking according to international standards as discussed by Gillion et al. (2000), or, most recently for minimum social security guarantees, by the International Labour Organization's Recommendation concerning national floors of social protection, 2012 (No. 202).
Most financially-effective reforms would not exclusively depend upon constitutional reform. It is important to note that reforming ordinary legislation or changing administrative practices can have fast, powerful, and long-lasting effects to reduce income inequality and promote social cohesion.9
One difficulty is the current nature of the debate in Brazil. Stakeholders commonly adhere to their own epistemic universe, ignoring other arguments. A more sober perspective suggests that the Constitution, on the one hand, privileges universalism and, on the other hand, protects corporative interests, especially among State employees and philanthropic institutions. Until February 2013, State employees were not subject to the same pension ceiling as private-sector workers. Philanthropic institutions are immune to taxation, including payroll taxes as employers, and are not monitored on a regular basis to evaluate their performance and accountability. Put bluntly, the Constitution did not create mechanisms to prevent the transfer of income to the better-off and sustains mechanisms that stimulate income concentration and a regressive tax model.
The possibility of accumulating benefits and pensions or old-age benefits along with earned income was described in detail by Matijascic and Kay (2008, 2013). We argue that there is no justification for:
Paying full and not partial benefits if a person works full time. Starting with Bismarck's legislation, pensions have been meant to replace income after one can no longer work;
Providing lifetime survivor benefits to those not in a position of economic dependency. Survivor benefits are intended for spouses that do not enter the labour market, school-age orphans, or the disabled;
Accumulating multiple retirement pensions and survivor pensions while also receiving income from full-time work. Such a situation is not a social right and diverts public resources from more productive uses;
Allowing premature retirement, which is not a social right and unduly burdens public finances.
According to Figure 5, over 25 per cent of pensioners work in full-time jobs, as is the case for more than 20 per cent of those older than age 65, accumulating both sources of income. From 1978 until 2012 the situation has not substantially changed. Nevertheless, among those aged 65 or older, there is a smaller contingent of pensioners that remain in the labour market in 2012. This situation probably means that increments of income for lower deciles stimulate the elderly to effectively retire when they are entitled to pensions.
If a person works full time, it does not make sense for them to receive a full retirement pension benefit. It is essential to emphasize that in Brazil there is no part-time regular old-age pension to facilitate a gradual exit from the labour force.
Most Brazilian social policy experts do not criticize the possibility of accumulating benefits while working full-time. On the contrary, what is commonly argued is that benefit levels are too low and that returning to work is necessary (Follador, 2013). However, the average old-age retirement pension benefit amounted to 64.4 per cent of average earnings for workers aged 16–64 contributing to social security in 2012, while for those that work and receive retirement pension benefits and survivor pension benefits it represented 96.3 per cent of average earnings. Moreover, according to the Brazilian National Household Sample Survey (PNAD, 1978–2012) data in 2012, for those who earned retirement pension benefits and received a regular salary, the figures were equivalent to a 106.2 per cent replacement rate and for those accumulating retirement pension, survivor pension benefits, and regular income from work the figure was equivalent to 180.8 per cent. It is clear that those who are better off can accumulate different sources of income, which concentrates income distribution.
The persistence of these rules, which have existed since the 1930s, is a clear demonstration that legislation needs to be understood in light of prevailing patterns and practices. It is precisely those receiving higher benefits, with higher replacement rates, that return to the labour force. In contrast, there are severe restrictions in accessing benefits for those workers with a more precarious link to the labour market. These restrictions include:
The requirement to contribute for 15 years and reach age 65 for men and age 60 for women to be eligible to receive a retirement pension in an urban area. These are strict requirements when compared to other countries (Table 2);
Individuals who become unemployed and cease contributing lose their insurance and give up their right to benefits in the event of disability or death (survivors), even if they have contributed for a lengthy period;
Access to non-contributory benefits is contingent on family per capita earnings up to one quarter of the minimum wage, which necessitates means-testing.
Matijascic and Kay (2010) show that if the poverty line were measured according to OECD methods (i.e. half of workers' median earnings), poverty in Brazil would be reduced by only 2 per cent, and not more than 30 per cent. Thus, the arbitrary value given to the poverty line is a determinant of the success or failure of policy.
Table 2. Entitlement requirements and minimum age for old-age pensions eligibility
Minimum contributions (does not include social assistance benefits)
In presenting the options for those with access to stable employment along with the options for those with fewer opportunities, there is little doubt: the financial impact of legislation concerning benefits in Brazil favours the better off, who received higher earnings from more stable occupations. Table 2 lists the legislation of a range of countries. The minimum requirements for receiving a pension benefit in Brazil is among the strictest, because it requires a contribution period of 15 years, while in other countries it may only require affiliation or residency or shorter periods of contribution.
Administrative comparisons are difficult without benchmarking, nevertheless, it can be observed that:
The large number of social security institutions is inefficient. Without monitoring the health status and capacity to work of beneficiaries, the risk of fraud increases (incidents of fraud are regularly reported in the media);
There are a large number of personnel devoted to administrative work and comparatively few working in areas that directly serve the public. This helps to explain negative views of public services; and,
Much needs to be done with respect to organization and methods. Services are slow, public works are continuously delayed, and State purchasing is slow due to very ineffective legislation.
Reform of the State is an urgent priority, and reforming social security will be a key component. For example, Brazil spends more than 3 per cent of total benefits spending on administrative costs, while the United States and Switzerland spend 0.7 per cent and 0.4 per cent, respectively (Matijascic and Kay, 2008).
Institutional or managerial changes can free up resources for other priorities, including raising benefit levels or improving public social service infrastructure in the interior of Brazil, where health services are more limited and the quality of service is often precarious. This translates into fewer years of healthy living and disproportionately high levels of ill health, as described by Matijascic and Kay (2008). Clearly, a new development strategy is needed to reduce income concentration and create job opportunities for the population. Social services can obviously be used as leverage to improve social development. Moreover, relevant pension reforms do not require constitutional amendments and can be achieved by means of ordinary legislation or administrative decisions.
Policy proposals: Towards a pension model adapted to the labour market
Brazil's pension system plainly has room for improvement. To base a pension system on payroll taxes in a low-wage economy with a low density of contributions and high levels of informal employment (with high levels of self-employment and domestic service work) is problematic. In turn, permitting the accumulation of benefits for those who remain employed serves to concentrate income and limits capacity to adopt policies to promote employment, especially youth employment. Permissive survivorship pensions represent a further misallocation of resources.
The pension system could achieve effective rates of coverage and improved equity with a first-pillar universal benefit, a second-pillar contributory benefit based on earnings, and a third voluntary pillar regulated by the State. As regards the first pillar, this would be a single non-contributory universal benefit for those who were no longer able to work, offering a means to achieve universal coverage with low administrative costs. Coverage would be made available to all who satisfied the age condition or who were assessed as disabled and incapable of work, so that if a person ceased contributing because they could no longer work, they would still have access to benefits, as is the case in many other countries. A contributory second-pillar benefit would permit income smoothing for those with higher earnings.
For those who do contribute regularly there could be easier access to credit from financial institutions, linked to total contributions. Firms, as well as individuals, would have access, and this would provide incentives for firms to follow best governance and disclosure practices. Pensioners may get easier and significantly cheaper access to credit if the periodic repayments for loans are deducted directly from their benefit payments. Additional credit measures could be considered for those who regularly contribute, as a way of recognizing that their income risk is far lower than for those who are not contributing regularly.
A simpler tax code would end the confusing and efficient practice of multiple taxes; Federal agencies and social security taxes frequently overlap. A fewer number of taxes and simpler rules would lead to greater efficiency. The diverse range of administrative authorities and tax rolls is inefficient and raises costs. A single registry would lead to administrative efficiencies and help prevent the current high levels of fraud. The monitoring and cross-checking of administrative records could free up a significant amount of resources.
The absence of a reserve fund makes the retirement system vulnerable to short-term financial conditions. A reserve fund is called for in the Constitution and should be implemented – it can also help smooth the demographic transition as the population ages. The decentralized units of government which manage benefits can also be used as a tool to audit and detect potential fraud. There should be incentives for local administrative units to benefit from recovered contributions or the cessation of benefits due to irregularities. Professional organizations, trade unions, and civil society movements may also support this effort by participating in campaigns to promote the importance of paying contributions.
The accumulation of benefits and salaries should be considered from a legal perspective. Benefits were designed to substitute for wages when a worker was no longer able to earn income. The possibility of accumulating salaries and benefits impedes policies to encourage or postpone retirement with the aim of improving labour market conditions.
Pension policies affect both social rights and public finance, and remain an intensive source of debate. Policy legacies from decades past continue to elevate costs while the failure to adhere to principles enumerated in the Constitution means that measures to reduce inequality and promote social cohesion are not undertaken.
Conclusions: What has changed and what has not?
The principal challenges facing Brazil's pension system stems from the fact that it combines social insurance with social assistance where coverage is nearly universal for the aged, but contributions are made by a limited segment of the working-age population. Under such circumstances, it is impossible to achieve financial equilibrium between contributions and benefits. The social security budget, which utilizes fiscal resources to fund benefits, is somewhat regressive, but addresses these conditions. Yet current rules permit the accumulation of multiple benefits and the granting of survivorship benefits that are extremely generous. These represent significant policy challenges that must be included as part of any administrative and institutional reforms.
Brazil's recent social policy achievements stem from the 1988 constitutional reform, which was the result of a broadly participatory process that impeded the adoption of structural reforms that might have left the elderly population more vulnerable. Yet, the reluctance to reduce generous contributory benefit entitlements that are a legacy from the past is an obstacle to more equity enhancing reforms.
Furthermore, the much discussed recent reduction in income concentration via cash transfer benefits and tax incentives cannot stimulate further progress in income distribution without more investment in economic and social infrastructure, and, perhaps even more so, better social services to assist a rapidly ageing population with relatively low education levels and poor health conditions.
In other words, to defend the status quo without acknowledging the need to improve both efficiency and equity is disingenuous, while it is also a mistake to adopt models from other countries that do not take into account the specific dimensions of Brazil's labour market (with its significant informal sector) as well as other economic, social, and demographic factors. As a precondition for reform, the range of distributional consequences of the current system, which contains both progressive and highly regressive features, must be systematically described and fully considered.
Social policies that are better adapted to social conditions – and the ending of pension entitlements that have nothing to do with social rights and which may contribute to the deterioration of social cohesion – are essential. Such reforms are necessary to modernize the State and improve labour relations in order to consolidate a new equity-improving development path.
Reform of the pensionable age requires a constitutional amendment. See footnote 9 for an explanation of which types of social security reform require constitutional amendments.
Giambiagi and Tafner (2007, 2010) also argue that the burden of Brazilian civil servants' pension schemes is not relevant to measure the impact on public finances. Nevertheless, data for 2004 presented by Palacios and Whitehouse (2006) show that Brazil spent about 4.8 per cent of the GDP on civil servants' pensions and was ranked first when compared to non-OECD countries, where the average was 1.3 per cent. Tunisia, Zimbabwe, and Turkey – the countries ranked second through fourth – spent about 2.5 per cent of GDP. In OECD countries, the average spent on civil service pensions was about 1.9 per cent, with the top two countries, Austria and Belgium, with expenditures of about 3.7 per cent of GDP. The new pension scheme for Brazilian public servants put in place in February 2013, creating pension funds for workers with incomes above the national ceiling, will take decades before they contribute to a reduction in the fiscal burden.
The social services described here included health and education, which were the only sectors analyzed across the member States of the Organisation for Economic Co-operation and Development (OECD). Not using policies targeted at specific segments of the population hides the impact of the improvement in well-being in societies that use these types of services. It is important to note that France, and above all the Nordic countries, have a wide range of social services recognized as being highly effective.
Esping-Andersen (1990) analyzed welfare states from Western Europe, Oceania, and North America. After calculating the degree of decommodification related to labour market dependency, Esping-Andersen concluded that the results presented three different models: Social Democratic, Conservative and Liberal. The Social Democratic model was characterized by a low level of dependence from the labour market and an extensive menu of non-contributory universal social services. The Conservative model relied on income protection and was targeted to workers belonging to dominant social professional groups, with a minor role for social services. Finally, the Liberal model relied basically on market provision and social protection is focused on those who cannot find regular jobs.
Quality is also a concern with educational policies, as Brazil's scores in the Program for International Student Assessment (PISA) are very poor, even when compared to countries with lower per capita income. Only 3.8 per cent of Brazilian students achieved a score 4 or higher in mathematics, or, conversely, more than 95 per cent of students aged 15 did not meet expected levels of performance. Performance in language and the sciences are also poor. Among South American countries, only Colombia and Peru received lower scores in mathematics, although Brazil's financial resources and state capacities are higher than those of most of its neighbouring countries (OECD, 2010).
If we analyze the employment-adjusted dependency ratio – which incorporates those leaving work temporarily due to illness, disability or unemployment, or being economically inactive – the proportion of dependant workers expands dramatically. For example, in comparing Brazil with two industrialized countries for which data is available, we get a sense of how dependency expands. In France and Spain, the population aged 65 or older compared to the population aged 15 to 64 in 1990 and 2003 was: 21.3 per cent and 24.8 per cent in France, and 20.7 per cent and 24.8 per cent in Spain (Whiteford, 2005). However if the employment-adjusted dependency ratio is used, the results for those same years were 102.4 per cent and 101.6 per cent in France, and 133 per cent and 105.6 per cent in Spain. For Brazil, using the same methodology for the years 1990, 2003, and 2012, the results would be 108.9 per cent, 102.4 per cent, and 84.3 per cent. This demonstrates how considering employment conditions is vital for developing effective public policies.
Constitutional reforms involving pension benefits are required for legislation on minimum and maximum benefits, minimum age, years of contribution and different types of benefit (such as old-age, survivor, sickness, maternity, unemployment, etc.). Specific eligibility criteria such as accumulating wages and benefits or qualifying for survivor benefits only require ordinary legislative reforms. Furthermore, rules about disability are legislated by the Federal Administration using decrees within the INSS. This is not to say that the issue does not involve very sensitive political problems. In 1997, President Cardoso issued a Decree that prohibited receiving pensions while being employed. At the time as the constitutional reform was under consideration in the Senate, President Cardoso came under intense pressure to revoke the Executive Decree (Matijascic, 2003), which he did in order to restart negotiations (Amendments 19 and 20, were approved in December 1998). This demonstrates that even a measure that did not require a constitutional reform – a rule prohibiting collecting a pension while being employed – can face intense opposition from vested interests (and how the judiciary would have responded will never be known).