Report No: AUS0001674 Enhancing Coverage and Cost-effectiveness of Brazil’s Unemployment Protection System: Insights from International Experience Social Protection and Jobs Global Practice, Latin America and Caribbean Region December 2020 © 2020 International Bank for Reconstruction and Development / The World Bank 1818 H Street NW, Washington DC 20433 Telephone: 202-473-1000 Internet: www.worldbank.org This work is a product of the staff of The World Bank with external contribution. The findings, interpretations, and conclusions expressed in this work do not necessarily reflect the views of The World Bank, its Board of Executive Directors, or the governments they represent. The World Bank does not guarantee the accuracy, completeness or currency of the data included in this work and does not assume responsibility for any errors, omissions or discrepancies in the information, or liability with respect to the use of or failure to use the information, methods, processes, or conclusions set forth. The boundaries, colors, denominations, and other information shown on any map in this work do not imply any judgment on the part of The World Bank concerning the legal status of any territory or the endorsement or acceptance of such boundaries. Nothing herein shall constitute or be construed or considered to be a limitation upon or waiver of the privileges and immunities of The World Bank. All of which are specifically reserved. Rights and Permissions The material in this work is subject to copyright. Because The World Bank encourages dissemination of its knowledge, this work may be reproduced, in whole or in part, for noncommercial purposes as long as full attribution to this work is given. Any queries on rights and licenses, including subsidiary rights, should be addressed to World Bank Publications, The World Bank Group, 1818 H Street NW, Washington, DC 20433, USA; fax: 202-522-2625; e-mail: pubrights@worldbank.org. ACKNOWLEDGMENTS This note was prepared by a team led by Matteo Morgandi (task team leader) and comprising Katharina Fietz, Malin Ed, Alison Farias and Michael Weber, and with inputs from Stella Carneiro, and under the strategic guidance of Pablo Acosta (Program Leader). The team would like to thank reviewers Indhira Santos, Vinicius Botelho, Truman Packard and Tiago Falcao for their insights and comments. Table of Contents OVERVIEW ............................................................................................................................ 1 I. INTRODUCTION .................................................................................................................. 6 II. UNEMPLOYMENT AND RECOVERY DYNAMICS SINCE THE 2014 ECONOMIC CRISIS .............. 6 III. CHARACTERIZING BRAZIL’S UNEMPLOYMENT PROTECTION SYSTEM ................................ 12 IV. BENCHMARKING BRAZIL’S UNEMPLOYMENT PROTECTION PROGRAMS WITH INTERNATIONAL COMPARATORS ......................................................................................... 19 SD BENCHMARKING ........................................................................................................................ 22 FGTS BENCHMARKING .................................................................................................................... 30 LABOR MARKET ACTIVATION OF SD BENEFICIARIES FROM INTERNATIONAL PERSPECTIVE ............................ 33 V. POLICY OPTIONS .............................................................................................................. 40 LEGISLATIVE CHANGES TO SD BENEFIT ............................................................................................... 42 MORE SYSTEMIC REFORM OF THE CURRENT PROGRAMS ........................................................................ 43 COMPLEMENTARY POLICIES TO EXTEND PROTECTION AND SERVICES TO THE EXCLUDED UNEMPLOYED ......... 45 REFERENCES ........................................................................................................................ 48 ANNEX 1: ADDITIONAL INFORMATION ................................................................................. 51 Figures FIGURE 1: UNEMPLOYMENT RATES AND GDP, 2012–2019 ....................................................................... 7 FIGURE 2: NUMBER OF UNEMPLOYED BY DURATION OF UNEMPLOYMENT ..................................................... 7 FIGURE 3: HIRING AND DISMISSALS IN FORMAL FIRMS 2010–2018 (MILLIONS) ............................................. 7 FIGURE 4: UNEMPLOYED (MILLION) DISAGGREGATED BY THEIR LABOR FORCE STATUS IN THE PREVIOUS SIX MONTHS 8 FIGURE 5: QUARTERLY AVERAGE TRANSITION PROBABILITIES BETWEEN DIFFERENT STATES OF THE LABOR MARKET (2018–2019) ..................................................................................................................................... 9 FIGURE 6: DIFFERENCE IN QUARTERLY AVERAGE TRANSITION PROBABILITIES IN THE LABOR MARKET ............... 10 FIGURE 7: EVOLUTION OF REQUESTS OF SD, JANUARY 2012–OCTOBER 2020 ............................................. 12 FIGURE 8: LABOR MARKET EXPENDITURE FINANCED BY FEDERAL GOVERNMENT (BLUE) AND EMPLOYERS’ LEVIES (ORANGE) IN 2018 ............................................................................................................................. 17 FIGURE 9: VALUE OF UNEMPLOYMENT PAYOUT, EXPRESSED IN MONTHLY WAGES, AFTER 24 MONTHS OF EMPLOYMENT .................................................................................................................................... 18 FIGURE 10: EFFECTIVE RECIPIENT RATE OF UNEMPLOYMENT BENEFITS: BRAZIL, REGIONAL BENCHMARKS (BLUE) AND ASPIRATIONAL COUNTRIES (ORANGE) .............................................................................................. 21 FIGURE 11: ACTIVATION ELEMENTS OF UNEMPLOYMENT BENEFITS ............................................................ 33 FIGURE 12: HAZARD RATE OF PARTICIPATING IN A JOB INTERVIEW OFFERED BY SINE SINCE THE DAY OF REGISTRATION IN SINE, FOR NON-BENEFICIARIES OF SD (LEFT) AND BENEFICIARIES OF SD (RIGHT) ................. 35 FIGURE 13: AN OVERVIEW OF POLICY OPTIONS ....................................................................................... 42 FIGURE 14: EXAMPLE OF UNEMPLOYMENT BENEFIT PAYMENT THROUGH COORDINATED FINANCING OF FGTS, MULTA AND SD: UNEMPLOYMENT BENEFIT BY MONTH - PRE UNEMPLOYMENT WAGE: 1.5 MWS; JOB TENURE: 24 MONTHS ....................................................................................................................................... 44 FIGURE 15: EVOLUTION OF REQUESTS OF SD BY EDUCATION, JANUARY 2012–OCTOBER 2020 ...................... 51 FIGURE 16: BENEFIT LEVEL AND REPLACEMENT RATE AT DIFFERENT WAGE LEVELS (BETWEEN 1 AND 3 MW) ..... 51 Tables TABLE 1: ELIGIBILITY CRITERIA AND PROGRAM RULES 2020 - SD ............................................................... 13 TABLE 2: RISK MANAGEMENT TOOLS IN BRAZIL FOR DIFFERENT FORMS OF WORK ........................................ 16 TABLE 3: EMPIRICAL LITERATURE ON SD UNINTENDED DISTORTIONARY EFFECTS .......................................... 18 TABLE 4: CHARACTERISTICS OF LABOR MARKETS IN BRAZIL AND COMPARATOR COUNTRIES FOR BENCHMARKING EXERCISE ........................................................................................................................................... 20 TABLE 5: INTERNATIONAL BENCHMARKING - COVERAGE CONDITIONS ......................................................... 23 TABLE 6: INTERNATIONAL BENCHMARKING ELIGIBILITY CONDITIONS ........................................................... 24 TABLE 7: INTERNATIONAL BENCHMARKING PAYOUT PERIOD ...................................................................... 26 TABLE 8: INTERNATIONAL BENCHMARKING REPLACEMENT RATE ................................................................ 27 TABLE 9: INTERNATIONAL BENCHMARKING FINANCING ............................................................................. 29 TABLE 10: INTERNATIONAL BENCHMARKING JOB SEARCH REQUIREMENTS ................................................... 35 TABLE 11: INTERNATIONAL BENCHMARKING SUSPENSION OF UNEMPLOYMENT BENEFIT ................................ 36 TABLE 12: INTERNATIONAL BENCHMARKING REASONS TO REFUSE JOB OFFERS............................................. 37 TABLE 13: INTERNATIONAL BENCHMARKING INCENTIVES TOWARD REEMPLOYMENT...................................... 39 TABLE 14: COUNTRY CLASSIFICATION OF THE UNEMPLOYMENT INSURANCE SCHEMES (OECD COUNTRIES) ........ 46 TABLE 15: SPENDING ON LABOR MARKET PROGRAMS BY TARGET GROUP AND PROGRAM TYPE ......................... 52 OVERVIEW The economic crisis induced by the COVID-19 pandemic underscored the importance of renewing social protection instruments to protect incomes from the shocks that are channeled through the labor market, in Brazil and around the world. The objective of this policy note is to inform the reform of Brazil’s unemployment protection programs, through a comparison of its salient features with international benchmarks, and by taking stock of the ongoing relevance of these instruments in Brazil’s changing labor market. The note builds on the Brazil Expenditure Review (World Bank 2018), and is informed by the World Bank’s recent white paper on adapting social protection instrument to a changing labor market (Packard et al. 2020). Brazil’s unemployment insurance, Seguro Desemprego (SD), is available to formal dependent workers1 conditional on complying with the required work history and being dismissed “without just cause”. These beneficiaries received on average a benefit of BRL 1,279 (US$243), 85 percent of the median wage of formal employees, for a mean length of 4.3 months. Dismissed workers also receive a cash benefit from the employer- financed individual savings account (Fundo de Garantia do Tempo e Serviço (FGTS)), and severance pay (Multas). SD and FGTS absorb almost three-quarters of Brazil’s (federal- and employer-financed) labor market expenditures totaling 2.3 percent of gross domestic product (GDP) in 2018. This amount is high compared to most regional and global benchmarks. Despite its cost, in 2019 only 17.7 percent of the average monthly unemployed (12.6 million) received unemployment benefits. 2 SD is key to protect formal workers from idiosyncratic shocks. However, on aggregate, the program plays a limited role in mitigating the social impact of crises, because those who are most hit by them tend to be ineligible. A large share of the workforce (informal workers, the formal self-employed) have no right to unemployment benefits, despite experiencing greater income volatility than formal employees. These workers can rely only on Bolsa Familia for minimum income protection, if eligible. Further, during recessions, formal firms in Brazil tend to adjust through a freeze in hiring rather than by increasing dismissals. After 2015, the chance of formal employees to lose their job decreased, while unemployment for all those already out of work or informal increased, particularly among youth. Last, the relatively short duration of SD does not support CLT workers during protracted periods of unemployment, which rise during crises. These dynamics explains why, after the 2014 crisis, unemployment increased but SD spending fell, while the number of eligible for Bolsa Familia continued to rise. During 1 Formal wage workers are defined by having a formal work permit ( Carteira de Trabalho) and are subject to Consolidation of Labor Laws (Consolidação das Leis do Trabalho (CLT)). 2 Compare this to 12.8 million unemployed on average per month during 2018. 1 the 2020 COVID-19 crisis, SD reacted more countercyclically, as formal firms had less time to adjust by reducing hiring. Nonetheless, for the same reasons above, the emergency cash transfer Auxilio Emergencial (AE) became necessary to protect the large share of SD- ineligible workers. If the recovery after the COVID-19 crisis mirrors the recovery of the 2014 crisis, net employment growth could occur in good part through jobs that do not provide access to SD and FGTS, such as formal self-employment and informal, thus leaving a large share of the population more vulnerable even when they return to work. For all these reasons, a broader toolkit of labor market policies to support the unemployed is urgently needed, especially after the end of AE. For Brazil to afford this, and given that total spending is already significant, improving the efficiency of the existing unemployment protection programs remain essential. This note contributes to this discussion through a detailed benchmarking of Brazil’s unemployment programs with regional and aspirational comparators. Among the considered benchmark countries, Brazil stands out for having one of the lowest coverage rates of the unemployed by unemployment benefits. The benchmarking also shows that Brazil’s unemployment programs are outliers in several aspects:  Generosity. SD’s replacement rate for the average benefit represents 80 percent of the wage before dismissal, with peaks of 100 percent for workers at the minimum wage (MW). This is above the replacement rate of any considered benchmark country. When adding to SD the income from Multas and FGTS, most dismissed workers are likely to receive the first month significantly more than their regular labor income.  Duration. In contrast to being generous, the payout period of SD (three to five months) is shorter than almost all comparator countries.  Eligibility. With 12 months of formal work out of the last 18 as a condition to receive SD for the first claim, access to the benefit is slightly harder in Brazil than other countries. This long contribution period excludes poorer workers and youth, who tend to have fragmented job histories. On the other hand, the contribution requirement for SD becomes less stringent after the first claim, which is not observed for other countries.  Financing. SD is also different than most of the comparators, as there is no clear linkage between contributions and expenditure. Financing for the SD fund (Fundo de Amparo ou Trabalhador (FAT)) derive from two taxes on formal firms’ revenues3, regardless of the size of payroll or to the frequency of dismissal in these firms. 3 The Programa de Integração Social (PIS), and Programa de Formação do Patrimônio do Servidor Público, (PASEP) 2 The high initial payout of unemployment programs, coupled with their low duration, results in incentives to overuse unemployment benefits, but in a suboptimal allocation of support that does not benefit those prone to long-term unemployment. Gerard and Naritomi (2019) for instance showed that, after receiving SD, recipients’ household expenditures grew by one third, but once benefits run out consumption falls significantly below pre-unemployment levels. To reduce moral hazard, most countries require availability to resume working in exchange for receiving unemployment benefits. Typical conditions include (a) proof of regular autonomous job search, (b) participation in proposed job interviews, and (c) availability for participation in active labor market policies (ALMPs). While the Brazilian legislation includes most of these elements, the SD law also contains more exceptions to refuse job offers than shown in international benchmarks. In addition, in Brazil autonomous job search efforts are little supported, nor are they incentivized or monitored. Empirical analysis from 2016 show that the few SD recipients who participate in job interviews organized by the National Employment System (Sistema Nacional de Emprego, SINE), do so only when the SD benefit is about to end. The analysis also found that most interviews do not result in a job match, indicating also that the quality of intermediation and vacancies has room to improve. A root cause for weak enforcement of ‘activation’ is the low spending on ALMPs and labor intermediation services, compared to any benchmark. In 2018, expenditures on labor intermediation in Brazil was equivalent to less than 0.01 percent of the combined spending on SD and FGTS paid for dismissals. IPEA (2017) estimates that even minor improvement in the effectiveness of ‘activation’ would result in large fiscal savings on SD. The 2021-2024 FAT budget, if confirmed, includes a moderate increase in federal grants to SINE offices. This would be a first step in the right direction, and very complementary with a technology-centered reform project of job matching and workers profiling (IPEA 2020). Addressing these challenges will require regulatory reforms to the existing unemployment benefits, and reinvestments of savings to develop modern labor market support systems that reduce moral hazard, but also serve the less protected half of the labor market, to whom almost no expenditures are devoted. This note focuses particularly on recommendations for the first piece of the puzzle – unemployment protection programs – presented according to varying degrees of complexity:  A reduction of the replacement rate of SD, and contemporaneous lengthening of the maximum benefit duration, could start addressing some of the incentive issues of the benefit, and at the same time enhance the coverage rate of the long-term unemployed. Consistent with international experience, the replacement rate could be reduced gradually after the first few months. The reform could be designed to be fiscally neutral. 3  To further improve the coverage of SD, the vesting period to be eligible for the first- time claim could be reduced, and the requirement for subsequent claims increased. Less weight could be put on uninterrupted contributory periods.  A review in the SD legislation of the allowed reasons to refuse a job offer, more in line with international standards, could attenuate the barriers to the implementation of activation by the labor offices. Benefit top-ups could be also provided to unemployed willing to immediately engage in job search assistance or ALMPs, which would serve as pilots for more comprehensive operational reforms.  A more comprehensive reform in line with international best practices would designate FGTS as the ‘first payer’ of income support after dismissal, distributed on a monthly basis. The SD benefit would be paid when income from FGTS account is depleted. FGTS would also need to be rewarded more in line with market rates. Simulations suggest that the reform results in shorter average unemployment duration, and in a lower deficit of FAT.  The costs of unemployment protection could be more directly associated with firms that generate it, through a reform of the financing system of labor market programs. One option would be financing SD through contributions from workers’ payroll taxes, as it is the case in most countries. Brazil could also consider the example of some countries that make employers’ contribution to the unemployment fund proportional to its use: firms that create more unemployment also cover more of its social costs. One way to operationalize this in Brazil would be to devote the current Multas largely to FAT. In addition, the FAT and FGTS could be fully dedicated to support the modernization of the labor market, including re-employment programs. Equally important will be complementary policies can broaden access to risk management tools in the labor market, for the populations excluded from SD, either by helping more workers formalize, or to support those in autonomous work arrangements. These policies are not dealt with in equal depth in this note, but mentioned for completeness:  One instrument to support formalization are targeted wage subsidies, suitable for workers that are below the minimum productivity threshold set by the MW, such as unemployed youth without work experience (See World Bank 2020).  In addition, regulatory reforms and simplifications of labor regulations can reduce the implicit costs and uncertainties of using labor contracts as well as increase formal hiring without fiscal costs. This effect was recently demonstrated through experimental evidence in South Africa, a country that like Brazil has a complex labor code (Bertrand and Crepon, 2020). 4  Human capital enhancing interventions, in the medium term, have longer lasting effects on workers’ employment and earnings than other programs. In Brazil, rolling this out this in scale may require a re-orientation of at least part of training institutions (Sistema-S), or the private sector, to work in synergy with unemployment benefits, including through performance-based financing (See World Bank 2021 for a review).  With the appropriate design considerations, Brazil could also offer instruments for financial self-insurance for the growing number of self-employed, through incentivized individual savings accounts. Such schemes exist in a handful of middle- income countries, and they are usually bundled with incentivized long-term savings for retirement of the self-employed. These programs adopt behavioral nudges, and specific rules for withdrawal and incentive to save, to build an additional layer of financial protection that the self-employed can rely on in hard times. As in most countries, introducing changes to labor market programs is also politically complex. This explains why in Brazil, and many other countries, the design of labor programs still reflects development models that assumed formal dependent employment as the main way to engage in the labor market, and which assigned to employers a large role in the provision of social welfare (See Packard et al, 2020). The COVID-19 crisis laid bare the limits of these premises to deliver protection, and the vulnerability of about half of the workforce in Brazil. This time may be the opportunity to renew the social contract that underpins the status quo. An improvement in labor market services and programs for a broader share of the workforce could generate the needed public support for a reform of the cash benefits treated in this note.4 4These complementary reforms were treated extensively in other recent publications. See Almeida, Silva, and Strokova 2015; Inter-American Development Bank 2018; Instituto de Pesquisa Econômica Aplicada 2016; World Bank 2018. 5 I. INTRODUCTION The objective of the policy note is to inform reforms of the unemployment protection system of Brazil, through a comparison of its salient features with international benchmarks, and by taking stock of employment dynamics since the 2014 crisis. Recommendations include varying degrees of ambition (and political complexity), from parametric reforms of the unemployment insurance Seguro Desemprego (SD) to more systemic reorganization of programs, and the extension of coverage to growing shares of non-dependent workers. The note builds on prior analysis of Brazil’s labor market programs undertaken as part of the World Bank’s expenditure review for Brazil (published World Bank 2018). The main novelties include a broader and deeper international benchmarking, as well as analysis of the unemployment dynamics since 2015, which are important to inform the optimal redesign of unemployment programs. The note was developed before the COVID-19 pandemic, and, while it takes stock of recent trends, it is not meant to be an assessment of Brazil’s emergency response or do justice to the emerging literature and data on the matter. As Brazil prepares for its economic recovery, the note aims to provide valuable information for a national policy debate around the much-needed modernization of the unemployment protection system. The note is organized as follows. The first part of the note takes stock of some of the key characteristics of Brazil’s labor market and unemployment dynamics since the 2014 economic crisis using the National Household Sample Survey Pesquisa Nacional por Amostra de Domicílios Contínua (PNADc) survey data. The centerpiece of the analysis is the comparison of the design features of SD to a range of international programs of reference countries, through a benchmarking exercise of its regulation, generosity, activation measures, and effective coverage rate. The final part of this note summarizes the key findings and provides recommendations. II. UNEMPLOYMENT AND RECOVERY DYNAMICS SINCE THE 2014 ECONOMIC CRISIS The 2014 economic crisis caused, with a lag, a spike in unemployment largely affecting young people. Total unemployment doubled from just below 7 percent in 2014 to a peak of nearly 14 percent in 2017. A full recovery never took place before the most recent COVID-19 shock, with unemployment still at 11 percent before the pandemic (Figure 1). The crisis’ impact, however, was heterogenous and particularly hurt young people, who experienced unemployment rates as high as 25 percent. Moreover, youth (ages 14–29) are also the most numerous in absolute terms due to their demographic weight, representing 54 percent of all unemployed in 2019. 6 Figure 1: Unemployment Rates and GDP, 2012–2019 Source: Authors based on PNAD Continua 2012–2019. The 2014 crisis also led to a change in the nature of unemployment, which became more permanent and disproportionally affected low-skilled workers. The number of long-term unemployed increased significantly: the total number of those out of work for 12 to 24 months more than doubled in the post-crisis period (Figure 2). Furthermore, there was a sharp rise in unemployment for low- and medium-skilled workers. Figure 2: Number of Unemployed by Duration of Figure 3: Hiring and Dismissals in Formal Firms Unemployment 2010–2018 (millions) Precrisis 2012Q2-2014 Crisis 2015-2017 Contract terminations Post 2018-2019Q2 Dismissals 1,800,000 35000 1,600,000 30000 1,400,000 25000 1,200,000 1,000,000 20000 800,000 15000 600,000 400,000 10000 200,000 5000 - 1 4 7 10 13 16 19 22 36 72 108 0 Months of unemployment 2010 2011 2012 2013 2014 2015 2016 2017 2018 Source: (2) Authors based on PNAD Continua 2012–2019. (3) Authors’ based on RAIS 2010–2018. Note: (2) The figures show the quarterly average number of unemployed, by duration in a month as declared by the respondent, in different periods (3) In Figure 3, dismissals are a subset of the total contract terminations. 7 The sharp rise in the number of unemployed after the 2014 crisis occurred among those who did not have a previous formal job, and thus were ineligible to SD. Panel data analysis suggest that a large share of the growing unemployed were either students looking for the first job, informal workers, or those who were already inactive (Figure 4). The number of unemployed who were formerly in jobs protected by SD did not vary substantially. Figure 4: Unemployed (million) Disaggregated by Their Labor Force Status in the Previous Six Months 6.5 6 5.5 5 4.5 4 3.5 3 2.5 2 1.5 1 0.5 0 Student Unemployed/Out of labor force Informal worker CLT employee Formal Self Employmed Source: Authors based on PNAD Continua 2012–2019. Note: The figure shows the labor status of the unemployed two quarters before any given period. This dynamic is consistent with the fact that formal firms in Brazil tend adjust through a slowdown in hiring, rather than by increasing dismissals. Analysis of administrative data (Relação Anual de Informações Sociais (RAIS)) from 2010 to 2018 shows that during the crisis employers reacted mainly through a hiring freeze, rather than through layoffs (which fell year-on-year). As Brazil was characterized by high job turnover, this resulted in large net job losses (Figure 3). Similar dynamics are also observed in the COVID-19 crisis, with most net employment losses in the second trimester of 2020 due to the freeze in new hires, rather than increase in dismissals (Box 1). As a result, the probability of falling into unemployment after 2015 disproportionately increased only for the informal workers and the self-employed, exacerbating labor market segmentation. Quarterly panel data in PNAD show that during the years after the crisis (Figure 6), all categories of workers, except formal employees, experienced greater risk of falling into unemployment, compared to pre-2015 times. In addition, this analysis shows that during the recovery the main way to transition out of unemployment was through informality and self-employment (Figure 5). The rapid growth in self-employed after the 2014 crisis puts even more pressure on revising the current unemployment protection system in Brazil. Comparing these transition rates with those in 2012-2014 8 (pre-crisis) (Figure 6) shows how the 2014 crisis exacerbated segmentation between formal dependent workers and the rest of the labor market.5 For all of these reasons, after 2015 higher unemployment resulted in increase pressure on social assistance claims (Bolsa Familia), while claims to SD steadily fell. As transition to unemployment increased more significantly for informal workers and independent workers, ineligible to receive SD, the SD claims fell as job rotation in formal firms declined (Box 1). On the other hand, after 2015, joblessness and poverty among those who were never in the formal sector in the first place began to rise and consequentially also the number of families eligible for Bolsa Familia. Figure 5: Quarterly Average Transition Probabilities between Different States of the Labor Market (2018–2019) Quarterly transition probabilities - 2018-2019 0.80 Formal Self Employment Formal Wage 0.60 Employment Inactive - Discouraged 0.40 Inactive - Others 0.20 Informal Self Employment - Informal Wage Formal Self Formal Inactive - Inactive - Informal Self Informal Unemployed Employment Employment Wage Discouraged Others Employment Wage Unemployed Employment Employment Source: Authors’ based on PNAD Continua panel data. 5 Since part of Brazil’s high turnover rate could be explained by worker’s decisions, it is possible that the increase in firing rates by employers was attenuated by a decrease in workers voluntary separations. 9 Figure 6: Difference in Quarterly Average Transition Probabilities in the Labor Market 0.20 Unemployed 0.15 Change in average probability of transition 0.10 Informal Wage Employment 0.05 Informal Self (pcerntage points) Employment - Inactive - Others (0.05) Inactive - (0.10) Discouraged (0.15) Formal Wage Employment (0.20) Formal Self Formal Self Formal Inactive - Inactive - Informal Self Informal Unemployed Employment Employment Wage Discouraged Others Employment Wage Employment Initial labor status Employment Source: Authors’ based on PNAD Continua panel data. During the COVID-19 pandemic, SD claims exhibited a temporary spike; however, much broader income support via Auxilio Emergencial (AE) remained necessary. In contrast to the previous crisis, during the COVID-19 recession SD claims increased, because firms required a deeper adjustment than what could be obtained through a halt in hiring (which also happened). This spike was especially pronounced once the emergency wage subsidy (Beneficio Emergencial Manutencao Emprego (BEm)) 6 began to expire (See Box 1 and Figure 7). World Bank simulations (Cereda et al, 2020) suggested that thanks to SD and the employer-financed individual savings account (Fundo de Garantia do Tempo e Serviço (FGTS)), the increase in poverty resulting from losses of formal jobs was mitigated, but also show that the top two quintiles benefitted the most from it7. For the majority of workers ineligible to SD, the Government of Brazil (GoB) had to introduce emergency income support AE, and this much larger program delivered the bulk of poverty mitigation during the pandemic. 6 BEm was established through Medida Provisória No. 936 to subsidize shorter work time by 25 percent, 50 percent, and 70 percent initially for three months and has been extended for up to 240 days. In return, the firm agreed to employ the worker afterward for the same duration as the benefit was used. The firm could also choose to suspend the contract for up to 240 days, with the condition to hire the worker for the corresponding time after the suspension ends. The benefit amount is equivalent to the SD benefit. The estimated cost of the program is BRL 51.2 billion (US$9.7 billion), of which BRL 28.97 billion (US$ 5.5 billion) has been spent as of October 2020. 7 Under an assumed GDP contraction of 8 percent, (Cereda et al, 2020) estimated job losses of 10.8 million formal workers for 2020, and an increase in total poverty rates by 18 percentage points (if unmitigated). The authors estimate that the combination of SD, FGTS and Multa buffer 30 percent of the resulting income losses. However, the simulation also points out that especially the two richest quintiles benefit from these instruments, as fewer workers in the bottom quintiles have formal jobs. 10 Box 1: The Response of SD after the COVID-19 Economic Shock and after the 2014 Crisis The response to the COVID-19 crisis in the Brazilian labor market is twofold. On one side, expansion of the unemployment insurance program SD in 2020 and through 2021 is playing a key role as the main safety net for formal sector workers. On the other side, the GoB launched the Emergency Employment and Income Maintenance Program BEm. By October 2020, the program recorded over 18 million agreements by over 1.4 million employers. Furthermore, additional measures such as the Saque Aniversário, and an FGTS emergency cash-out, were introduced. This measure makes it possible for every FGTS account with a balance to make a withdrawal of up to BRL 1,045 (US$198.5) per worker in accordance with a payment schedule. The extent to which the crisis hit the formal labor market can be seen in the peak of unemployment insurance claims in May 2020 (Figure 7). The number of monthly beneficiaries peaked in May 2020—from approximately 497,000 in March 2020 to 852,000 in May 2020. From March to May 2020, monthly new claims increased from 537,000 to 960,000 followed by a decrease in claims, reaching 461,000 in October 2020. More than half of those claimants of SD (53 percent) were earning only up to 1.5 monthly MWs before layoff, of which around 48 percent were women. Nevertheless, the creation of BEm—and especially its expansion—played an important role in the response strategy of the government, which is shown through a V-shaped recovery and a gain in formal employment in July, August, and September of 2020. At the peak of the COVID-19 crisis in June 2020, the labor market experienced a decrease in hiring by 18.3 percent and an increase in dismissal by 1.3 percent. However, the recovery started in the second quarter of 2020, showing a slightly positive trend with a current net balance of loss in formal employment in 2020 of 558,597 (1.4 percent). From January to September 2020, hiring decreased by 14.5 percent but interestingly dismissals also decreased by 4.1 percent in comparison to the same period of the previous year. The service sector suffers the most during the ongoing crisis with a reduction in formal employment by over 400,000, followed by the trade sector, which saw formal employment decreasing by over 340,000. The response of SD to the crisis during COVID-19—even after the meeting effect of BEm—differs from the reaction of the program in the previous 2014 crisis. In 2015, the labor market response was characterized by a fall in hiring rather than an increase in dismissal, which led to a fall in SD claims. In contrast, the demand shock for firms during the peak of the COVID-19 crisis was deeper and more concentrated and likely could not be buffered through a gradual slowdown in hiring (employment fell by 531,765 in 2016, compared to 1.2 million between March and June 2020). Even though SD claims fell in the third quarter of 2020, the peak in May was the highest since 2012 (Figure 7). 11 Figure 7: Evolution of Requests of SD, January 2012–October 2020 1,200 2014 Crisis COVID-19 crisis SD claims in thousands 1,000 800 600 400 200 0 Jun-12 Apr-13 Jun-17 Apr-18 Nov-12 May-15 Nov-17 May-20 Jul-14 Jul-19 Dec-14 Oct-15 Aug-16 Dec-19 Oct-20 Jan-12 Sep-13 Feb-14 Mar-16 Jan-17 Sep-18 Feb-19 SD Formal (up to 1.5 MWs) SD Formal (> 1.5 MWs) Source: Painel de informacoes do Seguro-Desemprego, Ministry of Economy, accessed December 16, 2020. Ministry of Economy, Secretaria de Trabalho 2020. Novo CAGED, Presentation, and Tresouro Nacional. Monitoramento dos Gastos da União com Combate à COVID-19, October 2020. III. CHARACTERIZING BRAZIL’S UNEMPLOYMENT PROTECTION SYSTEM SD is a constitutional benefit, in place since the mid-1980s, for employees with a formal contract and who are dismissed without justa causa (not at fault). To make a claim, the worker needs prior formal work history which, depending on the number of claims made in the past, ranges between 6 and 12 months. Furthermore, the worker should have no other form of labor income. The benefit is calculated based on the average of the past three salaries and is at least one minimum wage (MW). In case the beneficiary only received two month’s salary in his or her past job, the benefit will be calculated as the average of those two months, and if the beneficiary only received one month’s salary, the benefit will be based on that salary. Prior work history is also relevant for the payout period of the benefit. When making a first claim, the dependent wage worker needs a prior work history between 12 and 23 months out of the last 36 months to receive 4 months of SD benefit. To receive five installments of SD, the worker needs to have worked 24 months out of the last 36. (see Table 1). The stringent vesting period requirement for first-time claimants excludes the more vulnerable formal workers from SD access. Ongoing World Bank analysis of the formal employment registry (RAIS) and the social registry ( Cadastro Único) revealed that in 2014- 2016 the job-spells in the formal sector of individuals in Cadastro Único were, on average, less than 7 months per calendar year8, compared to about 10 months for the rest. This suggests that the more stringent vesting period for first-time claimants, in combination 8 Unpublished, forthcoming analysis by Sousa, Liliana et al (2018). 12 with the emphasis on continuous contributions, may exclude these low-income workers with intermittent careers from accessing SD. Table 1: Eligibility Criteria and Program Rules 2020 - SD Minimum Prior Minimum and Monthly Benefit Payout Period Based on Claim Work Maximum Replacement Rate Number and Worked Months in Requirement for Monthly Benefit Reference Period Eligibility Amount 1st claim: 12/18 Minimum: Up to BRL 1,599.61 1st claim: months 1 MW - BRL (US$303.8): 80% of 4 installments: 12–23/36 1,045.00 average past 3 months months 2nd claim: 9/12 (US$198.5) salaries 5 installments: 24/36 months months 2nd claim: Maximum: BRL 1,599.62– 3 installments: 9–11/36 months 3rd claim: BRL 1,813.03 BRL2,666.29: BRL 4 installments: 12–23/36 6/6 months (US$344.5) 1,279.69 + 50% of wage months above BRL 1,599.62 5 installments: 24/36 months Above BRL 2,666.29 3rd claim: (US$ 506.5): BRL 3 installments: 6–11/36 months 1,813.03 4 installments: 12–23/36 months 5 installments: 24/36 months Source: Labor Secretary, Ministry of Economy Brazil. SD is paid through the Fundo de Amparo ou Trabalhador (FAT) and is not related to the wage of the insured worker. The FAT is financed by taxes based on the firm’s total revenues (the Social Integration Program [Programa de Integração Social, PIS]/Training Program for Public Servants’ Heritage [Programa de Formação do Patrimônio do Servidor Público, PASEP]). The tax is not necessarily related to either the size of a firm’s payroll nor to the firm’s behavior in generating unemployment. In addition to the SD, the FAT also aims to finance salary bonus payment, labor training, and job creation programs. SD beneficiaries are also eligible for Brazil’s largest off-budget program, FGTS. This is an instrument managed by the government for mandatory saving with contributions paid by employers directly into individual accounts, which can be accessed in case of job loss. Reflecting the originally stated policy objectives of FGTS—to support workers in case of termination of the employment contract—involuntary, non-disciplinary dismissal, ’without just cause’ (sem justa causa) is the most frequent reason for withdrawals. However, over the course of the program’s history, the number of contingencies for which workers can draw on their mandatory savings has grown. But the program remains mainly focused on smoothing income in case of non-disciplinary, involuntary job loss. The FGTS in Brazil is not coordinated with the SD. In 2018, BRL 67 billion was withdrawn for dismissal without just cause, about 64 percent of the total for the year.9 An additional 9 The total payout in 2018 was BRL 111.4 billion (US$21 billion). 13 significant outlay relates to retirement age withdrawal, which is also related to work termination. The manner in which an account holder lost his/her job matters. The law requires employers who dismiss workers without just cause to pay a penalty. This penalty, or ‘multa’, is intended as a deterrent to dismissals and is analogous to the ‘risk rated’ contributions paid in the unemployment insurance program of the United States, which adjusts the ‘premium’ for the likelihood of dismissals in a particular sector. This additional penalty payment from employers can reach up to 40 percent of the total FGTS withdrawal. Brazil’s social protection instruments to manage labor market risks, as in most countries in Latin America, were designed under the assumption of formal dependent employment as the prevailing manifestation of work. This assumption, imported from industrial revolution Europe, crystallized in Latin America’s corporatist state models prevailing during most of the 20th century. This institutional legacy assumed that employers had the best capacity to administer risk management instruments for the working population (Packard et al. 2019). In the last two decades, Brazil made significant efforts to extend some protections to some of the informal workers, through the microentrepreneur regime Micro Empreendedor individual (MEI). 14 Table 2 illustrates the prevailing forms of work outside the public sector in Brazil and the level of access to risk management instruments. The table shows that segmentation was partially overcome for the risk of old-age poverty and disability (though with a persistent inequitable distribution of benefits, see World Bank 2018 for a detailed treatment). Unemployment protection is the area where the segmentation persists as only workers with a Carteira Assinada can access SD, FGTS, and severance pay.10 Other categories of workers, representing over half of the employed, can only rely on social program to weather income shocks. 10 40 percent of the payment goes to the worker. In the scope of the COVID-19 crisis, the additional 10 percent, which is normally paid to the government, has been waived. 15 Table 2: Risk Management Tools in Brazil for Different Forms of Work Form of work and Risks estimated number in Brazil (millions) Labor earnings Accident Disability Old-age volatility/unemployment poverty Formal employees (CLT, FGTS, SD, Multas Seguro RGPS Full RGPS public) (36.7 million) Accidente benefits benefits Trabalho Micro Emprededor BPC Basic Individual (2.5 million) RGPS benefits Trabalhador Autonomo Seguro RGPS Full or (6.2 million) Accidente benefits basic Means Tested Social Trabalho RGPS Assistance benefits Firm owners (Pessoa (= Bolsa Familia) RGPS Full RGPS Juridica) benefits benefits (0.6 million) Informal workers (self- BPC BPC employed and employees) (31.9 million) Source: Estimates from World Bank microsimualations model (BraSIM), based on PNAD Continua. Note: BPC = Benefício de Prestação Continuada; PJ= Pessoa Jurídica; RGPS = Regime Geral de Previdência Social. Number of workers based on World Bank BraSIM microsimulation model, using PNAD Continua 2017 monthly average for the primary job. As a result of this legacy, SD and FGTS typically absorb three-quarters of all spending on labor market programs in Brazil. When combining both employer- and budget-financed programs, the total spending on federal labor market programs reaches more than 2 percent of gross domestic product (GDP) per year, which is above the OECD average. At 0.53 percent of GDP in 2018, or BRL 36 billion, SD is the largest among federal expenditures (see Figure 8). Another salient characteristic of Brazil’s labor market programs is that they are skewed toward cash benefits and formal sector workers: within them, most spending is focused on unemployment protection for dependent wage workers. In fact, only 1.8 percent of labor market programs’ expenditure is accessible to workers who are not necessarily formalized (World Bank 2018). The implication of this distribution is that spending on active labor market programs (ALMPs) and spending on the majority of workers (who are either in the informal sector or self-employed) are low. These imbalances make Brazil stand out compared to benchmarks of similar spenders in OECD countries (see World Bank 2018 for the complete benchmarking). 16 Figure 8: Labor Market Expenditure Financed by Federal Government (blue) and Employers’ Levies (orange) in 2018 FGTS - Dismissal Without Cause withdrawals 67.87 LM Cash Benefits Seguro Desemprego, Defeso, Escravo 36.32 Abono Salarial 17.24 FGTS Complemento 5.03 Salario Familia* 2.00 Sistema S Training 24.62 LM Services Fed. Training (Bolsa Qualificação, Pronatec, Projovem) 0.66 Entrepreneurship, microcredit (Rural and Urban) 0.44 Labor Intermediation, Registries 0.17 Other (Economia Solidária, Agricultura Familiar) 0.02 0 10 20 30 40 50 60 70 80 BRL Billions, 2018 Source: Authors’ based on budget data from Portal Transparencia data for 2018. The system is generous in its replacement rate and minimum benefit, with the total value of unemployment payouts being higher than the pre-unemployment wage, especially for low-wage workers. Figure 9 depicts the income support that a worker with 24 months of tenure in a job can expect if dismissed from the combination of SD, FGTS, and Multa.11 For workers earning up to two MWs, the expected unemployment income support is equivalent to more than seven months of wages, distributed over five months (the majority at the time of dismissal). While the SD benefit declines in generosity based on wage, FGTS and Multas are constant, at around 2.8 previous wages (already offering a significant layer of protection for short-term unemployment). This abnormally high payout induces strategic behavior among a section of the unemployed, who have incentive to seek frequently unemployment, as shown by the growing body of empirical literature (see Table 3). The decline of total support, as wages rise, is a positive feature of the system, given that low-wage workers have the highest risk of long-term scarring and unemployment (Silva et al, 2021). 11 The graph assumes the minimum contribution history of 24 months to receive five installments of the SD benefit. 17 Figure 9: Value of Unemployment Payout, Expressed in Monthly Wages, after 24 Months of Employment 9 SD 8 Unemployment payout Multa (monthly wages) 7 6 FGTS 5 4 3 2 1 0 1 2 3 4 5 Pre-unemployment monthly wage (in MWs) Source: Authors’ calculation. Note: The calculation assumes a contribution history of 24 months to receive five installments of SD. Further, it assumes a constant wage level before unemployment and that the unemployed claims SD for the first time. The high payout at the time of dismissal contrasts with a relatively low duration of total payments. SD pays a maximum of five months for those with the longest contribution periods, which is two years. Of course, when summing up FGTS and SD, the total payout value is higher than five months, but its lump-sum distribution is suboptimal from the perspective of consumption smoothing (as shown by Gerard and Naritomi 2019), as the unemployed tend to underestimate the time it will take to find a new job. A three-to-five- month payment is slightly lower than other benchmarks and could be considered insufficient to protect the long-term unemployed. Furthermore, its mandatory contribution history, of at least 12-month work experience for the first claim, makes it especially difficult for the young unemployed to receive income support while being unemployed (see section SD in international perspective: results from a benchmarking exercise for a more detailed comparison of SD and international examples). Table 3: Empirical Literature on SD Unintended Distortionary Effects Study Main Findings Gerard and Naritomi (2019) Dismissed workers eligible for both SD and severance pay increase consumption at layoff by 30%, despite experiencing a 17% consumption loss after they stop receiving any benefits. This is explained by a present bias in workers in intertemporal consumption choices. Carvalho et al. (2018) The authors find evidence consistent with workers having the incentive to strategically induce their layoffs so they can collect SD benefits. They estimate it accounts for 11–13 percent of the average dismissal rates of eligible workers. Doornik et al. (2018) The authors find evidence consistent with UI positively affecting turnover in the formal sector. They estimate that eligibility for SD in Brazil increases unemployment inflow by 12 percent and that such 18 Study Main Findings behavior is related to workers shifting sectors toward informal employment. Gerard and Gonzaga (2016) The authors find evidence of strategic behavior of SD beneficiaries consistent with moral hazard and that part of the observed response is driven by informality. They find that most of the SD beneficiaries exhaust the benefits and that their probability of reemployment in the formal sector peaks right after exhaustion of the benefits. Furthermore, the estimated share of displaced formal workers finding any new job does not increase after SD exhaustion, while the share of those finding a formal job does, which they explain as beneficiaries already reemployed shifting sectors from informal to formal employment. Hijzen (2011) The author finds evidence of moral hazard and incentives for SD beneficiaries to work informally. He estimates that the SD benefits disproportionally reduce the reemployment rate in the formal sector compared to the informal in the first months of non-employment. Additionally, he estimates absence of a significant effect near the spike in the reemployment rate in the formal sector around benefit exhaustion, which is consistent with job losers taking up informal jobs while receiving SD. Gonzaga (2003) The author argues that SD in Brazil increases turnover in the formal labor market. He argues that it occurs due to incentives of the program toward collusion between employee and employer to induce the layoff to collect SD benefits. Chahad and Fernandes (2002) The authors find evidence that SD affects participation of non-head households due to a high replacement rate to which they are subjected to due to lower average wages. They estimate that SD benefits increase the duration and frequency of nonparticipation in the labor market. IV. BENCHMARKING BRAZIL’S UNEMPLOYMENT PROTECTION PROGRAMS WITH INTERNATIONAL COMPARATORS In this section, Brazil is benchmarked with two sets of countries: regional comparators with similar labor markets and a selection of aspirational peers from high-income countries. The objective of the benchmarking exercise is to provide a more nuanced comparison of Brazil’s system. To achieve this, two categories of comparators are chosen: (a) regional benchmarks in Latin America, which share more similarities with Brazil in terms of fiscal scope and capacity to implement unemployment systems, and (b) international best practice countries with universally relevant and aspirational parameters. The best practice comparators include Germany, Portugal, Republic of Republic of Korea, and Sweden and will from now on be referred to as ‘aspirational countries’. The chosen regional benchmarks (Argentina, Colombia, Chile, Ecuador, and Uruguay) are appropriate benchmarks because they have (a) a similar labor market context to Brazil, with a relatively high MW compared to the distribution of income, high 19 informality, and low labor market spending, particularly spending associated with ALMPs, and (b) a UI system in place. The inspirational comparators are different for Brazil in terms of labor market context, but each exemplifies different aspects of best practice approaches, discussed in the note. Table 4 shows key characteristics of the countries chosen as benchmarks and Brazil. Table 4: Characteristics of Labor Markets in Brazil and Comparator Countries for Benchmarking Exercise Brazil Regional Benchmarks Aspirational Countries Informality rate 45% Chile (30%); Sweden, Germany, and Colombia, Ecuador (60– Portugal (Less than 20%); 70%) Republic of Korea (20–30%) Unemployment 12.1% Colombia, Argentina (9– Portugal, Sweden (6–6.5%); ratea 10%); Chile, Uruguay (7–8%); Republic of Korea, Germany Ecuador (4%) (3–4%) Total Labor Market Active: 0.3 % Active: Argentina, Chile Active: Sweden (2.3%); spending as a % of Passive: 0.6 %b (0.2%); Uruguay, Peru, Germany, Portugal (1–1.5%); GDP Colombia, Mexico (0.0– Republic of Korea (0.5%) 0.1%) Passive: Portugal, Germany Passive: Uruguay, Argentina (1– 1.2%); Republic of Korea, (0.4–0.6 %); Chile, Peru, Sweden (0.3–0.4%) Colombia, Mexico (0.0%) Minimum wage 70% Colombia (89%), Chile (69%), Portugal (60%), Republic of relative to median Argentina (46%) Korea (53%), Germany (49%) wagec Note: a. International Labour Organization (ILO) 2018: 2018 Labour Overview Latin America and the Caribbean. Numbers from 2018. b. Numbers from 2018: active programs: for example, Abono Salarial, Salario Familia, Pronatec. Passive programs: for example, SD, FGTS. c. OECD statistics database; IMF selected issues paper on Argentina, 2017; Almeida and Packard 2018. MW relative to median wages of full-time workers, data from 2017. The share of unemployed receiving SD benefits in Brazil (effective recipient rate) is low with respect to most of its international comparators. The combination of the various parameters to access UI and the characteristics of the labor markets determine the effective coverage rate of the unemployed by existing instruments in Brazil and its comparators. This is captured in the ‘recipient rate’, which is the simple ratio between the average monthly UI benefit recipients and total unemployed. This shows the share of unemployed people covered by the unemployment benefit. In 2019, in Brazil, on average, only 17.7 percent of the people who reported being unemployed received UI. Administrative data show that Germany, Portugal, Republic of Korea, and Sweden have a recipient rate between 37 percent and 60 percent. In benchmark countries such as Argentina and Chile, 7 percent and 46 percent, respectively, of the unemployed received income protection (Figure 10), whereas in Brazil the rate decreased from 33.6 percent in 20 2015 to 17.7 percent in 2019. The fall in recipient rate is explained by a decreasing number of monthly UI beneficiaries between 2015 and 2018 (from 2.8 million to 2.3 million) contemporary as an increasing number of unemployed (from 8.3 to 12.6 million). The overall low recipient rate can be explained by the relatively short payout period combined with large groups of ineligible unemployed. A low recipient rate can be explained by a combination of factors that characterize Brazil’s labor market: (i) ineligibility of many unemployed to SD due to their previous form of work (as shown in Table 2 above), (ii) the high share of first-time jobseekers with a long transition from school to work (as shown in Figure 4 above), (iii) an increasing number of long-term unemployed including those whose benefits expired (as shown in Figure 2 above), and (iv) the presence of many workers with intermittent and short job spells. Figure 10: Effective Recipient Rate of Unemployment Benefits: Brazil, Regional Benchmarks (blue) and Aspirational Countries (orange) 70% 60% 60% 46% 48% 50% Receipient rate 40% 37% 30% 32% 30% 18% 20% 11% 7% 10% 5% 0% Argentina Brazil Chile Colombia Ecuador Uruguay Germany Portugal Republic Sweden of Korea Selected countries Source: Authors’ based on ILO Social Security Database 2019,OECD Social Benefit Recipients Database (SOCR) 2016 and IPEA data. Note: The numerator of the rate is equal to the average monthly number of beneficiaries of unemployment insurance benefits. The denominator equals the total number of monthly unemployed (OECD SOCR, 2016). The rates correspond to 2016 numbers. For Brazil numbers have been updated to 2019 values. 21 SD BENCHMARKING In all the countries considered, dependent wage workers are eligible for UI, while only in a few countries are self-employed workers eligible (though not in Brazil). Most Latin American countries, including Brazil, offer UI for dependent wage workers and exclude self-employed workers. The only country in the region where self-employed workers have access to unemployment protection is Colombia, which offers voluntary savings account schemes for self-employed workers (see section regarding FGTS benchmarking). In many European countries such as Germany, Portugal, and Sweden, voluntary coverage for formal self-employed workers is offered (see Table 5). The minimum required contribution history, or vesting period, for SD is slightly more stringent than most of the other international comparators. In all countries, a certain length of contribution history is required before receiving the unemployment benefit. The work history can range from 6 up to 24 months. The contribution requirements are slightly stricter in Brazil for the first claim but ease up the second and third time of unemployment. Some countries such as Argentina, Chile, or Uruguay differentiate between employment contracts and base the required work history on the type of contract. In Chile, for example, workers holding permanent contracts are required to put in more months of contribution than workers on fixed-term contracts (see Table 6). In most countries (including Brazil) the reason for unemployment plays a role when making a claim. Termination without just cause is the prerequisite for SD in Brazil. This means that an employee is terminated from a job not because of misconduct but because the employer has decided that it no longer needs the employee. Uruguay’s unemployment insurance has the same regulations. In most other countries (both the aspirational and regional benchmarks), involuntary job loss is the requirement to receive benefits. In Germany and Sweden, the receipt of unemployment insurance is not dependent on the reason for unemployment. The unemployed only has to register in a public employment office and actively search for a job to receive the benefit (see Table 6). SD relies on a rather short payout period compared to other countries. Overall, the unemployment benefit payout period is shorter for countries in Latin America than in the aspirational countries. In Argentina, the period varies depending on the contribution history and age of the unemployed. In Colombia, Ecuador, and Uruguay, the payout period is similar to the one in Brazil, ranging between five and six months. Countries outside Latin American have implemented longer payout periods. Germany, for example, offers the UI benefit up to 24 months. Republic of Korea and Sweden require a waiting period of seven days before receiving the UI benefit (see Table 7). The replacement rate of past wages provided by SD is high compared to international standards. The replacement rate of 80 percent for workers with pre-unemployment earnings up to 1.5 MW is high compared to its regional benchmarks in the region and to 22 more developed countries (see Table 8). This rate would be even higher when the additional income from FGTS and Multas is included (see Figure 9). Some countries such as Chile or Ecuador rely on decreasing the replacement rate over the payout period, which aims to incentivize a fast reintegration in the labor market. In Argentina, the replacement rate depends on the previous work experience, whereas in Germany unemployed with dependent children receive a higher replacement rate than unemployed without dependent children. The financing of the unemployment benefit system in Brazil differs substantially compared to other systems. In most countries, unemployment insurance relies on employer and employee contributions, in combination with governmental subsidies. In addition to this, the size of contributions are often based on employees’ gross earnings. In Argentina, Colombia, and Ecuador, for instance, the employer contributes between 0.9 percent and 4 percent of the monthly payroll to UI. In most of the aspirational countries, both the employer and employee contribute to UI. In Brazil, however, taxes on firm revenues finance SD with no participation of employees (see Table 9). Table 5: International Benchmarking - Coverage Conditions Country Instrument Coverage Private sector employees, including temporary and casual workers Exclusion: self-employed, domestic workers, public sector Argentina UI employees, private schoolteachers, and lecturers at private universities UI Private sector employees, domestic, fishermen Unemployment Brazil insurance savings Exclusion: self-employed account (UISA) (FGTS) UI (Subsidio de Individuals who are involuntary unemployed and not covered by Cesantía) Seguro de Cesantia Salaried paid workers if employed since 10/2002 (voluntarily Chile UISA (Seguro de coverage if employed before) Cesantía) Exceptions: government workers, domestic workers, apprentices, youth below 18 years, and self-employed UI Employed persons UISA (Cesantías) Mandatory coverage for dependent workers Voluntary coverage for self-employed and employees with monthly earnings above Col$10,765,508 Colombia UISA (supplemental voluntary individual account) UI Employed persons Ecuador UISA Exclusion: self-employed persons 23 Country Instrument Coverage Private sector employees in industry and commerce, rural workers, government employees with a fixed-term contract of at least two years, professional athletes, members of cooperatives, Uruguay UI and household workers Exception: self-employed persons Special system for bank employees Employed persons, including household workers, apprentices, and trainees and certain other persons, including recipients of sickness Germany UI benefits and persons raising a child Voluntary coverage for self-employed Portugal UI Employed person and certain types of self-employed Employed persons Republic of Employment Voluntary coverage for certain small businesses Korea Insurance Exclusions: household workers and family labor Voluntary income-related insurance: unemployed and self- Voluntary UI employed persons who are members of an unemployment Sweden insurance fund. Employed and self-employed persons not covered by the Social insurance voluntary income-related program. Source: ILO Social Security Database 2019; ISSA Social Security Around The World 2019; SSA Social Security Country Profiles 2019. Table 6: International Benchmarking Eligibility Conditions Country Instrument Contribution History Reasons for Unemployment/Additional Conditions 6 months in last 3 years, 90 days in 12 months for temporary workers, and Involuntary unemployed or quit for a Argentina UI 240 days in last 24 justifiable reason months for construction workers 12 out of 18 months for first time unemployed, 9 out of 12 months for Dismissal without just cause, not due to UI second time unemployed, misconduct or resignation Brazil and 6 out of 6 months for third time unemployed Involuntary unemployment, retirement, UISA (FGTS) — purchase property 12 out of 24 months - last UI (Subsidio 3 months need to be Involuntary unemployed, having exhausted de Cesantía) continuous with the same the UISA, up to 2 claims in the last 5 months Chile employer 12 months of UISA (Seguro contributions: permanent Unemployment, retirement de Cesantía) contract 24 Country Instrument Contribution History Reasons for Unemployment/Additional Conditions 6 months of contributions: fixed-term contract Enrollment in a family Register with an employment service, allowance fund for at participate in training and vocational UI least 1 year (2 years for rehabilitation, and have a certificate of self-employed) in last 3 employment termination. years UISA Dependent workers can only make (Cesantías) partial/full withdrawals under the following circumstances: For partial withdrawals:  For financing higher education of someone in the household  For housing acquisition or renovation  For land tax payments. For total withdrawals: Colombia Dependet workers:  Dependent workers: unemployed or retired, possible to finance family members’ tertiary education, purchase/renovate property  Mandatory military service Self-employed: No conditions for withdrawing the funds UISA Earning < 2 × MW: (supplemental contributed 10% of the voluntary insured average monthly individual wage in last year account) Earning > 2 MW: contributed 25% of insured average monthly wage in last year 24 months of contributions, including 6 continuous months of contributions UI Involuntary unemployed immediately before unemployment began, 2 Ecuador months of involuntary unemployment 24 months of contributions. The benefit UISA Unemployment, retirement is paid after two months of unemployment 6 months of 12 months or Not be the result of dismissal for disciplinary Uruguay UI 6 months of coverage + 5 reasons 25 Country Instrument Contribution History Reasons for Unemployment/Additional Conditions months of work if the worker is payed irregularly (special rules for rural workers + household workers) Registered at an employment office, actively Germany UI 12 out of 24 months seeking work. Involuntarily unemployed, registered at an Portugal UI 1 year in last 2 years employment office, and be capable of and available for work. Involuntary unemployment and not due to misconduct, a labor dispute, or the refusal of Republic EI 6 out of 18 months a suitable job offer; registered at of Korea employment security office and capable of and available for work Voluntary UI Registered as a job seeker at the public employment service and able and willing to Sweden Social 6 out of 12 months accept a suitable job for at least three hours insurance a day. Source: ILO Social Security Database 2019; ISSA Social Security Around The World 2019; SSA Social Security Country Profiles 2019. Table 7: International Benchmarking Payout Period Country Instrument Payout Period 2 months (if contributed between 6 and 11 months), 4 months (if contributed 12 and 23 months), 12 months (if contributed 24 and Argentina UI 35 months), and 18 months (older than 45 and children eligible for family allowance). 1st claim: 4 months (if contributed between 12 and 23 months of contributions) 5 months (at least 24 months of contribution) 2nd claim: UI Brazil 3 months (if contributed 9 to 11 months) 5 months (at least 24 months of contributions) 3rd + claims: 3 months: (6 to 11 months of contributions) 4 months (at least 24 months of contributions) UISA (FGTS) — UI (Subsidio de 1 year - Fondo Solidario: 3 months if contract is temporary, 5 Cesantía) months if contract is permanent (additional 2 months if the Chile national unemployment insurance rate is 1 percentage point UISA (Seguro de above the average of the last 4 years) (Fondo Solidario can be Cesantía) accessed 10 times in 5 years) Colombia UI 6 months 26 Country Instrument Payout Period UISA (Cesantías) Dependent workers can only request once per year a partial or total withdrawal depending on the reason for withdrawal. Self-employed/independent workers do not have conditions for making withdrawals. UISA (supplemental voluntary individual 6 months account) 5 months—after the 5th month the total balance of the individual UI Ecuador account can be withdrawn. UISA — 6 months (up to one year for workers ages 50 years or older and Uruguay UI up to two years for workers ages 58 years or older with at least 28 years of contributions) Germany UI 6–24 months Portugal UI Between 5 months and 18 months Republic of 90 days for those with 6–12 months of contributions EI Korea 240 days for those with more than 10 years of contributions 7-day waiting period Voluntary unemployment benefit: 80% of daily average income in UI (voluntary) Sweden last 12 months for first 200 days after a 7-day waiting period 70% for the next 100 days Social insurance Social insurance: 300 days after a 7-day waiting period Source: ILO Social Security Database 2019; ISSA Social Security Around The World 2019; SSA Social Security Country Profiles 2019. Table 8: International Benchmarking Replacement Rate Country Instrument Replacement Rate Minimum Maximum Benefit Benefit 6–11 months of contributions: 2 months: 50% of insured best wage in the last 6 months before unemployment (100% for rural workers) 12–23 months of contributions: 4 ARS 2,907.53 months: 50% of insured best wage in the (ARS 3,728 ARS 4,652.06 (ARS last 6 months before unemployment for rural 7,456.00 for rural Argentina UI (100% for rural workers) workers) workers) (0.37 24–35 months of contributions: month 9 (0.23 MW, MW, 0.6 MW for to 12: 37.5% of insured best wage in the 0.3 MW for rural workers) last 6 months before unemployment rural workers) (70% for rural workers) Extended period for person above 45 years with children eligible for family allowance 80% of average past 3 months salaries BRL 1,045 (1 Brazil UI BRL 1,813.03 up to 1,531 MW) 27 Country Instrument Replacement Rate Minimum Maximum Benefit Benefit 50% of average monthly earnings of 1,531.03 to 2,551.91 + lump sum of 1,224.82 Lump sum of 1,813.03 average monthly earnings higher than 2,551.96 Lump-sum account balance from current UISA (FGTS) employer Month 1–3: 17,338 UI (Subsidio de Month 4–6: 11,560 Cesantía) Month 7–12: 8,669 1st month: 70% (Fondo solidario: 50%) FS Fondo de Cesantía Solidario in Chilean pesos: 1st month: 1st month: 195,887 652,956 2nd month: 55% (Fondo solidario: 40%) 2nd month: 2nd month: 153,912 513,038 Chile 3rd month: 45% (Fondo solidario: 35%) 3rd month: 3rd month: 125,927 419,757 UISA (Seguro de 4th month: 40% (Fondo solidario: 30%) 4th month: 4th month: Cesantía) 111,936 373,118 5th month: 35% (Fondo solidario: 30%) 5th month: 5th month: 97,944 326,478 6th and following months: 30% 6th +: 83,951 6th +: 279,838 (from 1st to (From 1st to 6th 6th month: month: 2.07–0.88 0.67–0.27 MW) MW) Monthly family benefit, food voucher: UI value 1.5 MWs Lump sum of 1 month of the insured UISA annual salary paid for each year of (Cesantías) employment Colombia UISA Based on account balance (supplemental voluntary individual account) 1st months: 70% of average monthly earnings in the 12 months before unemployment UI Ecuador 2nd month: 65% 3rd month: 60% 4th month: 55%; 5th: 50% UISA Lump-sum account balance Between 66% and 40% of the insured’s UYU 5,574.33 UYU 61,329.58 Uruguay UI average earnings in the six months (0.35 MW) (3.9 MW) before unemployment 28 Country Instrument Replacement Rate Minimum Maximum Benefit Benefit 67% of net monthly earnings (60% Germany UI — without children) 65% of average gross salary during the 12 months period ending 2 months Maximum: 2.5 Minimum: before the start of unemployment times the monthly Portugal UI EUR 428.90 divided by 360 social benefit rate (0.61 MW) For self-employed persons, 65% of the (1.5 MW) reference earnings Minimum.: Republic 90% of EI 50% of average daily earnings of Korea minimum daily wage SEK 365 a day is paid if the insured was Maximum: SEK UI (voluntary) working 40 hours a week before 910 a day for the unemployment. first 100 days Sweden First 200 days: 80% of the insured's daily average income in the last 12 months Social insurance Next 100 days: 70% of the insured’s daily average income in the last 12 months Source: ILO Social Security Database 2019; ISSA Social Security Around The World 2019; SSA Social Security Country Profiles 2019. Table 9: International Benchmarking Financing Country Account Financing UI Employer funded: 0.92% or 1.09% of gross payroll (depends on the type Argentina of enterprise); minimum earnings to calculate contributions: ARS 4,009.94, no maximum. UI PIS/PASEP (tax on firm revenues) Brazil UISA (FGTS) Employee: 8% of wage Employer: 40% ‘multa’ of FGTS account balance, if dismissal is without just cause UI (Subsidio de Total costs financed through the Unified Daily Allowance and Cesantía) Unemployment Fund Employer: 1.6% of monthly covered payroll + 0.8% to the solidarity severance fund (2.8% of monthly covered payroll + 0.2% to solidarity Chile severance fund for fixed term contracts) UISA (Seguro Employee: 0.6% of monthly covered earnings + administrative fee of de Cesantía) 0.04% (fixed terms do not contribute) Government: 225,792 Unidad Tributaria Mensual (UTMs) to solidarity severance fund UI Employer: 4% of wage to Cajas de compensación familiar; 30% of it is devoted to the UI fund Colombia UISA Dependent work relationship: (Cesantías)  Mandatory for employer to contribute to the Cesantías: one month of the person’s annual salary per year 29 Country Account Financing Self-employed:  One month of the person’s declared earnings or wage per year UISA Dependent work relationship: (supplemental  A portion of contributions to mandatory account may go to voluntary voluntary account individual Self-employed: account)  A portion of contributions to mandatory account may go to voluntary account Ecuador UI Employer: 1% of gross payroll UISA Employee: 2% of gross earnings, Government: subsidies in special cases UI Social insurance: 15% employee + 7.5% employer of gross salary - Uruguay finance cash sickness and parental benefits, unemployment benefits, and family allowances UI Employee: 1.5% of gross earnings (maximum contribution base exists) Employer: 1.5% of gross earnings (maximum contribution base exists) Germany Self-employed: 3% of monthly reference value Government: financing of deficit Portugal UI 5.14% (employer + employee) EI Employee: 0.65 % of gross earnings (no minimum/maximum contribution base) Republic Employer: 0.9%–1.5 % of annual payroll (no minimum/maximum of Korea contribution base) Self-employed: 2.25% of declared wages (employment services only) UI (voluntary) Employee: membership fee for voluntary UI Employer: 2.64% of payroll Self-employed: 0.1% of earnings and membership fee for voluntary Sweden program Government: subsidies to basic program Social — insurance Source: ILO Social Security Database 2019; ISSA Social Security Around The World 2019; SSA Social Security Country Profiles 2019. FGTS BENCHMARKING Unemployment individual saving accounts (UISAs), similar to FGTS in Brazil, were implemented during the last century in many Latin American countries and in some places even the self-employed have access. UISAs have gained popularity in recent years as an alternative to the traditional UI system. Especially in Latin America, the idea of UISA attracted a lot of attention and different types of UISAs were introduced in several countries (for example, Argentina, Brazil, Chile, Colombia, Ecuador, Peru, and Uruguay).12 12 This analysis is based on the benchmarking tables above, Ferrer and Riddell 2011 and Vodopivec 2004. 30 In most cases, the UISAs in Latin America cover dependent workers. One exception is Argentina, where only construction workers are covered. In most of the regional benchmarks, the self-employed are excluded from the UISA. However, Colombia offers UISAs for the self-employed on a voluntary basis. UISAs are not offered in any of the aspirational benchmarks, but in some of those countries, for example, Sweden and Germany, the self-employed can access the regular unemployment benefit. While there are no requirements for the contribution period in Brazil, most of the regional benchmarks require a period of contributions to access the UISA. In Chile and Ecuador, unemployed individuals are required to have a minimum period of contributions to access the UISA. In Chile, this period is 12 months at the minimum and in Ecuador 24 months of contributions are required. To access the voluntary UISA in Colombia, an unemployed individual must have contributed a certain percentage of the insured average monthly wage. The contribution depends on the pre-unemployment wage. In Brazil, however, there are no specific requirements on contribution history to access the UISA. Apart from job loss, retirement is a valid reason to access the UISA in most regional benchmarks and in Brazil. In Brazil and Panama, in the case of job loss, involuntary dismissal is a prerequisite to access the UISA. In Argentina, Peru, Chile, Ecuador, and Colombia, a proof of dismissal is required, but there are no specific requirements on the reason of the dismissal. In addition to requiring proof of dismissal, Chile and Ecuador require a minimum contribution tenure to gain access to the UISA funds (around one year of contributions). For Brazil’s FGTS, retirement, chronic illness, and property purchase are also accepted reasons for accessing the benefit. This is also the case in Colombia and Panama. In Chile and Ecuador, apart from job separation, retirement is also a reason for accessing the UISA fund. In most countries, upon dismissal, the UISA pays a lump sum to the account owner. FGTS in Brazil pays out a lump sum of the account balance from the current employer. This is also the case in Ecuador and for the mandatory UISA in Colombia. In Chile, on the other hand, benefits are paid monthly, and the level is linked to past earnings, with a declining schedule. The benefit is first drawn from the individual account and upon depletion of a common solidarity account (see Box 2). In most countries in Latin America, the employer is mandated to contribute to the UISA, and employees can make voluntary contributions. Employers (and in some cases employees) are required to deposit a monthly contribution into an individual worker’s account. Normally in Latin America, only the employer makes contributions to the individual savings accounts. Unlike Brazil, most other countries in Latin America allow the employee to contribute to the UISA voluntarily. Chile is one of the few countries where the both the employee and employer contribute to the UISA. The employer contributions vary in Latin America. In Brazil, the employer contributions are relatively high: the 31 employer pays 8 percent of the monthly payroll to FGTS. This is in line with the employer contributions in Colombia where the employer pays one month of the employer’s annual salary to the mandatory UISA (approximately 8 percent per month). In Peru, Panama, and Chile, the employer contributions are lower. In Peru, for instance, the employer pays 8.33 percent of the monthly payroll every sixth month, and in Chile the employer pays 1.6 percent of the monthly payroll. In the case of self-employed individuals in Colombia, contributions are equivalent to one month of the person’s salary. Box 2: Unemployment Insurance Savings Accounts - The Case of Chile In recent years, UISAs have been introduced in several developing countries. These kinds of unemployment protection systems based on individual savings accounts do not risk pool between workers but pay out the savings accumulated by an individual in the case of unemployment. UISAs are beneficial in developing contexts because of their low fiscal cost and the low risk of moral hazard attributed to them. However, if a substantial proportion of workers cannot generate sufficient savings to draw from during unemployment spells, a UISA system might not be viable. One of the first countries to establish such an unemployment protection system was Chile. The UISA (Seguro de Cesantía) in Chile was established in 2002 and has since its introduction been a source of inspiration and a model for many countries. Seguro de Cesantía is financed by three actors—workers, employers, and the government. The system has two mechanisms of financing: (a) in the case of open-ended contracts, both the employer and the employee are obliged contribute to the saving account and (b) in the case of atypical contracts, only the employer is required to contribute to the savings account. The contributions each worker makes to his or her UISA constitute the worker’s personal savings and can only be withdrawn in the case of unemployment, termination of contract, retirement, or any other event in which the worker leaves his or her job. The contributions are 3 percent of the monthly salary for workers with a fixed term contract and has a cap at Ch$3.3 million (US$4,500) and the maximum contribution period is 11 months. Anyone who has contributed to the Seguro de Cesantía and remains unemployed can withdraw from the fund. However, workers who are self-employed, employers, and informal workers are excluded from the UISA system. Moreover, workers younger than 18 years and older than 65 years are excluded. To withdraw income support, workers must have a written contract and must have contributed for 12 months in the case of an open-ended contract and 6 months in the case of an atypical contract over the course of the last 24 months. Replacement rates decrease from 50 percent to 20 percent over the duration of seven months. If the worker switches from one job to another, the worker can choose whether he or she wants to withdraw the funds or leave them for future withdrawals. 32 The UISA in Chile is integrated with a risk pool account. Fondo de Cesantía Solidario is funded by the government as well as employers and acts as a complement to the Seguro de Cesantia. If the worker’s contributions to his or her savings account are insufficient, but the employer has contributed to the solidarity fund for 12 months over the past 34 months, the worker is eligible for payments from the fund. Only formal employment contracts that were entered after 2002 became part of the new unemployment system. Workers hired before 2002 can, on a voluntary basis, become part of the new system. In the case of not opting in, the workers with a contract before 2002 will be supported by the system ‘Subsidio de Cesantía’, which supports the unemployed with fixed cash amounts up to one year. This system is entirely financed by the Instituto de Previsión Social and Cajas de Compensación. LABOR MARKET ACTIVATION OF SD BENEFICIARIES FROM INTERNATIONAL PERSPECTIVE Integration of activation measures and unemployment insurance systems is crucial to support reemployment and to contain moral hazard. Activation is based on the multiple obligations principle—a principle that increases the responsibilities of the unemployed by requiring active job search combined with provision of support through cash benefits and support by the public employment service (PES). The key activation elements of the multiple obligations principle are (a) individualized action planning, which ensures that profiling and treatment of the unemployed are based on personal needs; (b) benefit design, which is incentive compatible with the activation objectives; (c) well-organized coordination between services and benefits; and (e) monitoring and evaluation of the measures (Figure 11) (Almeida et al. 2012). Figure 11: Activation Elements of Unemployment Benefits 33 International examples suggest three typical responsibilities for UI beneficiaries: (a) availability to participate in job interviews proposed by the PES, (b) autonomous job search, (c) participation in ALMPs when required. Evidence shows that well-designed policies can make a difference when it comes to transitioning out of unemployment (EC 2012). In some countries such as Republic of Korea and Sweden, job search requirements are set on a general basis, that is, it is required to apply for a set number of jobs on a monthly basis. In Germany and Portugal, however, the job search requirements are tailored on an individual level and documented in an individual job search plan. Those obligations might be paused during participation in ALMPs (for example, Republic of Korea or Portugal). Monitoring of the job search mainly happens through the written proof of the job search activities, such as through applications and/ or job interview participation (see Table 10). Comparted to the benchmarks, Brazil’s regulatory framework of recipients’ obligations is less stringent. When compared to other countries, SD has a weaker regulatory framework in terms of the co-responsibilities during benefit receipt on job search and job acceptance. In addition, implementation of the activation protocols is weak due to severe capacity constraints in the intermediation system. This is in good part also a function of how little Brazil is spending on services that support unemployed individuals to find jobs, as discussed earlier, and also of its regulatory framework. Similarly, in many countries noncompliance with activation measures results in the suspension of the benefit. Most countries differentiate between the refusal of a suitable job offer and the participation in ALMPs. On average, the sanctions are higher for the refusal of a job offer. In general, sanctions are executed through the suspension of the benefit. In countries such as Republic of Korea, Germany, and Sweden, sanctions increase with repeated misbehavior, whereas in Chile and Portugal the benefit is lost after the first refusal (see Table 11). Yet, countries do allow the refusal of job offers under certain circumstances. Valid reasons might be the proposed: (a) job does not fit the qualifications, (b) job pays a lower wage, or (c) the job requires an increase in commuting time. Further, the health and physical ability of the jobseeker is considered in most countries (see Table 12). In contrast, the administrative data in the National Employment System (Sistema Nacional de Emprego, SINE) indicates that job search obligations for SD beneficiaries are not enforced. Past World Bank analysis of SINE administrative data of 2016 shows that only 1 percent of the SD beneficiaries have a probability of being referred to a job interview on the day of registration. This contrasts with 10 percent of those who register in SINE but are not receiving SD. These low numbers are a first striking indication of the overall weakness of the SINE system in finding and offering enough vacancies for its clients. Second, the analysis found that in the year following registration, only 5 percent of the SD beneficiaries accepted to participate in a job interview, compared to 15 percent 34 of those without SD. The last striking fact is that those in SD who take part in a job interview do so only when benefits are close to expiration (Figure 12). This means that even when vacancies are available, the SINE offices have weak enforcement ability of job search participation until the time the beneficiary has a personal incentive to do so. Last, it is important to note that the matching rate of those who attend interviews has remained low, at around 24 percent in 2016, again showing the extent of the challenges faced by SINE as a whole. Figure 12: Hazard Rate of Participating in a Job Interview Offered by SINE since the Day of Registration in SINE, for Non-beneficiaries of SD (Left) and Beneficiaries of SD (Right) Source: World Bank team’s elaboration, based on Brazil Intermediacao Mao de Obra administrative data 2015–2016. A body of empirical literature confirms that this lack of monitoring results in abuse of unemployment insurance, and higher turnover in the labor market. Studies estimated that eligibility to the UI increases inflows to unemployment by 12 percent, and that the probability of reemployment peaks after the exhaustion of the benefit. Similarly, a significant number of interruptions of formal employment contracts happen at the time when the right to the UI benefit is matured; the same individuals return to employment when the benefit is exhausted (for example, Carvalho 2018; Gerard 2016; see Table 3 above). Table 10: International Benchmarking Job Search Requirements Country Availability for Work during Frequency of Job Search Documentation of Job Participation in ALMPs Monitoring Search Activities Chile Available for a job offer by  Must register at No requirements the employment service BNE and complete their full (Bolsa Nacional de Empleo, curriculum after 96 hours of (BNE)) claiming the benefit. 35 Country Availability for Work during Frequency of Job Search Documentation of Job Participation in ALMPs Monitoring Search Activities  No independent job search activities required. Germany Must be available to accept  Renewable of Must submit evidence of a job offer when integration agreements job search actions and the participation in a training every 6 months results of these activities Must continue with  For young List of applications, independent job search unemployed: renewable including date of while participating in a every 3 months application, name of training employer, and stage of application process Republic No demands on job  Requirement of List of all employers of Korea availability while reporting once every 4 contacted participating in ALMPs weeks to the job center. Should continue job search  Confirmation of job during participation search activities on the standard job seeking activity form Portugal No requirements on job  Individual and May need proof of availability during the collective information independent job search, participation in ALMPs sessions applications, attending job  During session interviews, and unemployed individuals correspondence with should provide evidence of employers job search activities.  Planning of actions is done quarterly (at least once per semester)  Unemployed individuals must participate in job search sessions. Sweden Must be available for work  Once a month Activity report to the PES, and actively searching for which states applications work during the and other activities related participation in ALMPs to the job search. Source: Immervoll and Knotz 2018, based on 2017 rules. Table 11: International Benchmarking Suspension of Unemployment Benefit Country Sanction Voluntary Refusal of Job Offers Refusal of ALMP Unemployment Chile No sanctions Benefits are suspended at Lose the remaining benefit the first refusal entitlement 36 Country Sanction Voluntary Refusal of Job Offers Refusal of ALMP Unemployment Germany No payment for 12 1st refusal: suspension of 1st refusal: suspension of 3 weeks. 3 weeks from benefit weeks from benefit 2nd refusal: suspension of 2nd refusal: suspension of 6 Period of entitlement to 6 weeks from benefit weeks from benefit unemployment benefit Subsequent refusals: 12 Subsequent refusals: 12 weeks will be cut by the weeks from benefit from benefit suspension time, at least by a quarter of the period of entitlement. Republic No unemployment 1st refusal: suspension of 1st refusal: suspension of 4 of Korea benefits 2 weeks from benefit weeks from benefit Subsequent refusals: Subsequent refusals: suspension suspension until job until job seeker attends seeker takes job Portugal Can register as seeking Suspension of benefit Suspension of benefit after 1st employment but cannot after 1st refusal refusal receive unemployment benefits Sweden No payment for 9 weeks 1st refusal: 5 days 1st refusal: 5 days 2nd refusal: 10 days 2nd refusal: 1 days 3rd refusal: 45 days 3rd refusal: 5 days 4th refusal: suspension of 4th refusal: 10 days benefit 5th refusal suspension of benefit Source: Immervoll and Knotz 2018, based on 2017 rules. Table 12: International Benchmarking Reasons to Refuse Job Offers Country Requirement Requirement to Accept Lower Requirement to Other Valid to Accept Wages Commute and Reasons for Jobs in Relocate Refusing Different Job Offers Occupations Chile Not allowed May refuse if the wage offered is The form the Illness to limit less than 50% of the last monthly unemployed worker availability wage earned. fills in at the BNE includes his/her availability/willingness to work in other parts of the country and commuting costs. Can reject an offer if this implies moving to another region. Germany Obliged to First 3 months: a job offer is not Commuting time may Health, take jobs deemed acceptable if the wage not exceed 2.5 hours. moral, or suggested by difference is more than 20% religious the federal Month 4–6: more than 30% reasons 37 Country Requirement Requirement to Accept Lower Requirement to Other Valid to Accept Wages Commute and Reasons for Jobs in Relocate Refusing Different Job Offers Occupations employment After 6 months: job offer is agency. The considered unacceptable only if, placement is taking account of job-related guided by expenses, the net income earned knowledge, in the position offered is lower skills, and than the amount of abilities of the unemployment benefit. unemployed. Republic May refuse Declination justifiable if the wage Refusal acceptable if Health, of Korea job if skills do offered is more than 20% less workplace does not mental, and not fit the than the usual wage level of the provide physical offered job. same type of work or for the accommodation series abilities same level of skills in the same and commuting is region. difficult. Suspension of benefit if recipient refuses the job offer from employment service. Portugal Must accept The limits regarding the wage Claimants are allowed Illness, offer that fits level are the following: to reject jobs when disability physical and  During the first 12  Time of educational months of unemployment commute exceeds 25% qualifications allowance, the claimants are of working hours, of the allowed to reject jobs that carry  Cost to unemployed. wages (gross wage) below the commute should not unemployment allowance + 10%. be more than 10% of  After a year of the gross monthly pay unemployment allowance, the or the travel expenses claimants are allowed to reject of the previous job. jobs that carry wages (gross wage) below the unemployment allowance. Sweden Claimants are allowed to refuse Claimants are allowed Illness, job if to reject jobs when family  The wage is less than  Time of reasons, if 90% of the daily unemployment commute exceeds 1.5 the benefit. hours per day and claimant  The job offer has been requires moving to promised a another place. job Source: Immervoll and Knotz 2018, based on 2017 rules. Besides sanctions, some countries also provide incentives to accelerate re-employment, either for employers or to UI recipients. Incentives toward reemployment can either be targeted toward the employee (for example, bonuses or financial support) or the 38 employer (for example, wage subsidies when hiring UI beneficiaries). These kinds of back- to-work benefits work as a complement to ALMPs in making it more attractive to accept a job offer (EC 2015). In Portugal, to incentivize the unemployed to accept a job offer that is paying less than the current amount of unemployment benefit received, in-work benefits are given to the unemployed who accept a job offer (see Box 3). In the Republic of Korea, to incentivize fast reemployment, a bonus is paid to those that find a job within 30 days after a job separation. On the employer’s side, many countries chose to subsidize the employer who hires a UI beneficiary. In Sweden and Germany, for instance, the hiring of long-term unemployed individuals is subsidized (see Table 12). The advantage of introducing a reemployment bonus is that it is less costly than a wage subsidy. However, while reemployment bonuses shorten the unemployment spells (Ahn 2018), little is known about its long-term effect on employment. Table 13: International Benchmarking Incentives toward Reemployment Country Policies toward Fast Reemployment Germany Wage subsidy paid to employer if he/she hires long-term unemployed  Eligibility: employers who hire difficult-to-place or long-term unemployed individuals  Benefit: wage subsidies and coaching and training subsidies  Payout: diminishing subsidy over 5 years (1st and 2nd year 100%, 3rd year 90%, 4th year 80%, and 5th year 70%) Republic Reemployment bonus paid to the employee of Korea  Eligibility: those who find a job 30 days (one month) before exhausting their benefits  Benefit: half of the residual UI benefit if 55 years or younger; two-thirds of the residual UI benefit if older than 55 years  Payout: lump-sum payment Sweden Special reemployment incentives  Benefit: wage subsidy to the employer - covers 85% of the wage cost up to a ceiling (EUR 100 per working day)  Target group: must be registered in the PES, have enrolled in the Job and Development Guarantee Program for at least 6 months, or have participated in the Youth Guarantee Program for at least 15 months  Payout: 24 months Wage subsidy paid to employer if he/she hires long-term unemployed  Eligibility: ages 21–55 years/older than 55: 6 months of unemployment; ages 27–55 years: 1 year of unemployment (registration requirements)  Benefit: subsidy of twice the employer social contributions  Payout: proportional to unemployment spell; maximum 5 years (10 years for over 55 years old) Box 3: Incentives toward Reemployment - The Case of Portugal In Portugal, the programs ‘Incentivo à Aceitação de Ofertas de Emprego’ and ‘Estímulo Emprego’ aim to incentivize fast reemployment. 39 Incentivo à Aceitação de Ofertas de Emprego offers financial support to unemployed individuals who accept a job offer paying less than the current amount of the unemployment benefit received. Unemployed individuals who have been receiving unemployment benefits for at least three months are eligible. The benefit—which has a maximum duration of 12 months—adopts a decreasing generosity structure. In the first six months the unemployed individual receives 50 percent of the unemployment benefit (or a maximum of EUR 500) followed by 25 percent of the unemployment benefit in the following six months. Estímulo Emprego is a subsidy to employers who hire unemployment benefit recipients for a minimum period of 12 months and offer vocational training to the new hire. The eligibility criteria only accept unemployed individuals who have been jobless for at least six months. With the aim to especially focus on the youth and elderly, the eligibility threshold decreases to two months of unemployment for people under 29 years or over 45 years. The benefit generosity depends on the employer’s regional location, the contract duration, and the circumstances of the unemployed. On average, the benefit structure is as follows: (a) Nine times the value of Social Support Index in the case of permanent employment contracts (b) Three times the value of Social Support Index in the case of fixed-term contracts (c) Additional ‘conversion bonus’ if a fixed-term contract is changed to a permanent employment contract (d) 10 percent more if the target audience is ‘hard-to-reach’ groups or jobs are located in an economically disadvantaged region. V. POLICY OPTIONS The result of the analysis of the unemployment protection system of Brazil can be summarized around three main challenges:  Most of those who experience unemployment do not qualify for SD, because of their modality of work or due to low contribution history. This can partly be explained by the eligibility requirements. Only formal dependent wage workers who were dismissed without ‘justa causa’ have access to SD and to make a first claim requires a contribution history of at least 12 months within the last 18 months. This already excludes a lot of workers who lack work experience or has a segmented contribution history in the formal sector. The criteria also leave self-employed and informal workers without protection. In combination with no coverage, these workers also face a higher labor market income volatility. 40  SD has a high replacement rate, but a relatively short payout period. This makes it more suitable for mitigating short-term unemployment. Relative to international comparators, SD’s replacement rate of 80 percent is on the higher end. UI programs’ face a trade-off between ensuring coverage against a broad set of risks and remaining solvent as well as incentive compatible. In Brazil’s case, the beneficiary of SD is granted three to five months of payments. This is considered low compared to other OECD countries where the payout period can extend for a year or longer. Even though the benefit is generous in the early stages, it is inadequate for long-term unemployment.  The windfall from three unemployment programs at the time of dismissal combined with the little enforcement of reemployment obligations result in perverse incentives to seek job separations and remain in the benefit until its exhaustion. The payout is, in fact, comparable to the full-time income of those working at the MW. This substantial replacement rate, that is not declining over time, may incentivize workers to frequently seek job separation. Some countries, such as Chile or Ecuador, rely on decreasing the replacement rate over the payout period to facilitate the re- integration of the unemployed in the labor market. In addition to this, compared to international best practices, Brazil’s spending on reemployment services is low, and the UI system lacks job search requirements tied to the benefit and is poorly designed to incentivize reemployment. Because workers face no job search conditionalities that can curb moral hazard, they exhaust the benefit. Addressing these challenges will likely require a combination of regulatory reforms, and the reinvestment of fiscal savings to develop modern labor market support systems. This note places particular emphasis on recommendations for the first piece, unemployment protection reforms, but also discusses in less detail complementary policies such as extending protection to a larger share of the uncovered labor force (see Figure 13). The equally critical aspects of how to improve labor intermediation and employment services is not treated here, but these would certainly be a priority for investment of savings accruing from the reform of cash benefits. A positive increase of SINE financing (from 30 to 200 million BRL per year) is planned for 2021-2024 (Ministerio da Economia 2020), in conjunction with the implementation of the new performance- based financing model of SINE offices and an ongoing upgrade of the matching and profiling system of the unemployed (IPEA 2020). 41 Figure 13: An Overview of Policy Options 1 2 3 Legislative changes to SD Systemic reform of the three Complementary policies to extend unemployment protection protection and services to the programs excluded unemployed Designate FGTS as a first- Extension of instruments to Access payer, followed by SD, to formal self-employed to conditions maximize protection and manage income volatility increase equity Activation contracts of SD Benefit Dedicate FAT entirely to beneficiaries with generosity and labor policy, and align FGTS performance-based duration return to market rates financing Policies to widen the access Financing: Link to CLT contracts and its Co-responsibilities contributions and benefits protections for the more directly currently informal workers LEGISLATIVE CHANGES TO SD BENEFIT Parametric reforms to SD could start addressing some of the incentive issues of the benefit and enhancement of coverage, without requiring operational changes.  The contribution period to qualify for the benefit the first time could be reduced to facilitate access for workers with shorter work spells (who are overrepresented among the poor). At the same time, the contribution period for the second and subsequent claim would increase, consistent with the principles of insurance design where more frequent uses should be discouraged (currently the second and third claim require less contributions than the first).  The average replacement rate of pre-unemployment wages could match the ranges observed in international benchmarks. With the savings obtained from the measures above, the duration of the SD benefit could be moderately extended beyond the current five months maximum, particularly for workers who have contributed longer periods since their last utilization of the benefit. This could be designed in a fiscally neutral way and provide greater financial support for more vulnerable workers with longer unemployment spells.  Also, in some countries the benefit assumes a declining generosity after a certain number of months, to continue financing minimum subsistence needs for the long- term unemployed. In Brazil, this process could be coordinated with social assistance 42 (Bolsa Familia) deliberately, to ensure families remain above poverty line when benefits fall or expire.  The SD legislation could be aligned with international experience also in terms of the exceptions to accept a job offer. Much of the activation system depends on the availability of intermediation services and listed vacancies. In addition to improving services, however, the legislation is also critical to ensure that beneficiaries are properly incentivized to collaborate with the activation process. Compared to all other benchmarks, the SD law would benefit from a stronger mandate for the beneficiary to conduct and produce evidence of autonomous job search, in addition to a streamlining of the reasons that allow beneficiaries to refuse job offers.  Top-ups could be provided to workers willing to immediately engage in monitored job search assistance or other ALMPs. The last SD reform in 2015 instituted the obligation to enroll in professional qualification courses and the possibility of suspension of the benefit in case of job refusal. The measures did not yield the expected results, and a study of this experience may also highlight additional areas that require strengthening. Instead of mandating participation, incentives could be provided to workers willing to immediately engage in monitored job search assistance or ALMPs, which would serve as pilots for more comprehensive operational reforms (see Box 3 above for examples). MORE SYSTEMIC REFORM OF THE CURRENT PROGRAMS A more comprehensive reform would rely on FGTS as a primary source of income support after dismissal, and phase in SD after its exhaustion. Countries concerned about informal labor markets have used an equivalent program to FGTS as the first line of financing for unemployment claims (See Box 2 above for the case of Chile). In this way, workers draw from their individual savings, with a monthly maximum. Only once FGTS is exhausted workers would receive funds from the risk pool (SD in Brazil). A key result of this reform is that in the event of unemployment, workers would receive less than they earned while working (fundamental to increase equity), but the total income protection could cover longer payout periods for those in need. Figure 14 provides an example of how such sequencing would look like for a worker with a salary of 1.5 MW. As UISAs may not offer adequate income protection for workers with intermittent working lives in the formal labor market, the SD (risk pooling) element remains important. This proposal has been fully developed and simulated based on Brazilian 2016 data (World Bank 2018) and found that unemployment duration would slightly fall, and the fiscal balance of FAT improve. This proposal would require changes to Laws 7.998/1990 and 8.036/1990 and potentially also constitutional amendments. 43 Figure 14: Example of Unemployment Benefit Payment through Coordinated Financing of FGTS, Multa and SD: Unemployment Benefit by month - pre unemployment wage: 1.5 MWs; job tenure: 24 months 1,400 100% Unemployment Benefit in BRL 1,200 80% Replacement Rate 1,000 80% 80% 80% 80% 75% 800 70% 60% 65% 60% 600 55% 40% 400 20% 200 0 0% 1 2 3 4 5 6 7 8 9 Months of Unemployment FGTS + Multa Financing Public Financing Benefit Replacement Rate The link between SD financing and benefits could be strengthened. Brazil could align to the standard practice of including capped contributions to UI in the workers’ payroll taxes, instead of using a turnover tax. In this way, firms with higher labor intensity would bear a more proportional share of the program’s cost. This shift does not need to equate to higher labor costs, because FGTS’ current 8 percent contribution rate is significant, and part of it could be used finance the FAT. Alternatively, the current Multas would become part of the revenue of the FAT, in this way compensate the fund for increasing rotation of employees. Brazil could also consider the example of some countries that modulate employers’ contribution on experience-based risks levels. Firms that generate more unemployment end up paying more of the total UI costs: this may help reduce the high levels of labor market turnover seen in Brazil, by reducing employers’ incentives to replace their workers frequently and disincentivize collusion schemes. This provision already exists in the SD legislation, but it was never applied. Brazil already has a similar provision in place for the Seguro Accidente Trabalho, part of the payroll taxes for wage employees, with sectors that experience higher work-related accidents paying higher contributions. The large funds generated by firm’s taxes and contributions for Brazil’s unemployment programs, FAT and FGTS, could also be dedicated to serve the modern needs of workers, and provide a competitive rate of return. Currently, the funds are also subsidizing other public policies not related to the labor market, for historical reasons. The surplus generated by FGTS and the FAT could be used to revamp modern ALMPs that prepare the workforce for the demands of the labor market, as well as for better digital and face-to- face labor intermediation service. With stronger ALMPs in place, curbing moral hazard would be more of a ‘win-win’ for all stakeholders, and a self-reinforcing virtuous cycle would foster labor productivity, which is the source of long-term economic growth. A 44 remuneration of savings in FGTS at market rates would also reduce pressure to seek frequent unemployment spells to cash in. The GoB has already taken several steps to reduce the total size of FGTS in 2019–2020, by allowing ‘birthday withdrawals’ and emergency withdrawals during the COVID-19 crisis. The FAT on the other hand continues to be subsidized. COMPLEMENTARY POLICIES TO EXTEND PROTECTION AND SERVICES TO THE EXCLUDED UNEMPLOYED Complementary policies are equally important to broaden access to risk management tools for the excluded populations. Three avenues are proposed in this note: (i) programs to increase protection of the formal self-employed, in its various legal forms (MEI, PJs, Trabalhadores Autonomos), which are a rising share of the formal labor force; (ii) complementary programs and reforms that make access to formal labor contracts more affordable, and thus more inclusive; and (iii) human capital enhancing interventions to increase income and employability. With the appropriate design considerations, Brazil could roll out protection programs for the growing number of self-employed, especially through incentivized savings schemes. The growing number of self-employed is part of the trend of rising diversification of formal work, with shorter employment durations and more dynamic labor relations. The main challenge of designing such schemes is that the self-employed are subject to frequent and difficult-to-observe revenue fluctuations, rather than employment spells. A fully developed proposal to extend unemployment protection to this diverse group is beyond the scope of this report, but some international references point at promising avenues. An incentive-compatible design consists in the development of savings schemes which can serve as an additional layer of protection to smoothen consumption for non- dependent workers. Pragmatically, Brazil could achieve this by opening access to FGTS to the self-employed, if properly incentivized by increasing remuneration of the fund to market rates. Workers could access it only in periods of demonstrated low revenue collection from their registered business. There is evidence that financial education has a positive impact on long-term saving behavior and that it overall has a positive effect on household financial well-being (Wagner & Walstad 2019). Thus, financial education would serve as an important complement to such saving schemes. Not only could it increase savings among self-employed, it might also have a positive impact on how they manage their income in case of dismissal. Saving schemes targeting the self-employed exist in a few countries, including Colombia, and they are usually bundled with incentivized long- term savings for retirement (Bosch et al., 2019). A dozen countries in the OECD also have unemployment insurance windows for the formal self-employed; these schemes would require an activation system of the unemployed to be viable, since they are open to adverse selection risks (Box 4). 45 Box 4: UI schemes for the formal self-employed in OECD A handful of countries in the OECD have introduced UI schemes for the formal self- employed (see Table 14). In some countries, this is compulsory and in others it is optional. UI schemes that are targeting self-employed tend to have stricter eligibility criteria than the UI for formal dependent wage workers, including having to prove bankruptcy or involuntary business closure. However, it is important to note that simply expanding formal arrangements is unlikely to work in middle-income countries such as Brazil where informality is high. Table 14: Country classification of the unemployment insurance schemes (OECD countries) Source: Based on Spasova et al 2017 Wage subsidies and regulatory simplifications can make the CLT contract more attainable for groups of the labor force that are now below the productivity level set by the MW. A targeted wage subsidy to reduce employers’ costs of hiring low-skilled and low experienced unemployed can fill an important gap in the current set of policies. Several countries in Europe (e.g. Germany, Austria and Hungary) have extended wage- subsidies to unemployed without contributory history and could serve as good examples. World Bank (2020) explores in full the international evidence on their effectiveness as well as limitations. In addition, regulatory reforms and simplifications of labor regulations can reduce the implicit costs and uncertainties of using labor contracts and increase formal hiring without fiscal costs. This was recently demonstrated through experimental evidence in South Africa, which shares with Brazil a complex labor code (Bertrand and Crepon, 2020). 46 When properly designed and implemented, human capital enhancing interventions yield, in the medium-term, lasting effects on workers’ employment and earnings. Meta- analyses in Kluve et al (2018) show that human capital enhancing interventions, in the medium-term, have longer lasting effects on workers’ employment and earnings than other programs. In Brazil, rolling this out this in scale may require a re-orientation of at least part of training institutions (Sistema-S), or the private sector, to work in synergy with unemployment benefits, including through performance-based financing (See Clarke et al 2021 for a review). “Activation contracts” of SD beneficiaries with performance-based financing could be an option to implement human capital interventions, with reduced risk for the state. These contracts are coordinated with unemployment benefits and pay for specific milestones (e.g. job placement and continued employment). As in most countries, introducing changes to labor market programs is politically complex agenda. An improvement in labor market services and programs for a broader share of the workforce could generate the needed public support for a reform of the cash benefits treated in this note. The COVID-19 crisis laid bare the limits of the current organization of labor market expenditures. This time may be the opportunity to renew the social contract that underpins the status quo. 47 REFERENCES Almeida, R., J. Arbelaez, M. Honorati, A. Kuddo, T. Lohmann, M. Ovadiya, L. Pop, M. Puerta, and M. Weber. 2012. “Improving Access to Jobs and Earnings Opportunities: The Role of Activation and Graduation Policies in Developing Countries.” Social Protection Discussion Papers and Notes 67610, Washington, DC: World Bank. Almeida and Packard. 2018. Skills and Jobs in Brazil: An Agenda for Youth. International Development in Focus. World Bank. https://doi.org/10.1596/978-1-4648-1293-4. Agencia Brasil. 2020. Cerca de 25% dos desempregados procurma de emprego há mais de dois anos: https://agenciabrasil.ebc.com.br/economia/noticia/2020-02/cerca-de-25-dos- desempregados-procuram-emprego-ha-mais-de-dois-anos. Ahn, T. 2018. “Assessing the Effects of Reemployment Bonuses on Job Search: A Regression Discontinuity Approach.” Journal of Public Economics 165: 82–100. Avlijaš, S. 2019. “The Dynamism of the New Economy: Non-standard Employment and Access to Social Security in EU-28.” LEQS Paper, volume 141. Barros, Ricardo P., Carlos Henrique Corseuil, and Miguel Foguel. 2000. “ Os incentivos adversos e a focalização dos programas de proteção ao trabalhador no Brasil.” Planejamento e Políticas Públicas 22: 3–45.Carvalho, C. C., R. Corbi, and R. Narita. 2018. “Unintended Consequences of Unemployment Insurance: Evidence from Stricter Eligibility Criteria in Brazil.” Economics Letters 162: 157–161. Bosch, Mariano; Caballero, Gustavo; Cofré, Fabián; González, Stephanie; Hand, Anne; Keller, Lukas; Silva Porto, María Teresa. 2019. How to Promote Retirement Savings for Low-Income and Independent Workers: The Cases of Chile, Colombia, Mexico and Peru. Inter-American Development Bank, Washington, DC. Gerard, F., and J. Naritomi. 2019. “Job Displacement Insurance and (the Lack of) Consumption- smoothing.” Working Paper no. w25749, National Bureau of Economic Research. Cabrero, G. R., A. Arriba González de Durana, V. Marbán Gallego, J. Montserrat Codorniu, and F. J. Moreno Fuentes. 2017. ESPN Thematic Report on Access to Social Protection of People Working as Self-employed or on Non-standard Contracts: Spain. European Commission, Brussels. Cereda, Fabio; Rubiao, Rafael M.; Sousa, Liliana D.. 2020. COVID-19, Labor Market Shocks, Poverty in Brazil: A Microsimulation Analysis. World Bank, Washington, DC. World Bank. Chahad, José Paulo Zeetano, and Reynaldo Fernandes. 2002. “Unemployment Insurance and Transitions in the Labor Market: An Evaluation of the Brazilian Program .” Brazilian Review of Econometrics, Rio de Janeiro 22 (2): 239–274. ISSN 1980-2447. Clarke, M. Sharma, M. & Bhattacharjee, P. 2021. Review of the Evidence on Short-Term Education and Skills Training Programs for Out-of-School Youth with a Focus on the Use of Incentives, Social Protection & Jobs: Discussion Paper No. 2013 48 Chłoń-Domińczak, A., A. Sowa, and I. Topińska. 2017. ESPN Thematic Report on Access to Social Protection of People Working as Self-employed or on Non-standard Contracts: Poland. European Commission, Brussels. Doornik, B. F. N. V., D. Schoenherr, and J. Skrastins. 2018. Unemployment Insurance, Strategic Unemployment, and Firm-worker Collusion. Strategic Unemployment and Firm-Worker Collusion. Eurofound. 2017, Exploring Self-employment in the European Union. Publications Office of the European Union, Luxembourg. EC (European Commission). 2012. Employment and Social Developments in Europe 2012 . Brussels. EC. 2015. Social Protection Committee on the Social Situation in the European Union: Annual Report 2014: Social Europe, Aiming for Inclusive Growth. Brussels. EC. 2017. Access to Social Protection for People Working on Non-standard Contracts and as Self- employed in Europe: A Study of National Policies 2017. Brussels. Ferrer, Ana M., and W. Craig Riddell. Unemployment insurance savings accounts in Latin America: Overview and assessment. 2011. 211-237. Fietz, Katharina. 2021 (forthcoming). Unemployment Insurance in Latin America and the Caribbean: A Comparative Review of Current and Leading Practices. Social Protection and Jobs Global Practice. Washington, DC: World Bank Group. Gentilini, Ugo; Almenfi, Mohamed; Orton, Ian; Dale, Pamela. 2020 . Social Protection and Jobs Responses to COVID-19: A Real-time Review of Country Measures . Washington, DC: World Bank. Gerard, F., and G. Gonzaga. 2016. “Informal Labor and the Efficiency Cost of Social Programs: Evidence from the Brazilian Unemployment Insurance Program.” NBER Working Paper 22608. Gonzaga, G., W. F. Maloney, and A. Mizala. 2003. “Labor Turnover and Labor Legislation in Brazil [with comments].” Economia 4 (1): 165–222. Hijzen, A. 2011. “The Labour Market Effects of Unemployment Compensation in Brazil.” OECD Social, Employment, and Migration Working Papers 119, OECD Publishing, Paris. https://doi.org/10.1787/5kg0prkh67d0-en. Immervoll, H., and C. Knotz. 2018. “How Demanding Are Activation Requirements for Jobseekers?” OECD Social, Employment, and Migration Working Papers 215. ILO (International Labour Organization). 2016. Non-standard Employment around the World: Understanding Challenges, Shaping Prospects. Geneva. ILO. 2018. Innovative Approaches for Ensuring Universal Social Protection for the Future of Work. Geneva. IPEA. 2017. Análise Da Contribuição Do Programa De Intermediação De Mão De Obra Para Os Gastos Com Seguro-desemprego E Para Arrecadação Previdenciária No Período Recente (2015- 2016). DISOC, No. 34. 49 IPEA. 2020. O novo SINE : mudanças do modelo brasileiro de intermediação de mão de obra. Mercado de Trabalho conjuntura e analise. ANO 26. Kangas, O., and L. Kalliomaa-Puha. 2017. ESPN Thematic Report on Access to Social Protection of People Working as Self-employed or on Non-standard Contracts: Finland. European Commission, Brussels. Ministerio da Economia. 2020. Avaliacao Financieira do Fundo de Amparao ao Trabalhador – PLDO 2021. Nota tecnica SEI 12912/2020/ME. Brasilia. Nelson, K., J. Palme, S. Burman, and M. Eneroth. 2017. ESPN Thematic Report on Access to Social Protection of People Working as Self-employed or on Non-standard Contracts: Sweden . European Commission, Brussels. OECD. 2018a. The Future of Social Protection - What Works for Non-standard Workers? Paris: OECD. OECD. 2018b. Policy Brief on the Future of Work. Paris: OECD. Pacolet, J., and L. Op de Beeck. 2017. ESPN Thematic Report on Access to Social Protection of People Working as Self-employed or on Non-standard Contracts. Luxembourg, Brussels. Packard, T., U. Gentilini, M.E. Grosh, P. B. O'Keefe, R.J. Palacios, D. A. Robalino, and I. V. Santos. 2019. Protecting All: Risk Sharing for a Diverse and Diversifying World of Work. Washington, DC: World Bank Group. Silva, J., Almeida, R. and Strokova, V. 2015. Sustaining Employment and Wage Gains in Brazil: A Skills and Jobs Agenda. Directions in Development. Washington, DC: World Bank Group. ISBN 978- 1-4648-0644-5. https://doi.org/10.1596/978-1-4648-0644-5. Silva, J. Sousa, L. Packard, T. and Robertson, R. 2021 (forthcoming). Crises and Labor Markets in Latin America and the Caribbean. Lessons for an Inclusive Recovery from the COVID-19. Pandemic. Washington DC: World Bank Group. SSA (Social Security Agency). 2019. Social Security Programs Throughout the World. https://www.ssa.gov/policy/docs/progdesc/ssptw/index.html. Vodopivec, Milan. Income support for the unemployed: Issues and options. The World Bank, 2004. Vroman, Wayne, and Vera Brusentsev. Unemployment compensation throughout the world: A comparative analysis. WE Upjohn Institute, 2005. Wagner, J. and Walstad, W.B. (2019), The Effects of Financial Education on Short-Term and Long- Term Financial Behaviors. J Consum Aff, 53: 234-259. https://doi.org/10.1111/joca.12210 World Bank, 2020. Brazil - The Contrato Verde Amarelo Wage Subsidy : An Initial Assessment. World Bank, Washington, DC. World Bank. 2018. Federal Spending on Labor Market Programs in Brazil, 2000-2018. Background Chapter for Brazil Expenditure Review. Washington, DC: World Bank. 50 ANNEX 1: ADDITIONAL INFORMATION Figure 15: Evolution of Requests of SD by Education, January 2012–October 2020 1,200 SD claims in thousands 1,000 800 600 400 200 0 Nov-12 Nov-17 Apr-13 Mar-16 Aug-16 Apr-18 Sep-13 Feb-14 Jul-14 Dec-14 Oct-15 Sep-18 Feb-19 Jul-19 Dec-19 Oct-20 Jun-12 Jun-17 Jan-12 May-15 Jan-17 May-20 Analfabeto Fundamental Incompleto Fundamental Completo Ensino Medio Incompletio Ensino Medio completo Superior Incompleto Superior Completo Source: Painel de informacoes do Seguro-Desemprego , Ministry of Economy, accessed December 16, 2020. Figure 16: Benefit level and replacement rate at different wage levels (between 1 and 3 MW) 2,000 120% 1,800 100% 1,600 Benefit amount BRL 1,400 Replacement Rate 80% 1,200 1,000 60% 800 40% 600 400 20% 200 0 0% 1 2 3 Pre-unemployment wage in MWs Benefit in BRL Replacement Rate Source: Authors’ calculation based on policy rules of 2020, Labor Secretary, Ministry of Economy, Brazil. 51 Table 15: Spending on labor market programs by target group and program type Passive support Active support Share Program Expenditure (2015 R$) Program Expenditure (2015 R$) Total of total Garantia Safra 380,604,551.00 Pronatec 2,439,323,144.00 Bolsa Verde 92,830,525.00 Rural technical assistance - ATER 93,951,794.00 Seguro-Desemprego ao 3,192,086.00 PAA (MDA) 20,596,651.00 Trabalhador Resgatado de Benefiting Condição Análoga à de mostly Escravo* 3,051,370,211.00 1.8% informal Economia Solidária 13,513,183.00 workers Support to productive inclusion 6,876,865.00 of family farmers and women rural workers Brasil Quilombola 481,412.00 Mais Aprendiz (youth - apprenticeship) 1 101,206,127,941.00 Mandated non-wage benefits2 14,820,000,000.00 FGTS Seguro-Desemprego 37,363,906,241.00 Abono Salarial 9,073,572,355.00 Seguro-Defeso* 2,310,619,343.00 Sistema S3 7,092,393,929.76 Seguro-Desemprego ao 33,357,798.00 Salario Familia** 2,630,000,000.00 Trabalhador Doméstico* Benefitting Mais Emprego-SINE 69,825,144.00 mostly or only Projovem-courses+financial 61,129,748.00 167,575,553,165.00 98.2% formal assistance workers Fomento/microgrants 6,831,066.00 Technical assistance /microcredit 183,529.00 to microentrepreneurs- PNMPO/Crescer/Agroamigo Total 141,390,638,485.00 29,236,284,891.00 170,626,923,376.00 Share of total 82.9% 17.1% Source: World Bank 2018. Based on BOOST national budget data for 2015. 52