Unemployment Insurance with Informal Labor Markets: Evidence from Brazil

Transcription

1 Unemployment Insurance with Informal Labor Markets: Evidence from Brazil Bernardus Ferdinandus Nazar Van Doornik David Schoenherr Janis Skrastins January 15, 2017 This Working Paper should not be reported as representing the views ot the Banco Central do Brasil. The views expressed in the paper are those of the authors and not necessarily reflect those of the Banco Central do Brasil. Abstract Using the universe of formal labor contracts in Brazil and an unexpected unemployment insurance (UI) reform, this paper examines how informal labor markets affect workers incentives arising from UI benefits. Exploiting a sharp discontinuity in the reform s effect, we find that eligibility for UI benefits increases the flow from formal employment into formal unemployment by twelve percent. This effect is mainly driven by workers in labor markets with a high degree of informality. Collusion between workers and their employers accounts for at least 20 percent of strategic formal unemployment. Firms appear to hire workers informally while they are eligible for UI benefits and rehire them formally when benefits are exhausted. Additionally, relaxing eligibility criteria for UI benefits leads to a higher formal labor supply in the presence of informal labor markets. Consequently, wages for formal jobs are lower for newly hired workers in more informal labor markets and relative to informal wages within the same local industry. These results suggest that incentive effects of UI benefits are stronger in the presence of informal labor markets and that UI benefits lead to higher formalization of labor. JEL Codes: J21, J22, J46, K31. Keywords: informality, labor supply, law and economics, unemployment insurance, collusion. We thank Henry Faber, Dimas Fazio, Camille Landais and seminar participants at the Banco Central do Brasil, Bank of Latvia, SSE Riga, and Washington University in St. Louis for many helpful comments and suggestions. Central Bank of Brazil, bernardus.doornik@bcb.gov.br Princeton University, 26 Prospect Ave, Princeton, NJ 08544, schoenherr@princeton.edu Washington University in St. Louis, One Brookings Drive, St. Louis, MO , jskrastins@wustl.edu

2 1 Introduction Providing social insurance against adverse shocks is one of the major functions of governments in developed countries, accounting for the largest fraction of public spending. 1 One of the most important government insurance programs is unemployment insurance (UI), which provides transfers to individuals who experience negative income shocks from loss of employment. While UI has been present in developed countries for decades, it is rapidly expanding to middle-income and developing countries only in recent years. 2 The existing academic literature has largely focused on the effects of UI on workers incentives in developed countries. 3 In contrast, empirical evidence on the effects of UI in middle-income and developing countries is scarce. 4 An important feature of these countries is the prevalence of informal labor markets. Since existing studies on UI in developed countries lack this dimension, it is unclear whether their findings can be extrapolated to mid-income and developing countries. The novel contribution of this paper is to bridge this gap and examine the incentive effects of UI in the context of informal labor markets. Informal labor markets provide an additional outside option for workers who consider formal employment. Extending UI benefits makes formal employment more attractive ex ante, as formal employment is a prerequisite to establish eligibility for UI benefits. the other hand, conditional on being eligible for UI benefits, the availability of informal labor markets may incentivize workers to exit formal employment and complement their UI benefits with informal income. The same effect may also reduce workers incentive to seek reemployment when they can work informally instead. Ultimately, informal labor markets generate an additional set of workers whose decision, to work formally or informally, may be affected by the design of UI programs. Thus, workers who would otherwise not be sensitive respond to UI when informal labor markets exists. As a result, informal labor markets may 1 Expenditures on social insurance account for about 60 percent of federal government spending in the U.S. (Gruber 2011). 2 See Holzmann et al. (2011) for data on unemployment insurance around the world. 3 See for example, Solon (1979), Moffitt (1985), Katz and Meyer (1990), Meyer (1990, 1995), Meyer and Mok (2007), Card and Levine (2000), Johnston and Mas (2015), and Landais (2015) for the US, and Card, Chetty, and Weber (2007), Lalive (2008), Schmieder, von Wachter, and Bender (2012, 2016), and Card et al. (2015) for Western Europe. 4 Gasparini, Haimovich, and Olivieri (2009) and Gonzalez-Rozada, Ronconi, and Ruffo (2011) document that transfers to unemployed individuals in Argentina made people less likely to take on formal work; Amarante, Arim, and Dean (2013) show that an extension of unemployment insurance benefits in Uruguay increased unemployment spells with no effects on future wages; Gerard and Gonzaga (2014) report that extending unemployment insurance benefits leads to longer unemployment spells, although they argue that reemployment rates are low in general even in the absence of unemployment benefits. On 1

3 exacerbate the incentive effects of UI benefits. To shed light on this question, we exploit an unanticipated UI reform in Brazil that significantly tightened eligibility criteria for UI benefits. Brazil constitutes an ideal laboratory for several reasons. First, Brazil is a middle-income country where informal labor markets are prevalent. 5 Second, the country is very heterogeneous, providing ample variation in labor market informality across cipalities and industries. This allows us to examine differences in the effects of UI for different degrees of labor market informality. Finally, data availability and quality are excellent even by developed country standards. 6 The nature of the reform provides a sharp discontinuity in the loss of eligibility for UI benefits. 7 Prior to the reform, which came into effect in March 2015, workers with an employment history of at least six consecutive months were eligible for three months of UI benefits. To obtain the same benefits after the reform, workers must demonstrate a longer employment history. Applying for benefits for the first (second) time requires formal employment for 18 (12) months during the last 24 (16) months. 8 Thus, a large fraction of workers with a tenure of six months does not qualify for UI benefits after the reform. In contrast, workers with a tenure of five months are not eligible for benefits both before and after the reform. This discontinuity motivates our main identification strategy, a differencein-differences methodology, in which we compare changes in the incentives for workers with tenure of five and six months. 9 Importantly, the reform only affects workers eligibility for UI benefits, but does not affect firms contributions to the UI program. This allows us to identify the effects of UI benefits on workers incentives free from changes in firms demand for formal labor. We start our analysis by examining how UI benefits affect workers incentive to exit formal employment. Our findings indicate that the UI eligibility has strong effects on unemployment inflow. Particularly, unemployment inflow relatively drops by 0.43 percentage points for 5 According to the International Labor Organization, 36.8 percent of all workers were employed informally in For comparison, in Europe, informal labor markets account for 17.4 percent of the total labor market during (Hazans 2011) 6 The Ministry of Labor and Employment collects comprehensive and high-quality micro-level employment and wage data for the entire population of formal workers in Brazil. 7 Importantly, the policy change was sudden and unanticipated. The results are not affected by the two months transition period between the announcement and the implementation of the reform (Section 5). 8 To qualify for benefits, workers must not have successfully applied for UI benefits during the previous 16 months both before and after the reform. See Section 2.2 for more details about the reform. 9 Based on a wide range of observable characteristics, workers with a tenure of five and six months are indistinguishable. 2

4 workers who lose eligibility for UI benefits after the reform, which is equivalent to a percent decline in unemployment inflow. It is important to note that this strategic inflow into formal unemployment has not been documented for developed countries. 10 Exploiting cross-sectional variation in labor market informality across industries and cipalities, we assess whether informal labor markets exacerbate the effect of UI benefits on unemployment inflow. We find that the drop in unemployment inflow after the reform is almost exclusively driven by workers in industries and cipalities with large informal labor markets. Specifically, we find that a ten percentage point increase in the share of informal employment in a given industry or cipality corresponds to an about three percent higher inflow into formal unemployment when workers are eligible for benefits. This suggests that informal labor markets induce workers to exit formal employment when they are eligible for UI benefits. Next, we examine whether UI benefits affect reemployment probabilities after layoff. Consistent with the existing literature, we find that dismissed workers return to formal employment more quickly when they lose eligibility for UI benefits. Specifically, for dismissed workers who become ineligible for UI benefits after the reform, the probability to return to formal employment within three months of layoff relatively increases by about three percentage points compared to workers that are ineligible for benefits before and after the reform. Higher inflow into and delayed outflow from formal unemployment could be driven by workers incentives without involvement of their employer. For example, workers may shirk to induce layoff when they become eligible for UI benefits. Alternatively, it is possible that workers and employers collude using UI benefits as a government subsidy to be shared amongst them. Firms can layoff workers when they are eligible for UI benefits, employ them informally while they receive benefits, and hire them back when UI benefits run out. Collusion would imply that workers are more likely to return to their previous employer before the reform when they exploit UI benefits strategically. This is what we observe in the data. Before the reform workers who are laid off just when they become eligible for UI benefits are significantly more likely to be rehired by their previous employer when benefits run out. This firing-rehiring pattern by the same firm is mostly driven by industries and 10 One implication of strategic unemployment inflow is that examining unemployment duration for people around an eligibility threshold is subject to selection concerns that may lead to an overestimation of UI incentive effects. 3

5 cipalities where informal labor markets are prevalent. We find no such pattern after the reform for workers who lose eligibility for benefits. This suggests that collusion between employers and employees is an important channel through which informal labor markets lead to higher inflow into formal unemployment. The previous analysis examines how informal labor markets affect workers incentives conditional on qualifying for UI benefits. UI programs may also affect workers decision to enter formal employment ex ante. An UI regime that facilitates qualification for benefits makes formal employment more attractive ex ante by increasing expected future cashflows from formal employment. This is particularly relevant in the presence of informal labor markets when workers decide between working formally or informally. On examining workers decision to enter formal employment ex ante, we find that formal hiring relatively decreases in more informal labor markets when the likelihood of future UI benefit eligibility is lower. Specifically, a ten percentage points increase in labor market informality corresponds to a drop in new hiring by 0.15 percent of the formally employed workforce in a given industry after the reform. Finally, we examine changes in formal wages following the reform. Our findings suggest that the reduction in formal hiring is driven by a decrease in the supply of formal labor. Formal wages relative increase when more informal employment options are avilable. Specifically, wages for newly hired formal workers relatively increase by 0.15 percent for a ten percent increase in labor market informality. Furthermore, we find that this increase in wages is not driven by changes in economic conditions or productivity in certain industries or cipalities after the reform. Survey data on both formal and informal wages shows that within the same industry and cipality, wages increase for formally employed relative to informally employed workers by more than three percent when qualifying for UI benefits becomes more likely after the reform. Together, these results suggest that informal labor markets have three distinct effects on how UI benefits affect workers incentives. First, conditional on eligibility for UI benefits, informal labor markets lead workers to exit formal employment. Workers collude with their employers and collect government transfers before returning to formal employment. Second, the option to remain employed informally further reduces search intensities to reenter formal employment compared to markets without informal labor markets. Finally, UI benefits induce workers to enter formal rather than informal employment in order to establish eligibility for UI benefits, leading to lower formal wages in equilibrium. 4

6 The combination of effects we observe in the data imposes high hurdles for alternative explanations. Due to the sharp discontinuity of the UI reform s effect around a six months tenure threshold, any alternative story would need to explain a change in behavior for workers with a tenure of six months relative to workers with a tenure of five months that exactly coincides with the month that the reform takes effect. With respect to the cross-sectional differences in the reform s effect depending on the level of labor market informality, any factor that might be correlated with labor market informality and could affect how workers respond to UI benefit eligibility needs to i) be correlated with labor market informality both geographically and across industries, ii) both lead to higher ex-post inflow into formal unemployment and higher ex-ante inflow into formal employment, iii) explain higher collusion between employers and employees to strategically exploit UI benefits, and iv) explain differential effects on employment and wages for formal and informal jobs within the same cipality and within the same industry. We perform several additional robustness tests to further ensure that the results are driven by differences in labor market formality. First, we control for cyclical patterns by examining the year before the reform. None of the effects are present in the previous year, confirming that the results are not driven by cyclical patterns. Second, we confirm that workers do not substitute to other forms of job separation, such as voluntary departures or job changes, after the reform. Third, we show that neither informal industries nor informal cipalities are differentially affected by the recession in Brazil during our sample period. The paper contributes to several strands of literature. Most importantly, the paper documents how informal labor markets interact with the incentive effects of UI benefits. Some recent studies analyze UI programs in middle-income and developing countries (Gasparini, Haimovich, and Olivieri 2009; Gonzalez-Rozada, Ronconi, and Ruffo 2011; Amarante, Arim, and Dean 2013; Gerard and Gonzaga 2014). However, these papers do not directly examine how differences in labor market formality influence the effect of UI programs with the exception of Gerard and Gonzaga (2014), which is the closest to our paper. They examine whether the size of informal labor markets affect the incentive effect of UI benefits on unemployment outflow using fifteen years of unemployment data in Brazil. In contrast, we examine how the presence of informal labor markets changes the effects of UI benefits on ex ante labor supply, wages, and unemployment inflow and outflow. Additionally, we exploit an unexpected reform that affects the eligibility for UI benefits. This reform occurs after the sample period in Gerard and Gonzaga (2014) and allows us to provide a sharper identification of the effects 5

7 of labor market informality on how workers incentives are affected by UI benefits. The paper also provides new insights into the impact of UI benefits on hiring and wages. Most studies in developed countries find no or a slightly negative effects on reemployment wages (Card, Chetty, and Weber 2007; Lalive 2007; van Ours and Vodopivec 2008; Centeno and Novo 2009; Degen and Lalive 2013; Johnston and Mas 2015; Schmieder, von Wachter, and Bender 2016). 11 Since workers incorporate expected future benefit payments in their optimization, they are more likely to enter formal employment when future eligibility for UI benefits is more likely. This positive effect on formal labor supply leads to lower formal wages relative to informal wages ex ante. It is important to consider the relevance of our results beyond the specific context in Brazil. First, our data spans the entire population of formal employees in the private sector in Brazil. Thus, the results are not subject to any selection bias. Second, our findings are obtained during a severe recession in Brazil. Schmieder, von Wachter, and Bender (2012) show that incentive effects from UI are significantly weaker during recessions in Germany. This suggests that the results we find are rather conservative estimates of the effects of UI insurance during non-recessionary period. Additionally, evidence on the effects of UI benefits in recessionary periods are of particular interest, as they provide a fiscal stimulus during recessions and are often extended during downturns (Rothstein 2011; Valletta 2014; Farber and Valletta 2015; Kroft and Notowidigdo 2016). 12 Finally, while informal labor markets are less prevalent in developed countries, non-negligible informal labor markets exit in those countries as well. Hence, we think that the results in this paper are relevant and informative beyond the specific context of this paper. 2 Institutional Background and Data This section provides information about Brazil s UI system and the UI benefits reform implemented in March Nekoei and Weber (2015) find a slightly positive effect on reemployment wages. 12 The magnitudes from our main tests for the full sample are comparable to the magnitudes found in the literature, for example Card, Chetty, and Weber (2007). 6

8 2.1 Unemployment Insurance in Brazil In Brazil, every formal worker is required to hold a working card. It is mandatory for employers to sign workers cards whenever a worker is hired, promoted, or dismissed. This information is reported to the Ministry of Labor. Hiring an employee formally rather than having an informal relationship with the employee involves costs. Formal employees are entitled to a the minimum wage. Payroll taxes amount to twenty percent to finance the public pension system plus eight percent for workers seniority account (FGTS). 13 Other mandatory contributions such as the social integration program (PIS) and contributions to social security funding (COFINS) depend on the industry that the firm operates in. These contributions are paid from net profit and sales. Funding for the UI system stems from these contributions. Laying off a worker is costly. If a firm dismisses a worker without a justified reason, it must pay an additional fifty percent of the total contribution that has accumulated in an employee s FGTS. Laying off workers with a valid legal justification does not involve penalties. However, this is rare (only 3.5 percent of all layoffs) since the hurdle to provide sufficient evidence is high, for instance proving low productivity is hard to prove, and judges tend to rule in favor of employees. UI applies to formally employed workers in the private sector. 14 Benefits are paid for three to five months, depending on workers time in formal employment. 15 In 2015, the monthly payment ranges from 1 to 1.76 minimum salaries, depending on the average pre-layoff wage. 2.2 UI Benefits Reform Eligibility criteria for UI benefits did not undergo a major reform from 1994 to To be eligible for UI benefits prior to March , a worker had to be employed over a consecutive period of at least 6 months prior to layoff, had to be fired without a justified reason, may not earn any other formal labor income, and may not have sucessfully applied for UI benefits during the previous sixteen months. 13 This account can be withdrawn when a worker retires, is fired, or suffers from a serious illness. 14 These are all employees that are subject to consolidation of labor laws (Consolidao das Leis do Trabalho). 15 The unemployment insurance benefit is paid for a minimum of three months and a maximum of five months: three payments if a worker was employed between six and eleven months in the last 36 months, four payments if a worker was employed between 12 and 23 months in the last 36 months, and five payments if a worker was employed for at least 24 months in the last 36 months. 7

9 In December 30, 2014, the parliament passed a provisional measure that tightened eligibility criteria for UI benefits. The new criteria were set to be enforced from March 1, While it was anticipated that UI would be reformed at some point, both the sudden implementation and the content of the new law were fully unexpected. 16 The main driver for the quick implementation and the tightening in eligibility criteria was the desperate attempt on part of the government, to find sources of consolidation for reducing the growing budget deficit. The size and duration of UI benefits was not altered. The eligibility criteria were tightened for all dismissed workers who applied for UI benefits for the first or second time. 17 Both groups had to have longer pre-layoff employment history to be eligible for UI benefits. In particular, workers who applied for the first time required documented employment of at least 18 months in the 24 months prior to layoff. Workers who applied for the second time required 12 months in formal employment during the last 16 months. 18 This measure was turned into law in July 2015 with only minor adjustments. While stricter than the old requirements, these requirements were somewhat laxer as compared to the ones in the provisional measure. Specifically, a first time applicant had to have at least 12 months of employment in the last 18 months. A second time applicant had to have at least 9 months of employment in the last 12 months. 2.3 Data In this paper, we use two large restricted-access matched employee-employer administrative datasets from Brazil, RAIS (Relao Anual de Informaes Sociais) and CAGED (Cadastro Geral de Empregados e Desempregados). While the RAIS database records information on all formally employed workers in a given year, CAGED records the flow of employees, that is information on hirings and layoffs. Ministry of Brazil. Both datasets are maintained by the Labor All formally-registered firms in Brazil are legally required to report annual information on each worker that the firm employs. RAIS includes detailed information on the employer (tax number, sector of activity, 16 Estadao Politica, December 29, 2014, Forca Sindical nega ter sido consultada sobre ajuste em beneficios. 17 Our data does not allow us to identify how many times a worker has applied for UI benefits in the past because the data is truncated. 18 See Figure 1 for a summary of eligibility criteria under the different regimes. 8

10 establishment size, geographical location), the employee (social security number, age, gender, education), and the employment relationship (wage, tenure, type of employment, hiring date, layoff date, reason for layoff, etc.). We use data from RAIS for the period from By the end of 2014, the database covers about 50 million formal employees. The information contained in CAGED is similar to the information reported in RAIS. For example, it includes the hiring and layoff date, the reason for layoff, wage, etc. We use CAGED for 2015 to complement the data from RAIS as RAIS is currently only available up to Both datasets contain a common identifier for each worker, which remains with the worker throughout his or her work history as well as the tax identifier of the worker s employer. Combining both datasets allows us to trace the duration of formal employment for each individual. We exclude all public sector employees, since they do not participate in the UI program. For our main identification strategy, we focus on employees with a consecutive formal work history of five or six months in a given month. Additionally, we exploit information on the location of the firm (cipality), its two digit industry classification (National Classification of Economic Activities), and information on employees occupations (Classificao Brasileira de Ocupaes) for our empirical analysis. Our main empirical specification compares the period after the announcement but before the implementation of the UI reform (January February 2015) and the period after the implementation of the reform (March April 2015). In Table 1, we confirm that workers with a tenure of six months, who are directly affected by the reform, and workers with a tenure of five months, who are not directly affected by the reform, are indistinguishable in terms of observable characteristics before the implementation of the reform in March For a person to be employed for five or six consecutive months during January to April 2015, this person must have been employed before the announcement of the reform. Thus, initial employment decisions are not subject to strategic selection in anticipation of the reform. We find that both groups of workers are virtually identical in terms of age, average salary, gender, university education, the size of the firm they are employed at, and the industries that they are employed in. They do, however, differ in terms of the probability of becoming unemployed in a given month. In particular, a worker with six month tenure is by ten percent more likely to become unemployed and 27 percent more likely to be laid off by the firm without a justified reason. Thus, while workers with tenures of five and six months look identical based on observable characteristics, unemployment inflow 9

11 differs significantly, which will be a focus of our empirical analysis. To exploit cross-sectional variation in labor market informality, we combine the linked employer-employee data from RAIS and CAGED with information on labor market informality from the Brazilian census in The census survey asks whether or not an individual has a job, and whether or not this job is formal. We classify workers as informal if they do not work on a formal contract (i.e., a signed worker s card). 19 The census groups workers in twenty different industry classifications (see Table 2). 66 percent of domestic services employees are working informally. The most formal industry, electricity and gas, has only 5.5 percent of informal workers. In terms of geographic variation in informality, most cipalities fall within the range of 20 to 70 percent of labor market informality (Figure 2). Importantly, informality is not clustered in some parts of Brazil but is prevalent throughout the country (Figure 3). Finally, we also take advantage of the National Household Sample Survey (Pesquisa Nacional por Amostra de Domiclios). This quarterly survey collects information on formal and informal employment and salaries for the working age population in 20 cipalities that are the respective state s capital cipalities. 3 Empirical Strategy This section outlines the empirical strategy to assess how UI benefits affect workers incentives the presence of informal labor markets. 3.1 Labor Market Informality We exploit two sources of variation in labor market informality. First, we exploit crosssectional variation in informality across industries. As documented in Table 2, there is a wide heterogeneity in labor market informality across industries in Brazil, ranging from 5.56 percent in electricity and gas to percent in domestic services. Second, we exploit variation in labor market informality across cipalities in Brazil. Brazil is a very heterogenous country with large geographical variation in labor market formality (see Figures 2 and 3). 19 We verify that the results are robust to alternative definitions of labor market informality provided in the census. 10

12 Naively comparing differences in workers incentives across industries and cipalities is problematic, as workers incentives may differ across industries and cipalities for reasons other than labor market informality. To control for such spurious effects, we exploit a UI benefits reform that generates time-series variation in UI benefits eligibility. The design of the reform induces a sharp discontinuity in changes in UI benefit eligibility. Workers with a tenure of six months lose eligibility for UI benefits after the reform, whereas workers with a tenure of five months are ineligible both before and after the reform (see Section 2.2 for more details). Thus, while the reform equally applies to all industries and cipalities in Brazil, whether a given worker is affected depends on her employment history. The sharp discontinuity in the reform s effect allows us to examine changes in unemployment inflow and outflow for workers just above the eligibility threshold (six months tenure) and workers just below the threshold (five months tenure). Any general differences in workers incentives in industries and cipalities with different levels of labor market formality should apply to workers with tenure of five and six months before and after the reform. However, for workers with a tenure of six months, UI benefits eligibility affects their incentives only before the reform. This allows us to identify how UI benefits affect workers incentives in the presence of informal labor markets, controlling for general differences in workers incentives across industries and cipalities. 3.2 Unemployment Inflow and Outflow We start by examining changes in unemployment inflow after the implementation of the reform for all workers with a tenure of five or six months by estimating: P [u unjust ] it = α + β 1 Eligible it + β 2 Reform t (1) +β 3 Eligible it Reform t + ɛ it where P [u unjust ] it is a dummy variable that takes the value of one if worker i is laid off in month t, and zero otherwise. 20 The dummy variable Eligible it takes the value of one for workers with a tenure of six months, and zero for workers with a tenure of five months. The dummy variable Reform t takes the value of one for the two months after the reform, and zero for the two months before the reform We only consider layoffs that are legally unjustified, as workers only obtain UI benefits in this case. 21 We ensure that the results are not driven by cyclical effects in Section

13 The parameter of interest is β 3, which is informative about the effect of UI benefits on unemployment inflow. The difference in unemployment inflow for workers with tenure of five and six months after the reform, when both groups of workers are not eligible for benefits, provides us with the undistorted difference in unemployment inflow between both groups of workers. The coefficent β 3 compares this undistorted difference to the difference in unemployment inflow between both groups of workers before the reform when workers with a tenure of six months are eligible for UI benefits. Since the reform eliminated eligibility for workers with tenure of six months, a negative value of β 3 implies that UI benefits lead to higher unemployment inflow when workers are eligible for UI benefits, and vice versa. We apply the same identification strategy to estimate the effect of informal labor markets on unemployment outflow by replacing the dependent variable with P [e 3] it, a dummy variable that takes the value of one if worker i is reemployed within three months after being laid off, and zero otherwise. The three months time-period is motivated by the fact that UI benefits are available for at least three months after layoff for eligible workers. Here, a positve value of β 3 implies lower unemployment outflow in the presence of UI benefits, and vice versa. To formally assess how unemployment insurance benefits affect workers incentives in the presence of informal labor markets, we estimate: P [u unjust ] it = α + β 1 Eligible it + β 2 Reform t (2) +β 3 Eligible it Reform t + β 4 Informal +β 5 Eligible it Informal + β 6 Reform t Informal +β 7 Eligible it Reform t Informal + ɛ it where Inf ormal is the share of informal employment in a given industry or cipality, all other variables are defined as before. We can further saturate equations (1) and (2) with month, cipality-month, cipalityindustry-month, and cipality-industry-occupation-month fixed effects to control for location-specific, industry-specific, and even occupation-specific shocks that may affect workers with with tenure of five and six months differentially. Parameter β 7 is informative about the effect of UI benefits in labor markets with high levels of informality relative to labor markets with lower informality. A negative value of β 7 implies that UI benefits have a stronger effect on unemployment inflow when labor market informality is higher, and vice versa. 12

14 3.3 Formal Employment and Wages In addition to unemployment inflow and outflow, conditional on eligibility for UI benefit, the prospects of future eligibility for UI benefits may also affects workers incentives to enter formal employment in the first place. After the reform, qualifying for UI benefits is substantially more challenging. To assess the effect of the reform on workers ex ante incentive to enter formal employment, we compare changes in hiring rates in industries and cipalities with different levels of labor market informality: W orkers Hired t = α + β 1 Reform t + β 2 Informal (3) +β 3 Reform t Informal + ɛ t where W orkers Hired t is defined as the number of workers hired in a given industry (cipality) in month t scaled by the number of workers employed in the respective industry (cipality) in December 2014, the month before the reform was announced. Since the reform was announced in January 2015, and workers incentives to enter formal employment should be affected from the time they are aware of the reform s effects, we define the Reform t dummy as one from January Informal is the share of informal employment in a given industry or cipality. Coefficient β 3 measures the relative change in the number of workers hired in more informal industries (cipalities) after the reform compared to the period before the reform. Finally, to assess whether changes in hiring are driven by changes in the demand for or the supply of formal labor, we examine changes in hiring wages in industries (cipalities) with higher levels of labor market informality relative to changes in hiring wages in industries (cipalities) with lower levels of labor market informality by replacing the dependent variable in equation (3) with the log of the average hiring wage in a given industry (cipality) in month t. Additionally, we use data on wages in informal jobs from the quarterly PNAD survey, to directly compare the effect of UI benefits on wages in formal and informal labor markets: log(wage) it = α + β 1 Reform t + β 2 F ormal Job it (4) +β 3 Reform t F ormal Job it + ɛ it where F ormal Job it takes the value of one if worker i is formally employed in quarter t, and 13

15 zero if worker i is informally employed in quarter t. We can saturate equation (4) to compare changes in formal and informal wages within the same industry (industry-time fixed effects), the same cipality (cipality-time fixed effects), or the same industry in the same cipality (cipality-industry-time fixed effects) at a given point in time. 4 Results This section presents the empirical results. We document how UI benefits affect workers incentives in the presence of informal labor markets. Specifically, we assess how UI benefits affect workers incentives to enter formal employment ex ante and to exit formal employment ex post in labor markets with different levels of informality. 4.1 UI Benefits Reform and Unemployment Inflow Figure 4 depicts the probability of being laid off for workers with different tenure for each month from January to April While there are no significant changes in unemployment probabilities for workers with a tenure between one to five months, for workers with tenure of six to sixteen months the probability of being laid off significantly decreases after the reform. In particular, there is a sharp drop in the probability of being laid off for workers with tenure of six months who lose eligibility for UI benefits after the reform, whereas there is no change in unemployment inflow for workers with a tenure of five months who were ineligible for UI benefits even before the reform. We confirm the insights from the graphical analysis statistically in Table 3. Controlling for time-series variation in unemployment inflow (month fixed effects) in column I, we find that unemployment inflow relatively decreases by 0.44 percentage points for workers with tenure of six months compared to workers with a tenure of five months. Further saturating the specification with cipality-month fixed effects to account for local shocks in column II, the effect remains almost identical with 0.47 percentage points. The results are not affected by controlling for industry-specific local shocks (cipality-industry-month fixed effects) in column III. Even within the same occupation in the same local industry (cipalityindustry-occupation-month fixed effects), we find that unemployment inflow drops by 0.41 percentage points for workers with six months tenure compared to workers with five months tenure (column IV). 14

16 The relative decrease in unemployment inflow for workers with six months tenure after the reform is not compensated for by an increase in alternative forms of layoffs. Table 4 shows that there is no significant increase in voluntary layoffs (columns I IV). Similarly, there is only a mild and marginally significant increase in job changes after the reform for workers with a tenure of six months (columns V VIII). The increase in job changes accounts for less than five percent of the decrease in layoffs. Thus, net layoffs adjusted for job changes are virtually identical qualitatively and quantitatively compared to the raw results reported in Table 3. Brazil provides large heterogeneity in labor market informality across cipalities and industries. In Figure 5, we split the sample into workers employed in industries with above (top panel) and below (bottom panel) median levels of labor market informality. The graphical evidence reveals that higher unemployment inflow for workers with six months tenure before the reform is driven by workers in industries with above median levels of informality. For these workers, we observe a substantial drop in unemployment inflow in March and April when they lose eligibility for UI benefits. In contrast, for workers in industries with below median levels of informality we observe no significant change in unemployment inflow. In Figure 6, we split workers into those employed in cipalities with above (top panel) and below (bottom panel) median levels of informality. We find that in cipalities with above median levels of informality unemployment inflow decreases by more than one percentage point for workers that lose eligibility for UI benefits after the reform. In cipalities with below median levels of informality, the magnitude of the effect is less than a third as large as for cipalities with above median levels of labor market informality. Figure 7 plots the drop in unemployment inflow for workers with six months tenure relative to workers with five months tenure (x-axis) and the share of informal labor (y-axis) for each industry. The plot shows that the trends in figure 5 are not driven by individual industries but there is a clear relationship between the drop in unemployment inflow after the reform and labor market informality across industries. 22 While it is infeasible to plot the same graph for all cipalities due to the large number of cipalities in Brazil, the dashed line in Figure 7 shows the regression line from regressing the relative change in unemployment inflow for workers with six and five months tenure on the share of informal employment across all cipalities. The results in Figures 5 to 7 suggest that informal labor markets induce workers to terminate formal employment when they qualify for UI 22 We confirm that none of our results is driven by inidividual industries by performing all tests dropping each industry individually. 15

17 benefits. Informal employment opportunities allow them to simultaneously receive income from informal work and UI benefits. In Table 5, we formally assess how informal labor markets affect workers response to UI benefits. The top panel shows the results for variation in labor market formality at the industry level. We find that a ten percentage points increase in labor market informality leads to a 0.10 percentage points stronger decrease in unemployment inflow after the reform (column I). Controlling for local shocks that are specific to workers affected by the reform, the effect is almost identical with 0.12 percentage points (column II). Additionally controlling for local industry shocks leaves the effect virtually unchanged with 0.09 percentage points (column III). When we further add controls for shocks to specific occupations within a local industry, the magnitude of the effect is similar with 0.10 percentage points (column IV). Finally, accounting for unemployment inflow patterns of workers with six months tenure at the industry (column V) or local industry (column VI) level leaves the results qualitatively unchanged with a magnitude of 0.11 and 0.08 percentage points, respectively. The bottom panel lists results of the same analysis exploiting variation in labor market formality at the cipality level. A ten percentage points increase in labor market informality leads to a 0.13 percentage points stronger decrease in unemployment inflow after the reform compared to the pre-reform period accounting for industry and cipality-specific shocks (column I). To account for the fact that industry-specific patterns that differ for workers with tenure of six months compared to workers with tenure of five months, we control for industry-level shocks specific to workers with different tenure in column II. We find that the results are virtually unchanged with 0.11 percentage points. Even after controlling for local industry shocks, we find that unemployment inflow decreases significantly more by 0.10 percentage points per ten percentage points increase in labor market informality (column III). When we additionally control for occupation-specific local industry shocks in column IV, the effect is 0.08 percentage points. Controlling for unemployment inflow patterns of workers with six months tenure at the cipality (column V) or cipality-industry (column VI) level leaves the results unaffected with 0.11 percentage points each. Together, the results in Table 5 show that UI benefits have a stronger effect on unemployment inflow in the presence of informal labor markets. 16

18 4.2 UI Benefits Reform and Unemployment Outflow Figure 8 depicts reemployment probabilities conditional on unemployment duration for workers laid off during the months from January to April 2015, separately for workers laid off with a tenure of six months (top panel) and workers laid off with a tenure ot five months (bottom panel). For workers with six months tenure at the time of layoff, reemployment is significantly less likely to occur during January and February when they are eligible for UI benefits, compared to March and April when they are no longer eligible for UI benefits. In contrast, for workers with five months tenure at the time of layoff, unemployment outflow does not change significantly from January and February to March and April. The results in Table 6 show that reemployment probabilities within four months after layoff increase by 3.44 percentage points more for workers with a tenure of six months, compared to workers with five months tenure (column I). The effect is slightly weaker with 2.97 percentage points when we compare workers in the same geographical area (columns II). Further restricting the comparison to workers within the same local industry does not affec the results with 2.94 percentage points (column III), as does comparing workers within the same occupation with 3.17 percentage points (column IV). 4.3 Collusion The previous results suggest that, in the presence of more informal labor markets, the availability of UI benefits induces workers more strongly to exit and stay out of formal employment. This could be driven through different mechanisms. Workers may switch from formal to informal employment without involvement of their employer, for example by eliciting layoff through shirking. On the other hand, firms may collude with workers to extract rents from the government, for example by hiring them informally while they are eligible for UI benefits. To identify whether collusion between workers and their employers occurs, we explore whether firms that lay off workers just when they become eligible for UI benefits rehire these workers after benefits run out. For this test, we examine the probability of being rehired four to seven months after a layoff, since this is when benefits run out and we would expect strategic rehiring to occur. We follow our main identification strategy comparing dismissed workers with six months of tenure at the layoff who lose eligibility after the reform to those 17

19 with five months tenure who are always ineligible. 23 The results are gathered in Table 7. Column I shows that before the reform the probability to be rehired by the same employer four to seven months after layoff is about one percentage point higher for workers with a tenure of six months compared to those with five months of tenure. After the reform, when both types of workers are ineligible for UI benefits, the difference in rehiring by the same firm four to seven months after layoff vanishes. The results hold even within the same local industry (cipality-industry-month fixed effects) in column II, and within the same occupation within a local industry (cipality-industryoccupation-month fixed effects) in column III. These results suggest that there is collusion between workers and their employers. Firms formally fire workers when they qualify for UI benefits and formally rehire them when benefits are exhausted. In columns IV to VII, we examine whether collusion between workers and their firms is concentrated in industries and cipalities with large informal labor markets. The results in columns IV V show that firms in more informal industries are significantly more likely to lay off workers when they are eligible for benefits to rehire them after benefits run out. Specifically, a ten percentage increase in labor market informality leads to an about 0.8 percentage points increase in strategic layoff and rehiring. The results are similar at the cipality level with slightly higher magnitudes (columns VI VII). This cross-sectional evidence suggests that informal labor markets provide firms and workers with an opportunity to collude. informal relationship. They can extract rents through UI benefit payments, while maintaining an Finally, we are interested in what fraction of the lower unemployment outflow of eligible workers is driven by the type of collusion described above. This allows us to understand how prevalent collusion is. The additional number of workers that does not return to formal employment while qualifying for UI benefits amounts to 3.5 percent of all laid off workers. Subsequently, an additional 0.7 percent of all laid off workers are rehired by the same firm when their benefits run out. Put differently, about 20 percent of workers who strategically remain unemployed when on benefits are rehired by the same employer. This suggests that 20 percent of strategic unemployment involves direct collusion between employees and firms For this test, we cannot compare workers laid off in March to April against January to February 2015, as their rehiring probabilities are influenced by the reform, which is announced in December Instead, we compare workers laid off in March and April 2015 to workers laid off in March and April 2014 as these workers are laid off and rehired within the same regime, respectively. Additionally, examining workers laid off during the same calender months directly controls for cyclical effects. 24 Note that these are difference-in-differences estimates where we compare the mean probabilities for 18

20 Clearly, the collusion that we pick up here captures only part of all possible types of collusions. For example, several firms and employees as a group could engage in collusion in a way that our test would not be able to identify as collusion. However, our analysis identifies the easiest to implement and the most direct form of collusion. Thus, we believe it is plausible that we pick up a large fraction of the possible collusion in the data. 4.4 Formal Employment and Wages In this section, we examine whether the reduction in expected future UI benefits has a differential effect on workers decision to enter formal employment ex ante in the presence of informal labor markets. In the absence of informal labor markets, higher expected future income from formal employment may lead the marginal worker to prefer formal employment over unemployment leading to an increase in formal employment. When informal labor markets exist, in addition to unemployed workers, workers that would otherwise have chosen informal employment may shift to formal employment when qualifying for UI benefits is easier. Figure 9 depicts the time-series evolution in average formal hiring scaled by total employment in a given industry (top panel) and cipality (bottom panel), separately for labor markets with above (solid lines) and below (dashed lines) median levels of informality. 25 We adjust all plots for industry-calender month fixed effects to take out cyclical fluctuations. In January 2015, the month after the reform was announced, we start to see a drop in formal hiring in industries with above median levels of labor market informality relative to industries with below median levels of labor market informality. The evolution in hiring continues to diverge over the next 12 months with formal hiring in highly informal industries continuing to relatively decrease. We observe the same time-series pattern for cipalities sorted based on labor market informality. 26 workers with six months tenure against workers with five month tenure before and after the reform. The difference between workers with five and six months tenure after the reform provide us an estimate of what fraction of workers would have returned to employment within three months in the absence of benefits as both types of workers do not qualify for benefits after the reform. Then we compare this difference to the pre-reform period to gauge the fraction of worker with six months tenure staying unemployed strategically before the reform. 25 To control for industry-driven hiring trends in aggregated cipality-level hiring, we adjust cipality-level data for industry composition to capture labor market informality not explained by industry composition. 26 The patterns are unchanged when we consider net hiring rates, that is the difference between newly hired and fired workers. 19