52737 World Bank Employment Policy Primer December 2009 No. 13 ThE ImPacT of NoN-WagE BENEfITs oN JoB QualITy aND laBor markET ouTcomEs IN ThE DEvEloPINg WorlD: WhaT Do WE kNoW? Brooke Helppie and Mario Macis One indicator of job quality in the developing world of worker relative to another may cause employment is the extent to which workers are covered by non- composition effects, and benefits applying differently by wage mandated benefits. Mandated benefits are a class firm size may distort firm, and thus economic, growth. of labor market regulation that requires employers to In general, the higher the cost of providing mandated provide their workers with some form of non-wage pay benefits, and the less equally distributed that cost is, the or benefits. These benefits are typically fully financed greater will be the distortions caused by these benefits. by the employer or the employer-worker pair, and Unfortunately, empirical research on the effects administered at the firm level. Some examples include of mandated benefits on labor market outcomes in health insurance or healthcare, parental leave, sick leave, developing countries is still very scarce and, thus, it is vacation leave, vacation bonus pay, "thirteenth month" hard to draw general conclusions. Moreover, because wages, and transportation or reimbursement for work- of the heterogeneous nature of mandated benefit laws, related transportation costs. and the multi-layered institutional settings in which This note summarizes the existing knowledge and they are implemented, there is probably no universal research on the effects of providing higher mandated message about the effects of such laws. We can learn benefits on labor market outcomes. In particular, we from economic theories and existing studies, such as focus on the evidence available for developing econo- those expounded in this note, but much more research mies. We discuss the potential advantages and drawbacks must be done to increase our understanding of the real- of providing mandated benefits (as implied by econom- world effects of mandated benefits. This note sets out ic theory), summarize some of the existing empirical to introduce some of the theory and existing literature literature on the impact of mandated benefits, and give about mandated benefits, and to draw attention to the some country-specific examples of existing programs. need for more research about the impact of this type While laws imposing some mandated benefit may of policy. yield needed benefits to particular groups of workers, or ease government budget concerns, their implementation costs and Benefits must be weighed carefully. Theory predicts that a benefit to all covered workers, funded by firms at the worker- Governments may choose to mandate benefits as level, may lead to increasing informal employment. Spe- an alternative to providing a social good or service to cially-targeted benefits that change the costs of one type the entire population, or because the need for a particu- T h e I m p a c t o f N o n - Wa g e B e n e f i t s o n J o b Q u a l i t y a n d l a b o r m a r k e t o u t c o m e s i n t h e D e v e l o p i n g Wo r l d : W h a t d o w e k n o w ? lar benefit only exists among the working population. reduce direct wages paid to workers, but the effect on Because mandated benefits appear to be revenue-neu- employment is ambiguous and depends on the degree tral, they are often implemented in lieu of a govern- to which workers value the benefit. ment-funded program.1 However, if mandated benefits require per-worker Also, since mandates generally state a minimum lump sum contributions, employment may shift away benefit or outline of requirements, these programs can from part-time or low-wage workers because the cost to be implemented at the firm or individual level in a employ them is disproportionately increased. The com- manner that best fits the needs of the employer and the position of employed workers may also be affected by employee. This flexibility can allow for a benefit to be mandated benefits when payments are only required for administered in a more efficient manner than it could a subset of workers. For example, the cost of maternity be at a federal or state level, both because the ability of leave programs may be paid only for workers who are firms to tailor benefits to the needs of their employees women of childbearing age. Then, these workers may be can reduce overall cost, and because non-government discriminated against in the labor market, resulting in administration avoids inefficiencies due to the political lower employment rates, wages and/or formality status nature of budgeting government programs. of this group of workers. Another potential advantage to mandated benefits Interactions with other labor market institutions over government-administered programs is that work- and the state of the economy are also important. For ers may more readily understand that contributions instance, the presence of a minimum wage would com- for these benefits compose part of their compensation plicate the analysis from Graph 1: it might further package, as compared to a payroll tax (also part of reduce employment while mitigating the decline in the compensation package) to fund a government- provided benefit. This may have positive effects on labor supply. When considering whether to impose a mandated Figure 1: Constant hourly cost of mandated benefit, however, it is also important to consider the benefit, no binding minimum wage drawbacks associated with these. Some of the potential Real wage LS (excluding costs of mandated benefits are best illustrated in the bene ts) LS' context of basic economic theory, and are discussed in Employee valuation 1 the next section. Another important aspect to consider is enforcement: mandated benefits may be difficult to enforce because the mode of implementation can vary w Employee valuation 2 greatly from firm to firm. For example, quality of health- w' LS'' care or health insurance provided by individual firms may be difficult for government inspectors or auditors w'' Cost to employer to assess. Mandated benefits must be enforced to have an effect, and enforcement is not costless. LD LD' L L' L'' Labor Theoretical considerations Employee valuation 1 indicates labor supply e ects with low valuation of the bene t. Employee valuation 2 represents valuation of bene ts above the cost to the employer. Point (L, w) represents Standard theoretical models2 of the effects of man- equilibrium employment before a bene t is mandated. Point (L',w') indicates a possible equilibrium dated benefits on the labor market are similar to those under low employee valuation, and point (L",w") shows a possible outcome if a mandated bene t is highly valued by workers. for payroll taxes. An important difference is that the value workers place on mandated benefits can mitigate the negative employment effects of such programs by 1 Since no regulation or institution operates in a vacuum, this reducing the direct wages workers require to supply assumption may often be false. their labor. A standard, static labor market model with 2 See Boeri, Helppie and Macis (2008), pages 20-24 for a more a mandated benefit may look like the graph below. The thorough discussion of the theoretical considerations related to impact of a mandated benefit program is expected to mandated benefits. 2 T h e I m p a c t o f N o n - Wa g e B e n e f i t s o n J o b Q u a l i t y a n d l a b o r m a r k e t o u t c o m e s i n t h e D e v e l o p i n g Wo r l d : W h a t d o w e k n o w ? wages for those workers who remain employed.3 The dicted disemployment effects of mandated benefits as case of a binding minimum wage is particularly salient compared to payroll tax-funded programs in standard in developing countries, where de jure minimum wages models. Because of the important differences between apply to a large portion of the workforce. payroll tax funded benefits and mandated benefits, this Further, effects of mandated benefits may be dif- note excludes discussion of research on payroll taxes ficult to predict: mandating that firms provide health from the empirical digest.5 insurance to workers may greatly change the size and Mandated Employer-Provided Health Insurance: As scope of a country's insurance industry, the percent- an alternative to nationalized provision of health care age of health care dollars required for administrative or insurance, policymakers may opt to mandate provi- costs, and the need for oversight. Mandated employ- sion of health insurance through employers. A large er-financed maternity leave could bias hiring against number of studies using US data have been written in women of child-bearing age or increase the gender pay recent years, as this country attempts to improve access gap. to healthcare. One recent paper constructed an estimate predicting that 0.2 percent of workers, and 1.4 percent of Empirical evidence unemployed workers in the United States might become unemployed with such a mandate (Baicker and Levy, The effects of many types of mandated benefits on 2007). One third of uninsured workers earn wages close labor markets have not been studied systematically in enough to the minimum wage that their wages cannot the case of developing countries. Early work on man- be fully adjusted to absorb the cost of health insurance, dated benefits generally used aggregated time-series so they would be put at risk of unemployment by a data or data in cross-section only. These papers gener- health insurance mandate. This is an example in which ally found ambiguous impacts of mandated benefits workers who would be targeted by such a mandate on wages and employment. Some recent researchers would suffer disproportionate job losses, underscor- have attributed these unclear findings to measurement ing the importance of attention to policy targeting and error and biases arising from regression techniques, impact analysis. and have attempted to improve upon these by using We did not find studies of this type of mandate micro-level longitudinal datasets from individual for developing countries. However, because the Unit- countries. Indeed, given that countries differ in insti- ed States' minimum wage is relatively non-binding, tutions, culture, rule of law and business cycles, analyz- compared to those of many developing countries, the ing the effects of mandated benefits within countries labor market consequences of such mandated benefits seems a solid approach to beginning to understand in developing economies could be yet more dramatic. their effects. Pregnancy and Maternity Coverage: A focus of the Given the paucity of studies on mandated ben- mandated benefits literature, in both developed and efits in developing countries, it can be useful to draw developing countries, has been on pregnancy and mater- on studies of government programs funded by payroll nity policies. taxes. Theory suggests that payroll taxes and mandated A study in the early 1990s found a large degree of benefits will have similar effects on the labor market. wage-shifting in the United States as a result of anti- The costs of programs are generally borne to some discrimination legislation that mandated health insur- extent by both firms and workers. Sometimes the bur- den is borne by some firms or industries more than oth- 3 The above discussion focused on a single labor market with stan- ers, depending on the applicability of programs or their dard labor supply and demand curves and assumed homogeneous funding mechanisms.4 workers. Relaxing the assumption of perfectly competitive markets However, mandated benefits may also be more high- in this model, also a consideration in developing country literature ly valued by workers than government benefits funded can result in workers capturing more rent through benefits. De la by payroll taxes, because the connection between con- Rica and Lemieux (1994) provide a more rigorous model of such decisions. tributions and benefits is more transparent. Addition- 4 Anderson and Meyer (1997) ally, there may be efficiency benefits to non-government 5 For more detail on some payroll taxation studies, see Boeri, Help- programs. Both of these factors may reduce the pre- pie and Macis (2008). 3 T h e I m p a c t o f N o n - Wa g e B e n e f i t s o n J o b Q u a l i t y a n d l a b o r m a r k e t o u t c o m e s i n t h e D e v e l o p i n g Wo r l d : W h a t d o w e k n o w ? ance coverage of childbearing costs. The implication enforcement of the minimum wage, yet again under- of this finding is that workers bear most of the burden scoring the importance of considering enforcement and of these laws. Interestingly, this paper also found an incentives to operate informally when developing and increase in hours worked by women of prime child- analyzing the impact of mandated benefits. bearing age (Gruber, 1994). Another study found that Unemployment Insurance: Investigation into the Canadian policies that guaranteed time off and a right effects of taxes that vary at the firm level, through to return to the same or a similar job for new mothers the example of the experience-rated unemployment increased labor force participation of women before insurance system in the United States, found that while childbirth and increased rates of return to the same market-level tax is largely borne by workers, firm- employer after leave (Baker and Milligan, 2005). These level variation in taxes appears to be borne by firms, findings appear to confirm the predictions of basic both in terms of wages and employment (Anderson theory that substitution away from costly groups of and Meyer, 1997). It is important to note, however, workers (women of childbearing age) can be mitigated that employment reallocation across firms and micro- through increased productivity by these workers (for level deadweight losses may exist even if no aggregate example, by increased labor force attachment in antici- employment effect is apparent. This is an important pation of receipt of maternity benefits) or more firm- idea, and one that underscores the desirability of careful specific human capital accumulation (made more likely firm-level studies. by increased likelihood that a woman will return to her prior job after leave). conclusion and Policy recommendations Developing country evidence is scarce, but does sup- port some of the same findings. A study of the impact of Policymakers around the developing world are con- women's work hours and maternity benefits in Taiwan cerned with identifying policies that enhance the quality show increased labor force participation in both hours of jobs being created. In particular, they hope to design and employment, and no wage effects (Zveglich and van labor market regulations that simultaneously provide der Meulen Rodgers, 2003). Of particular importance the socially needed non-wage benefits without intro- in the case of developing countries, however, is that ducing large distortions in the labor market. Mandated statistically significant effects were found only after the benefits can be an important potential tool in govern- laws began to be enforced. Likelihood and extent of ments' battles against poverty and exploitation, and help enforcement should be considered when evaluating the achieve higher levels of employment and training. They effects of mandated benefits in developing countries, have the potential to increase workers' welfare through where rule of law may be weaker than in more developed more valuable or better-tailored compensation pack- countries. ages, and to reduce waste by allowing for local solutions The "Thirteenth Salary," vacation pay and other to nationwide issues. On the other hand, mandated required bonuses: In many countries--both more and benefits may encourage firms and workers to operate less developed--Christmas bonuses, vacation pay at in the informal economy, reduce wages, and may even greater than 100% of the wage, and other required reduce employment of the groups of workers they were bonuses comprise and important part of the wage struc- designed to help. ture. To date, surprisingly little research has examined The literature on the provision of mandated benefits the labor market impact of such compensation, espe- is one area in the labor law and enforcement literature cially in less developed countries. However, one study, which has not yet been thoroughly or systematically on the impact of Ecuador's 13th, 14th, 15th and 16th explored. Often conflicting conclusions in this literature salaries, plus several other mandated bonuses and pay- do not yet allow us to draw broad conclusions. However, ments, was conducted in 1997 by MacIsaac and Rama. because the nature of the benefits is also very hetero- Their analysis showed that the total compensation pack- geneous across countries in terms of policy structure, age averaged just 18% more in private sector jobs com- implementation and enforcement, it is also very difficult plying with labor regulations than otherwise identical to draw these broad conclusions. jobs, indicating a high degree of wage shifting. This large As with any policy, mandated benefits may interact degree of wage shifting is probably enabled by weak with the existing institutional structure of a government 4 T h e I m p a c t o f N o n - Wa g e B e n e f i t s o n J o b Q u a l i t y a n d l a b o r m a r k e t o u t c o m e s i n t h e D e v e l o p i n g Wo r l d : W h a t d o w e k n o w ? Box 1: Policy Considerations Box 3: Mandated Benefits and Earnings When evaluating a potential mandated benefit policy, the following questions may be Composition in Brazil worth consideration: Component Percent Total 1. Who are the targeted groups? Basic wage 100 2. Is this policy likely to cause one group of workers to be substituted for another? Annual bonus 8.3 108.3 3. What value are workers likely to attach to the benefit? Vacations 11.3 119.6 4. How much will this policy cost firms? Severance fund contribution (FGTS) 8 127.6 5. Will it cost firms a set amount per worker, or a proportion of wages? Other mandatory benefits* 10 137.6 6. How might this policy incentivize firms or workers to operate in the informal economy? Total pay (basic wage + mandatory benefits) 137.6 7. How might this policy distort existing markets for the benefit being mandated? SESI, SENAI, SEBRAE (employer associations) 3.11 140.7 8. What will enforcement of this policy cost, and how effective is it likely to be? INSS** + Accident Insurance + Education + INCRA 24.7 165.4 9. What institutions, laws and other markets will affect the implementation and success of Source: Amadeo, Gill, Neri (2000), p. 7. this program, and how? * Includes family allowances, pregnancy leaves, transport subsidies, etc. Cannot be calcu- lated for all workers, since they depend on gender, type of work, sector, etc. ** Workers contribute with 8%, 9% or 10% of the wage to social security depending on the wage. Box 2: Mandated Benefits and Earnings Composition in Ecuador (Sucres per Hour as of September 1994) workers, but is generally predicted by standard eco- Component Legal minima nomic theory to reduce other wage payments (reduc- 1. Base wage 412 ing the positive impact on wages) and also to reduce 2. Tips and overtime 0 employment. 3. Voluntary bonuses (vacation, Christmas, firm anniversary, etc.) 0 Particularly relevant to developing countries is that 4. Voluntary allowances and payments in kind (food, housing, generous mandated benefits may raise incentives for clothing, etc.) 0 workers to seek jobs in the informal sector or for firms A. Basic earnings (=1 + 2 + 3 + 4) 412 5. Teen payments 196 to operate informally, thus increasing the size of the 6. Compensation bonus 294 unregulated economy. They also may also distort the 7. Cost of living bonus 659 composition of the economy, if they are implemented 8. Transportation allowance 71 differently by firm size or industry (e.g., Anderson and B. Mandated benefits (=5+6+7+8) 1,220 Meyer, 1996). C. Take home pay (=A+B) 1,632 D. Social security contribution by worker 43 references E. Payroll taxes 56 F. Total labor cost (=C+D+E) 1,731 Amadeo, E.J., Gill, I.S., and M.C. Neri. (2000). Brazil: Source: MacIsaac and Rama (1997), Table 2 The Pressure Points in Labor Legislation. Getulio Note: In the calculation of the legal minima, a month was supposed to include 170 hours of work and no overtime. All legal figures correspond to the most general regime. Vargas Foundation Economics Working Paper 395. Anderson, P. M. and Meyer, B. D. (1997). The Effects of Firm Specific Taxes and Government Mandates with an Application to the U.S. Unemployment Insurance Program. Journal of Public Economics, 65, and economy to produce unintended effects. Policymak- 119­145. ers and researchers should strive to consider interactions Baicker, K. and Levy, H. Employer health insurance with other institutions when predicting or analyzing the mandates and the risk of unemployment. National effects of any law. Bureau of Economic Research Working Paper 13528. Policymakers should also consider the goals of man- Baker, M. and Milligan, K. (2005). How Does Job-pro- dating a benefit. For example, a "thirteenth month" tected Maternity Leave Affect Mothers' Employment salary may be intended to increase the pay received by and Infant Health? NBER Working Paper 11135. 5 T h e I m p a c t o f N o n - Wa g e B e n e f i t s o n J o b Q u a l i t y a n d l a b o r m a r k e t o u t c o m e s i n t h e D e v e l o p i n g Wo r l d : W h a t d o w e k n o w ? Boeri, T., Helppie, B. and Macis, M. (2008). Labor Gruber, J. (1994). The Incidence of Mandated Mater- regulations in developing countries: a review of the nity Benefits. The American Economic Review, 84 evidence and directions for future research. World (3), 622­641. Bank. SP Discussion Paper 0833. MacIsaac, D. and Rama, M. (1997). Determinants of De la Rica, S, and Lemieux, T, (1994), Does public Hourly Earnings in Ecuador: The Role of Labor health insurance reduce labor market flexibility or Market Regulations. Journal of Labor Economics, encourage the underground economy? Evidence 15(3:2), S136­S165. from Spain and the United States? In Blank, R., Zveglich, J. E., and van der Meulen Rodgers, Y. (2003). Social protection versus economic flexibility: is there The Impact of Protective Measures for Female a trade-off? (pp. 265­299). Chicago: University of Workers. Journal of Labor Economics, 21 (3), 533­ Chicago Press. 555. The findings, interpretations, and conclusions expressed herein are those of the author(s), and do not necessarily reflect the views of the International Bank for Reconstruction and Development / The World Bank and its affiliated organizations, or those of the Executive Directors of The World Bank or the governments they represent. 6