发展中国家劳动力市场机构：不同观点的时序证据

W_F_s I jq o0 POLICY RESEARCH WORKING PAPER 1940 The Structure of Labor The informal sector behaves as an unregulated Markets in Developing entrepreneurial sector rather Countries than the disadvantaged segment of a dual labor market. Overall, it expands in Time Series Evidence upturns and contracts in downturns, though there is on Competing Views oevdnofuigt some evidence of queuing to enter the formal sector. William F. Maloney The World Bank Latin America and the Caribbean Region Poverty Reduction and Economic Management Unit June 1998  POLICY RESEARCH WORKING PAPER 1940 Summary findings Competing conceptions of the large, unprotected, This is particularly relevant in the debate about "informal" workforce in developing countries differ establishing common labor standards in regional trade greatly in their implications for the labor reform agreements. considered to be essential complements to trade Maloney reexamines the traditional view of the dual liberalization and "fair" competition in international labor market by studying the dynamics between the trade. formal and informal sectors across a business cycle and a Traditionally, the informal sector is viewed as the period of trade liberalization in Mexico (1987-93). disadvantaged segment of a dual labor market segmented He shows conventional comparisons of earnings, even by legislated or union-induced rigidities and high labor across time, to be unreliable tests for segmentation. As an costs in the protected (or "formal") sector. In this view, alternative, he shows that transitions on informal the size of the informal sector is a testament to the employment, the size of the informal sector, and levels of inefficiencies in labor allocation and the magnitude of mobility to be procyclical, increasing with upturns, and required reform. In cyclical downturns, the informal decreasing with recessions. He tests for, and finds, sector is thought to absorb displaced workers from the however, some evidence of queuing to enter formal formal sector (with informal earnings falling relative to employment. those in the formal sector) and then to contract again Overall, he contends, the informal sector behaves as an during recovery as the queue for "good jobs" shortens unregulated entrepreneurial sector rather than the again. disadvantaged wing of a dual labor market. There is A recent, related view postulates a long-term trend in evidence of increased subcontracting over time, with which large enterprises, confronted by heightened global trade liberalization, but it is not clear that workers are competition, increasingly subcontracts to unprotected worse off as a result. workers as a way to reduce costs and gain flexibility. This paper - a product of the Poverty and Economic Management Unit, Latin America and the Caribbean Region - is part of a larger effort in the region to reexamine the role of the informal sector. The study was funded by the Bank's Research Support Budget under the research project "The Informal Sector in Mexico" (RPO 680-59). Copies of this paper are available free from the World Bank, 1818 H Street NW, Washington, DC 204.33. Please contact Marta Cervantes, room 18-095, telephone 202-473-7794, fax 202-522-0054, Internet address mcervantes@worldbank.org. The author may be contacted at wmaloney@worldbank.org. June 1998. (29 pages) The Policy Research Working Paper Series disseminates the findings of work in progress to encourage the exchange of ideas about development issues. An objective of the series is to get the findings out quickly, even if the presentations are less than fully polished. The papers carry the names of the authors and should be cited accordingly. The findings, interpretations, and conclusions expressed in this paper are entirely those of the authors. They do not necessarily represent the view of the World Bank, its Executive Directors, or the countries they represent. Produced by the Policy Research Dissemination Center  The Structure of Labor Markets in Developing Countries Time Series Evidence on Competing Views William F. Maloney* I thank the Mexican National Institute of Statistics, Geography, and Information (INEGI) for the use of the data. INEGI is in no way responsible for any incorrect manipulation of the data or erroneous conclusions drawn from it. I also thank Lee Alston, Hadi Esfahani, Kevin Hallock, Steve Kamin, Christopher Pissarides, and Doug Steigerwald for helpful comments, and Taimur Baig for research assistance.  L Introduction The competing conceptions of the large, unprotected or "informal" work force in developing countries differ greatly in their implications for the labor market reforms deemed essential to complement other liberalization measures,' as well as issues of "fair" competition in international trade. The traditional view sees informality as the disadvantaged segment of a dualistic labor market segmented by legislated or union-induced rigidities and high labor costs in the protected or "formal" sector.2 The large size of the sector thus testifies to the extent of inefficiencies in labor allocation and the magnitude of required reforms. Downward formal wage rigidity also implies strong predictions about sectoral interactions across time. In cyclical downturns, the informal sector is thought to absorb displaced formal sector workers, informal earnings falling relative to those in the formal sector, and then contract during recoveries as lhe queue for good jobs shortens again. A recent and related approach also postulates a long term trend where large enterprises, confronted by heightened competition in global manufacturing, increasingly sub-contract to unprotected workers as a means of reducing costs and gaining flexibility.3 The issue is thus particularly relevant to the debate over establishing common labor standards in regional free trade agreements. 'See, for example, World Bank (1995). 2 The Harris-Todaro (1970) model is perhaps the traditional statement of this view. See also Sabot (1977), and Mazumdar (1983). An alternate view within the segmentation literature sees dualism as arising endogenously from efficiency wage type considerations that induce larger irmns to pay remuneration above market clearing (See Stiglitz, 1974; Esfahani and Salehi-Isfahani, 1989). See Rosenzweig (1988), Fields (1990), Thomas (1992), Tokman (1992), Portes and Schaufiler (1993), Portes (1994) for overviews. 3 See Piore and Sabel (1984), Beneria (1989), Port=s, Castells and Benton (1989), Portes and Schauffler (1989), Tokman (1992) and most recently articles in Blank and Freeman (1994). The U.S. Department of Labor has sponsored two studies on the informal sector: Worikers Without Protections: Case Studies of the Informal Sector in Developing Countries (1993) and The Infornal Sector in Mexico (1992).  However, there is some support for an alternate conception that sees the lack of protection as one dimension of an unregulated, yet dynamic, sector of small scale entrepreneurs many of whom enter the sector voluntarily and who choose, and are able, to remain largely outside the formal regulatory structures.4 Labor markets may therefore be broadly integrated, and the existence of unprotected labor is not in itself evidence of segmentation. Arguably, workers contemplating self- employment would wait for an auspicious business climate before leaving a protected job to launch their enterprise and would be more likely to fail during a downturn, inverting the counter-cyclical patterns of entry and exit predicted by the dualistic view. As. will be argued, the benefits of being informal may extend even to subcontracting relations which are generally thought to benefit only the employer. Research on the informnal sector has largely relied on case studies, or static comparisons of earnings differentials which tend to show that formal sector workers earn more given their human capital. But to date, the data have not existed to study sectoral interactions across time. This paper takes advantage of an extraordinary longitudinal data set from Mexico that permits an examination of the dynamics among the various subsectors of the labor markets from 1987-1993. This period encompasses a complete business cycle, a far reaching trade reform, and pursuit of regional integration through NAFTA. The paper first argues that traditional earnings comparisons are not reliable measures of segmentation, even when viewed across time. It then examines movements in relative sector sizes, the role of the various sectors in generating and absorbing the unemployed, and the changing patterns 4See, for example Hart (1972), de Soto (1989), Tokman (1992), Fields (1990), Tunham and Er&cal (1990), Portes and Shauffier (1992). 2  of mobility among sectors across the business cycle. It also explicitly explicitly tests for queuing to enter formal sector employment. Overall, the second "entrepreneurial" view emerges as the better first approximation. The last section offers preliminary evidence on the incidence of subcontracting and its impact on worker welfare by tracking the composition of the labor force after trade liberalization in a more comprehensive way than has been possible to date. IL Data The Mexican case is well-suited to the study of dualism, informality, and subcontracting. First, roughly 40% of the urban labor force works without labor protections. Second, an archaic and onerous system of labor regulation, dating back over half a century, grants little flexibility and drives up labor costs. The Constitution conceives of the employment relationship as a lifetime contract and workers may only be fired under extreme circumstances and at great cost. For their part, workers lose generous severance pay and may lose their pensions if they quit, and enjoy no unemployment insurance should they be laid off. Taken at face value, the labor code thus discourages the mobility necessary to ensure an efficient allocation of workers and jobs, and constrains employers in their management of personnel. Legally mandated non-wage compensation drives a wedge of 30-60%, substantially above those for other OECD countries. Third, the radical trade reforms beginning in 1986 and the subsequent pursuit of regional inl:egration through NAFTA converted a relative closed economy to one competing aggressively in both import and export markets. The resulting adjustment of its labor market can offer some preliminary observations on the issue of globalization and subcontracting. The National Urban Employment Survey (NUES) conducts extensive quarterly household 3  interviews in the major metropolitan areas and is available from 1987 to 1993. This permits the generation of time series of relative earnings for each sector at quarterly frequency. The NUES is also structured so as to generate panels which allow tracking a fifth of each sample across five quarters. Workers are matched by position in an identified household, sex, level of education, and age to ensure against generating spurious transitions. These panels permit sketching patterns of mobility among sectors that can be tracked across the seven year period. The analysis restricts itself to men aged 16-65 with a high school education or less. While the term "informal" suffers from overly broad and imprecise usage, it refers here to owners and workers in firms under 16 employees who do not have social security or medical benefits and are therefore not protected.' Four sectors of paid work are studied that include formal salaried employment and three "informal" sectors: The self-employed including owners of informal firms; the informal salaried, those working in informal firms; and contract workers, those who do not receive a regular wage or salary, but who are paid as a percentage, by piece, on commission, or fixed contract. They may be affiliated with a larger firm that provides raw materials, but work independently and are those most likely to be involved in subcontracting relations. There is also a residual "other" category (not shown) that includes firm owners who are protected, and owners and unprotected workers in firms of over 16 people. The remainder of the interviewed population (to total 100%) is divided into three classes of individuals who are not working: those out of the labor 5 There appear to be workers at larger firms who do not report having benefits, however this is stretching our idea of informality beyond the smaller scale industries usually contemplated and may, in addition, be due to measurement error. Since it is often the wife of the worker who is at home at the time of the survey, it is entirely possible that they are not informed about their husbands true employment conditions. STPS also argued that there may be some very transitory workers for large firms not covered in violation of the law. 4  force, not currently working and not searching; those studying; and the unemployed.6 Movements among sectors are best captured in transition matrices that present the conditional probability of finding a worker in sector j at the end of the period given that the worker began in sector i, Pij, as well as initial and terminal sector sizes, PL and PJ. As this would require 24 large matrices to cover the sample period, figures land 2 instead graphically present elements of these matrices serially. Figure 1 shows the evolution of the sectoral composition of the interviewed population PL using the entire sample. Figure 2 plots the probability of transition into and out of the formal sector standardized by the probability that in a random reshuffling a worker would finish in the terminal sector (relative sector size) P?j. Figure 3 presents median hourly real earnings, adjusted for human capital.7 m. Overview of the Period. Using the median hourly real earnings in each sector in figure 3, and the macro-aggregates in tables 1-3, we can divide the period 1987-1993 broadly into three phases, recession, recovery, and slowdown. Phase I: As Lustig (1992) details, 1986 and 1987 were years of deep recession surrounded by the uncertainty of the success of the stabilization and reform programs. The balance of payments The definition of unemployment differs slightly from the official which includes those searching for work. It includes this group, plus those who are waiting for the response to an application, a call from an employer that is expected in 3 months or less, the next cycle of work, and those who consider that there is currently no work or that they would not be given work if they applied. It then subtracts those in school who are looking for work because they already occupy the category of "Studying." This generates a statistic that broadly tracks, although is above the official rate tabulated in table 1. 7The medians were calculated conditional on experience, experience squared, education and education squared. 5  crisis in mid-1985 led to extreme contractionary measures which were undermined by the 50% fall in oil prices in 1986. GDP fell by 3.8% and manufacturing wages fell another 5.90/o. Before turning up again, real wages had fallen a cumulative 39% since the beginning of the crisis. As figure 3 suggests, the recession had a depressive effect on earnings through 1987, which, in all sectors, show great downward flexibility. During this period, Mexico embarked upon the first wave of dramatic trade liberalization: From 1985 to 1990, maximum import tariffs fell by 50% while import licensing fell to a quarter of previous levels.8 The depreciated peso and depressed wages, however, initially provided some cushion of competitiveness. Phase II. As table 2 shows, economic growth began a moderate recovery in 1987, 1.86%, that would peak at 4.46% in 1990. In December, the Economic Solidarity Pact, a joint agreement of government and formal representative of labor, and the business sectors was implemented that featured a comprehensive incomes policy supported by reduction ofthe fiscal deficit, tighter monetary policy, and trade liberalization. 12 The Pact successfully reduced inflation to 1.2% per month in the second half of 1988. Although employment growth was moderate and concentrated heavily in the construction sector, in 1990 open unemployment fell to 2.6%, its lowest level since 1976 (see tables 1 & 3). As figure 3 shows, from 1988:1 to roughly 1990:3, earnings in all sectors grew concomitant with the tightening labor market, especially in the self-employed and contract sectors where they rose over 30%. In 1990, the government initiated discussions of a free trade agreement with the United States. s See Lustig (1992) and Maloney and Azevedo (1996) 2In February 1988, the governuent effectively froze public prices and the exchange rate, and minimum wages were raised 3% then fixed. For its part, the private sector committed to not raising prices. 6  Phase m: The economy began to soften from 1992-1993 with growth rates slowing dramatically to .45% in 1993, sustained only by continued growth in the non-manufacturing sectors. The official measure of unemployment again rose to 3.4% as manufacturing shed workers and job creation in the construction and commerce sectors slowed. Earnings in all informal sectors declined in absolute terms and relative to formal sector wages which stagnated. The next two sections test the competing hypotheses outlined previously by examining first, the behavior of earnings differentials and second, patterns of mobility across the period. IV. Relative Earnings Movements and Segimentation. Figure 3 reveals clear patterns in relative earnings. First, seemingly contradicting the segmentation arguments, self-employment pays generally better than formal sector employment. Second, the similarity of the behavior of contract worker earnings suggests that these two sectors are closely related. Third, informal salaried workers appear to receive consistently less than the formal sector wage. The literature on dualism has relied almost exclusively on these kinds of cross-sectoral earnings comparisons, finding higher formal sector earnings to be evidence of segmentation."3 However, the interpretation of earnings differentials cannot be this straightforward since the specific characteristics of work that pertain to or even clefine the formal and informal sectors affect earnings in each sector and make it unclear what the magnitude or sign of the differential should be even in an unsegmented market. In a market with no distortions, earnings in the informal sector should rise 13 See Rosenzweig (1988). Somewhat unusually, Marcouiller, Ruiz, and Woodruff (1994) find higher mean earnings in the Mexican informal sector than the formal. 7  above that in the formal sector to compensate for the expected value of benefits received by formal sector workers. Similarly, formal sector workers would require compensation for taxation which informal sector workers may often avoid. Formal salaried work and informal self-employment may also differ in hours worked, degree of risk taken, degree of independence, and costs of capital invested that may further drive a wedge between reported earnings. Informal salaried workers are frequently young and related to informal entrepreneurs and thus their earnings may be net of training costs or unobserved payments in kind. In the absence of information on these factors, the magnitude of the distortion-free differential cannot be known a priori and the interpretation of the raw earnings differentials reported in previous studies as evidence of segmentation becomes less clear. In fact, it is arguable that the customary strategy of establishing segmentation by observing earnings differentials should be inverted. There is some reason to believe that Phase II represents a period absent of segmentation: As Bell (1994) and Davila (1996) have argued, the minimum wage was not binding and the historically low unemployment rates were held by Lustig to be the result of extraordinary downward flexibility in (formal sector) wages. It may therefore be reasonable to consider the observed differentials in figure 3 as reference values for unsegmented markets that account for all the effects detailed above. If this is the case, the assumption that a differential of zero represents an unsegmented market underlying many previous studies is probably unjustified. There is, however, substantial variance in differentials across the period. Earnings in the closest substitute to formal salaried work, informal salaried work, rise from being a minimum of 69% of formal salaried wages in 1988:1 to a maximum of 83% in 1990:3 before falling to 77% in 1993:4. It might be argued that if the value of the unobserved arguments remains constant over time that the initial narrowing of 14% represents evidence in support of the dualistic view: In a downturn, 8  displaced formal sector workers compete in the informal sector thereby expanding the differential between rigid formal sector and market clearing infornal sector earnings. But, it could also be argued that since the Pact was intended as a means of moderating salary increases, it capped them below market clearing during phase IS yielding a premium to working in the unregulated sectors whose earnings would not have been affected, The relative movements may also reflect the existence of skills or capital that are sector specific over the short run, without necessarily implying inferiority of a particular sector. Earnings of an informal skilled carpenter working on his own would rise dramatically in phase II with the increase in construction activity. Installed capital in a workshop or small store may also lead to self- employed workers preferring to earn subnormal profits in a recession if there is an expectation that in an upturn there will be supernormal profits. The reverse may also be true. A worker with substantial firm specific capital, or as Lucas (1978) postulated, little entrepreneurial ability, may ride out a wage stagnation in the formal sector should he doubt that he would earn more as a small businessman over the long run. In sum, neither the absolute levels of differentials, nor their movement across time yield conclusive evidence in favor of either the dualistic or integrative hypotheses. V. Sectoral Behavior and Patterns of Transition. However, the evolution of sector sizes and the patterns of worker mobility offers can. While offering some support to the standard dualistic view, the evidence strongly suggests that the informal sector may be a desirable destination for many workers. To begin, figure 1 shows that informal self- employment attains its maximum share of the work force at the peak of the recovery in 1990. 9  Informal salaried employment, while less dramatic, also locally peaks in 1990. The lag behind self- employment may be due to small firms hiring only after being established for a period. Contrary to the bulk of the literature on informality, it is the informal sector that is expanding as unemployment falls and growth picks up, while formal sector employment falls. Contract work shows an initial rise after moving out the recession, but overall behaves counter-cyclically as the dualistic view would suggest. However, until 1991, it also broadly tracks movements in formal sector employment, whose employment share also behaves counter-cyclically until 1991. The similarity of the behavior of contract and self-employed earnings suggests that both may represent equally successful types of entrepreneurship, but serve different clientele: contract workers with larger formal sector firms, the self-employed with a distinct market, perhaps more oriented toward services or smaller customers. The non-work sectors also show strong cyclical behavior. Unemployment reaches its minimum at the height of the expansion in 1990. Being out of the labor force appears to be a luxury, rather than the repository of discouraged workers, increasing as unemployment falls, and falling again as the economy worsens. The share of the labor force working as unpaid workers also behaves counter-cyclically suggesting that this subsector, roughly 2% of the workforce, may function as a safety net: when the economy worsens, perhaps paid family members, or those out of the labor force, work for free. There is no obvious explanation for the secular decline in the share of those in school. Second, the transition probabilities standardized by terminal sector size (P,/P) suggest that the increase in activity leads to greater, and relatively symmetrical, mobility between the formal and informal sectors. Mobility between formal salaried workers and the self-employed both rise as the economy strengthens going into 1990, but transitions into self-employment overall seem to rise more than the reverse (figure 2). The patterns of mobility are somewhat more obscure between the formal 10  salaried sector and the informal salaried and contract sectors. In the latter, there are broadly symmetrical rises in movements between secitors going into the upturn although movements into contract work become relatively more "fluid" after 1990 than the reverse transitions that dominated earlier. Movements between formal and informal salaried work, again, heighten in both directions in 1990 although any symmetries are less easy to detect. Overall, rather than a unidirectional flow back into fornal sector employment with economic recovery, there appears to be accelerated re- matching across all sectors, although, again, most clearly between the self-employed and formal sectors. Finally, statistical correlations ofintersectoral mobility and unemployment do not suggest that the primary function of the informal sector is to absorb displaced labor during downturns, although they do provide some evidence of rationing. T'he coefficients in table 4 are those from the standard exponential form of a multinomial logit. __ =-p' (1) ' i, where the vector ,Bj measures the degree to which an increase in open unemployment (ji) increases the probability of a worker going to sector j relative to the probability of staying in sector i for all sectors. Several findings at odds with the traditional dualistic view emerge from the table. First, movement into unemployment rises in downturns from all sectors, both formal and informal, and, with the exception of infornal salaried work, significantly. Movements out of unemployment decrease in downturns into all sectors of paid work, although significantly only for the formal salaried, and almost at the 10% level for the informal salaried. Both findings suggest that, contrary to the dualistic 11  hypothesis, you can "lose" ajob in the informal sector and it is not obviously easier to find informal work in a downturn. The one anomaly is informal salaried work where from OLF, school, and unpaid work, there is evidence of counter-cyclical entry, although only significantly from OLF. Looking at transitions within paid work, several relations are also striking. First, while it is true that movements into formal salaried work fall as unemployment rises from every sector, this is also true for self-employment, confirming the procyclical behavior of sector size found earlier. Only for the transition from formal salaried to contract work is there a significant positive coefficient. Even informal salaried work shows a lower probability of entering from self-employment, and only a very insignificant positive relationship from formal salaried employment. Combined with the previous finding that all sectors contribute to unemployment, it is difficult to sustain the view that overall, the informal sectors are the safety net absorbing dismissed formal sector labor. The increased relative movement from both school and OLF into unemployment suggests that, as much as workers losing positions, unemployment is comprised of those in school and OLF who begin to search for employment to augment their family incomes. This motivation would also explain why, from every sector, except OLF, workers are less likely to return to school, significantly so for the unemployed, unpaid, self-employed and informal salaried. The results are consistent with the findings of Revenga and Riboud (1993) who argue that most labor market adjustments in Mexico occur through dramatic falls in earnings in all sectors and thus relatively little labor is actually shed. However, the fall in real incomes does drive those out of the labor force to actively seek other sources of income, thereby driving up unemployment. 12  Is there evidence of queuing? Although the logit specification captures gross tendencies during cyclical downturns for movements among the four modes of work, we can test explicitly for rationing by recasting Abowd and Farber's (1982) test for union-induced segmentation in a time series context."4 At a given moment in time, a worker in current sector "c" will desire to switch to an alternate sector "a" if he expects a gain in utility which is a function both of earnings and of non-wage benefits of working in the sector such as those discussed in section IV. Since many ofthese factors-- independence, medical insurance, or other labor protections-- are not pegged to the wages, these effects are assumed to be constant across time leaving the differential rates of growth of sectoral earnings as determining the relative desirability of each sector. Together, the desire to enter the alternate sector is Irt = aaWat acWc - YFac + 2 ,:t @a~c)(2) F-lt =(Eat - ect)() Where yF. is the unchanging utility arising from differing non-wage benefits between sectors and W,, W, hat are the forecasted values from the standard Mincerian earning equation Wp=t =PCt + ect (3) Wat = 4Jat + Eat where X is a vector of worker characteristics. This permits proxying for unobserved earnings in the alternate sector and avoiding-bias in the estimates of aC.` The time subscript on 3 reflects the fact that across time similar worker characteristics may lead to different earnings in each sector. This may 14 See Dickens and Lang (1985) for an alternate test of queuing using endogenous switching models. 15 Replacing only the unobserved alternate wage with the forecasted variable, but using the realized current wage has the potential to bias alpha since e, is likely t. be correlated with W,. For this reason, both wages are replaced with the forecasted values. 13  result from any number of other structural or temporal factors including the level of economic activity. This equation, however, yields only the unobserved desire to move, which may be thwarted by any factor causing queuing or that affects the probability of being offered ajob in the new sector once it is desired to seek it. Since the unemployed are by definition, those looking or unable to find jobs, this probability of being offered a job is a function of the state of the labor market, ,u. y2, = X ,+ R 2t (4) Particularly during cyclical downturns, the informal sector is generally posited to serve as the reserve army of those unable, although willing, to take a job in the formnal sector. The unobserved latent variable, Y2 determines whether or not a worker is selected from the queue, a queue which may be of zero length in a period of high economic activity. The probability that an individual will be observed in, for example, the alternate sector is the probability that the worker desires to work in that sector (is in the queue) and is selected: Pae = P[