Skip to navigation

Skip to main content

Research & Seminars

contact us

Employment Trends in Mexico: Reversing a 15-Year Loss? -- Agustín Escobar Latapí



Agustín Escobar Latapí

1.- INTRODUCTION................. 1

II.- The period under consideration................. 3

III.- The demographics of labor absorption....................... 4

IV.- Social Security employment.................... 7

V.- Impact of formal job growth on composition of the total labor force.......... 8

VI.- Final remarks.. 9

References..................... 11

Starting in 1976, the Mexican economy showed signs that it was not able to create the formal[1] jobs needed by ever-larger cohorts of job seekers. This job absorption gap became evident for all to see (or experience) after 1982, when formal jobs stagnated and record numbers of persons entered informal occupations or left for the United States. But this is changing. This paper explores the significance of major improvements in formal job creation rates from 1996 to 2000.

Labor absorption is only one of Mexico's two main employment problems. The other is income from labor, which has also dropped by more than half in real terms since 1981. But my position in this paper is that an economy that generates sufficient formal jobs is much better able to provide well being to its population, even if salaries do not improve dramatically, than one that does not. Formal jobs pay the taxes that fund the health care and pension systems through Social Security and other government services through employment taxes. They allow individuals to retire from the labor force at a reasonable age, thus diminishing pressure on employment and old-age poverty (a phenomenon Mexico should plan against). They fuel higher national savings rates through now-privatized retirement accounts. Lastly, formal employment generation points at labor demand from modern, formal enterprises, characterized by higher-than average productivity, and thus better able, over the medium term, to increase wages in real terms.

This thesis is meant as a partial contribution to the discussion of the current and future size and nature of Mexico – U.S. emigration. Formal job generation by itself is not enough to change Mexico – U.S. emigration. Other factors such as demand in the U.S., the wage gap, social networks and a culture of emigration also matter. But it is, in my view, a major contributing factor to slow this flow both directly, because formal jobs play a retaining role, and indirectly, because formal jobs should lead to general income improvement.

Aside from "imperfections" and segmentation derived from the formal – informal split and from regional and inter firm heterogeneity, trends in the Mexican labor market may not seem to conform to usual market behavior. Institutional changes and intense pressures from the government, the Bank of Mexico and international financial agencies have contributed to maintain labor income at low levels via, among other instruments, high interest rates that stifle productive investment and semi-mandatory limits on wage increases. It is worth noting, however, that these pressures, though overwhelmingly real, have succeeded because of the growing size of Mexico's labor reserve in agriculture and in the informal economy. A changing balance in labor supply and demand, therefore, may lead to changes in wage levels in spite of these pressures, provided growth comes mostly from higher-than average productivity enterprises. The two economic factors most closely related to labor emigration (the job gap and the wage gap), therefore, are closely linked, and change in one could lead to change in the other.

This trend – based diagnosis is not one of unqualified optimism. First, as stated, labor income levels have recovered only slightly, and in a few occupations only, although the general income structure shows some improvement. Second, towards the end of the period formal manufacturing job growth depends more on maquiladora employment growth, with non-maquiladora manufacturing employment losing some momentum[2]. Since domestic consumption is growing, the reason would seem to be a slackening of investment in non-maquiladora manufacturing, probably related to rising imports (a consequence of an overvalued peso) and to political uncertainty during the first months of the year 2000. It is significant to note that, until June 2000, maquiladora employment growth continues unabated[3], in spite of the overvaluation of the currency. In other words, rising comparative costs of Mexican labor had not hurt in-bond employment. The weakness in this growth seems more closely related to internal market dynamics (a larger share of imports in total inputs and consumption). Although a more detailed economic analysis of Mexican industry is called for, the implication is that non-maquiladora manufacturing employment needs to regain its previous rate of growth, possibly through increased competitiveness gained through a lower peso value and/or productivity gains.

The third and final cautionary note has to do with the regional restructuring of Mexican economy and society. Most formal employment growth is taking place in the West, the North and along the U. S. Border. This trend is accelerating the regional demographic and migration shifts first observed in 1980-90 (Corona 1993, Escobar, Bean and Weintraub 1999). These changes will, in the medium term, create a large population in cities with easy access to the U.S. border.

The current analysis is based on the National Employment Survey, the Urban Employment Survey and Ministry of Labor Statistics. My interest and part of the analysis, however, stem from previous qualitative and quantitative work on the Guadalajara, the national, the urban and other labor markets in Mexico.

Some limited references shall be made to the previous period (1981-1995), in order to highlight the changes now taking place. A lengthier treatment of this period (and its sub-periods) is available elsewhere (Escobar, Bean and Weintraub 1999, Cortés 1997).

II.- The period under consideration.
In labor and population analyses, it is usual to work with pre-defined, usually decennial, periods. Recent Mexican history, however, has clearly defined periods that have nothing to do with decennial analysis. Roughly speaking, there are

A period of marked instability and no job generation from 1982 to 1987, followed by a short crisis.

A period of growth (1988-1991) that slows gradually and stops by 1992, with stagnation until 1994, followed by a major crisis in 1995.

Finally, a period of rapid growth in formal employment and in the national economy from 1996 to the year 2000. This last period sees faster and larger job growth than any other in absolute terms, and is the best in relative terms since the end of the 1970's.

This last period is one of recovery. Specific levels depend on the variable used, but in some cases this recovery places today's labor market structure and composition back at levels of 1990-1, or 1985-6. By no means have the main variables recovered to their levels of the early eighties.

Employment dynamics during the two first periods matter, because they were the basis for a very pessimistic outlook many authors, myself included, held until recently. Stated briefly (Escobar and González de la Rocha 1995, Escobar 1997), during the first phase (1982-1987) the weak growth rates of formal employment led to a more – than - proportional rise in unprotected, informal employment (González de la Rocha and Escobar 1986, González de la Rocha 1994, García and Oliveira 1994). Severe income drops from informal occupations followed, because of an excess of job seekers. This in turn led, between 1991 and 1994, to the creation of a large group of discouraged job seekers. In other words, towards 1994 Mexico was close to the rise of large-scale chronic unemployment and social exclusion. By this term we referred to the growing inability of kin and friendship networks to provide the kind of help that would place people in jobs and integrate them into family and community life. Since State Welfare mechanisms are extremely weak, poverty was worsening, both because labor income was falling and because growing sectors of would-be workers had no income.

My view in this paper is that Mexico avoided large-scale problems of social exclusion thanks to the revitalization of formal employment growth, although this growth is not distributed evenly throughout Mexico and this creates significant regions where social exclusion is a problem.

III.- The demographics of labor absorption
Two main factors define labor absorption rates: employment and the population of working age. Based on projections of the population of working age and sustained but modest economic and job growth rates, we (Escobar, Martin, Donato and López 1998, Escobar and Martin 1999) forecast that Mexico would reach a point in which formal job growth equaled or surpassed the growth of the population of working aged towards 2004 – 2006. Our general perception still seems correct since Mexico is coming closer to a point of equilibrium in demographic and job growth rates. But it needs to be refined on the basis of 1) changes observed in 1996 – 2000, and 2) strict projections of the labor force carried out by CONAPO (2000).

According to CONAPO, job demand was 1.16 million per year during 1995 – 2000, and will be 1.09 million per year during 2000 – 2005. These figures are higher than previous estimates (Escobar, Martin, Donato and López 1998, Escobar and Martin 1999) based on the projection of growth of the population of working age. This is because they forecast rapid growth of participation rates (mostly women's), as observed between 1981 and 1995, and also because workers are remaining in the labor force longer because they cannot afford to retire. Small numbers of labor force withdrawals reduce vacancies and increase demand. Higher participation rates and longer working lives result from falling real incomes and families' need to pool more incomes in order to get by, and from the weakening of social security mechanisms, itself the outcome of rising informality in employment between 1981 and 1995.

Participation rates, however, rose much less after 1995, as shown in Tables 1 and 2. Women's participation rates are rising slowly (0.4 % in 4 years), as are men's (1% in 4 years and close to postindustrial levels)[4]. The participation rates of the population aged 12 – 45, which grew continually from 1981 to 1995, are stagnant or falling.

Less growth in participation rates equals less growth in job demand. On the basis of the above considerations, growth in demand would be lower than the CONAPO figures. Can this slower growth of participation rates be expected to last or to become stabilized?

During 1981 – 1995, growth in demand was the outcome of rapid growth in the population of working age and increasing participation rates. Participation rates increased partly to offset losses in labor income. During the eighties, this rise in participation rates was particularly strong among married women with children and little education (González and Escobar 1986, García and Oliveira 1994). They were partly successful in diminishing income falls. According to Cortés (1997), the increase in inequality observed between 1989 and 1994 would have worsened considerably had households not increased their earners. One conclusion therefore is that labor income should remain stable or rise, for this part of "forced demand" to remain stable or fall[5].

Secondly, participation rates grew because young women are having fewer children and they are having them later. Fertility rates fell quickly over the past 20 years. The tendency is still for them to fall, but less rapidly than in the past. Their participation rates are no longer rising, and they do not therefore play a major role in rising demand.

Finally, the participation rates of the 45 + age group are still rising. In this group, both men's and women's rising participation rates matter. But women seem to be working beyond their forties because of low total household incomes. This group may withdraw from employment if total household incomes rise, either through the jobs of other household members or through an improvement in pension payments. This last depends on allocating special pension funds for persons who worked under the old "pooled" pension system[6]. An indication of the level of need of these older households is evident in data from ENIGH 1996 in Corona (2000). This study found that a large portion of the households receiving remittances from relatives in the United States are comprised of elderly persons and children, with a gap in the middle generation, precisely those in the United States.

In sum, job demand will grow strongly because of high population growth rates in the seventies and early eighties, which result in large numbers of young people entering the labor force, and because of rising participation rates among older persons. But the slower expansion of participation rates during the past five years means total demand should be slightly lower than the CONAPO figures (below one million). An additional factor contributing to the definition of job demand levels will be emigration, which is "absorbing" at least 230,000 persons per year. If emigration grows in the short term, this will have direct consequences on diminished job demand. The current U.S. job surplus (growth above demand levels), together with the large worker reserve in informal employment accumulated in Mexico between 1981 and 1995, may indeed result in more emigration over the short term.

Finally, since most participation rate growth was counter-cyclical, that is, aimed at counteracting income falls, formal job growth and income growth should help slow this growth further. However, over the longer term, if this favorable development prevails, we could see a new, pro-cyclical rise in participation rates, arising from the expansion of more attractive jobs.

IV.- Social Security employment.
The second factor in absorption levels is job creation. From 1981 and until 1995, the weak performance of formal jobs meant most jobs were created either by persons themselves, or by small-scale enterprises. Both the government, which was a major mover in the modernization of employment in Mexico from 1940 to 1980, and the corporate sector absorbed few workers. An expansion of personal services employment, informalization, unpaid workers, and a falling proportion of waged and salaried workers marked this period.

Since January 1996, however, Mexico has created an average of 700,000 IMSS (Mexican Institute of Social Security)-registered jobs per year, as opposed to 300,00 between 1990 and 1994, and 200,000 between 1990 and 1995 (a year of absolute formal job loss). Formal employment growth was spread fairly evenly over most branches of industry. Table 3 shows how this growth is distributed, as well as the growth rates of employment in each major industry. In absolute terms, in the five years starting in August 1995 and ending in July 2000, an average of 329,000 jobs were created annually in manufacturing (including maquiladora employment), 145,000 in construction, 117,000 in commerce, 147,000 in personal and producer services, and smaller industries also expanded[7].

The single most important change in total IMSS affiliation between 1984 and 1995 was the falling participation of manufacturing employment in the total. Since 1995 this is the sector expanding most rapidly, even if maquiladora employment growth is subtracted. This recovery is high and continuous. Manufacturing employment was the most formal of all industries in private sector employment until the early eighties, comprising 35% of all IMSS jobs. After 1988, it rapidly lost ground to other branches of the economy, as firms streamlined or failed. By 1994 it accounted for only 28% of all IMSS employment. The rapid growth of maquiladora and non-maquiladora employment growth since late 1995 means it is now back to the relative levels of the mid-eighties.

In summary, formal employment growth between August 1995 and July 2000 is significantly higher than average labor force growth during the past 20 years. This growth has an observable impact on the composition of the total labor force. Towards the end of the period (August 1999) there seems to be a tendency for formal employment growth to slow somewhat, but the usual end-of year peak may return growth to the levels typical of this fast-growth period, provided no major changes take place in inflation-related policies.

V.- Impact of formal job growth on composition of the total labor force.
If estimated labor force growth (hovering at about one million per year) and reported formal job growth are correct, then the absorption gap has narrowed very significantly for the past five years. But both results could be mistaken. They could be artifacts of erroneous assumptions or the result of measurement error (the population looking for jobs could be underestimated, or "permanent" affiliation to the IMSS could be overestimated because of failure to report lay-offs or similar events).

It is therefore necessary to test the accuracy of both population and job calculations on an independent basis. A narrower absorption gap should result in the rising proportion of waged positions in the overall occupational structure, and a relative fall in other categories. This is indeed the case.

The data in the last two tables come from Mexico's current urban employment survey (or ENEU), that is carried out every quarter in 47 cities. In general, most assessments in Mexico and abroad have concluded this is a very reliable survey.

Table 4 depicts the shift in the composition of the total labor force according to "position", i.e. the relationship of the person to employment. This is classified into 1) waged, salaried, piece-rate and commission work, 2) employers, 3) own account work, 4) unpaid work (usually for a family member) and "other". This table shows a loss of importance of the first category from 1989 and until 1995, and then a rapid increase. The other category changing markedly is that of unpaid workers, which rose from 1989 and until 1995 and has since fallen very rapidly. Other positions show less change.

Finally, table 5 lists the proportion of all jobs falling under the major categories of "underemployment", which should diminish markedly after 1995. This includes workers reporting less than 15 hours of work per week, less than 35 hours, persons working in establishments employing 5 or fewer persons, wage workers without benefits and, finally, workers earning less than the minimum wage. As can be seen, these categories show mostly a worsening from 1987 to December 1995, and then they all diminish in relative terms (some may have grown in absolute terms, since the labor force is growing rapidly). The percentage of all the labor force working in small establishments is the one showing least change. The last column (workers earning under the minimum), however, should be read with caution. The purchasing power of the minimum wage fell during the Salinas years (1988-1994). Under Zedillo, it has not fallen any further, but is still very low (about 4 dollars per day). This last column, therefore, does not show a real improvement from 1987 to 1995, and only a marginal one since.

The employment survey, therefore, does show significant changes in the overall composition of the Mexican labor force from 1995 and until the year 2000. Still, this does not mean that IMSS employment has generalized. Total IMSS-affiliated employment totals 12.5 million, and a further 3 million workers are affiliated at the public sector ISSSTE. The urban Mexican labor force comprises approximately 30 million persons. Roughly half of all Mexican urban employment, therefore, is still not "formal" by this definition. Over 60% of all workers would still be classified as underemployed if measured by their ability to purchase the minimum "basket" of goods and services. In other words, most Mexican urban jobs provide incomes below the poverty line.

VI.- Final remarks.
Mexico's crisis of 1982, and ensuing instability, restructuring and recurrent crises, not only meant job-related incomes fell, it also led to a stagnation of formal jobs, a substantial widening of the gap between population and formal employment growth, soaring informality and emigration to the United States. Informal employment rose 80% from 1980 to 1988. The Mexico-born population of the United States grew at a rate of 6.7% per year during the 1980's. My perception towards 1993-4 was that Mexico would soon enter a period in which social exclusion, arising from the inability of workers to sustain the non-workers, would be the major new trend in employment dynamics.

From 1996 and until the year 2000, however, Mexico has experienced a marked recovery in the creation of formal jobs which has already exerted some changes for the better in the overall composition of the Mexican workforce. This has not extended to improvements in real incomes, but is very significant in the sense that labor absorption has narrowed the gap between population and job growth. The quality and incomes of these formal jobs are not comparable to those of the 1970's. Both job stability and real incomes are lower, which means a number of these formal jobs can still be classified as precarious. But a renewal of formal employment growth implies a strengthening of job-related and social security institutions as well as net increases in domestic investment. Formal jobs in cities away from the border, on the other hand, have been shown to retain workers in Mexico.

In spite of this marked improvement, a number of difficulties are evident. First, this improvement is limited to urban areas. Second, most formal job growth has taken place in the West, North, and northern border cities, thus bypassing the Center (which still harbors Mexico's largest urban concentration) and the South. Not surprisingly, most "new" Mexican migration to the United States comes from Southern Mexico.

And finally, the future of this trend is uncertain. While credit for this expansion should probably be shared by the impetus created by incorporation into NAFTA, sustained U.S. economic growth and domestic economic discipline, large numbers of corporate employers are questioning the expansion of their operations in Mexico for various reasons, among which the most frequently mentioned are:

1) A serious yearly uncertainty concerning the fiscal regime, often entailing difficult negotiations with authorities;

2) The recent overvaluation of the peso, which has driven up wages and salaries (in dollar terms) and led to rising imports which compete with the production of domestic and foreign firms established in Mexico;

3) The scarcity and high cost of domestic credit, both for the enterprises themselves and for their middle-class employees, which blocks further investment, particularly by middle-sized enterprises;

4) Lack or inadequacy of education, health and other social security institutions, which entails a higher real cost of living than apparent in the prices of goods and services;

5) An lastly, a visible stress or insufficiency in urban service infrastructure in the cities absorbing this rapid employment growth (water and sanitation, transport, electricity supply, etc.).

There are no reasons to expect a sudden end to the rapid growth of formal employment in Mexico. But these stresses need to be addressed in order for this growth to continue and to become a true process of development. Nevertheless, Mexican emigration to the United States would have been larger, and the migrants poorer, had previous trends not been reversed, at least partly, for the past five years.


CONAPO (Consejo Nacional de Población) 2000 La situación demográfica de México 2000 Mexico City: Author

Corona, Rodolfo 1993 "Migración permanente interestatal e internacional. 1950 – 1990" Comercio Exterior 43 (8).

Corona, Rodolfo 2000 "Monto y uso de las remesas en México" paper delivered at "Diálogo Migratorio México – Estados Unidos" held at Secretaría de Relaciones Exteriores, Mexico City, October 18.

Cortés Cáceres, Fernando 1997 La distribución del ingreso en México en épocas de reforma y estabilización y reforma económica Ph.D. Dissertation, CIESAS Occidente, Guadalajara.

Escobar Latapí, Agustín 1990 "Estado, orden político e informalidad" Nueva Antropología 37.

Escobar Latapí, Agustín 1991a "De la informalidad al vacío" Antropológicas 9.

Escobar Latapí, Agustín 1991b "El nuevo Estado mexicano y el trabajo informal" en Alonso, Jorge, Alberto Aziz y Jaime Tamayo (Eds.) El nuevo Estado mexicano. Tomo I. Economía. Mexico City: Nueva Imagen.

Escobar Latapí, Agustín 1997 "El final de la Transición. México, 1982- 1996" paper delivered at the 3rd Conference of the Asociación Latinoamericana de Sociología del Trabajo, Sao Paulo, September

Escobar Latapí, Agustín and Mercedes González de la Rocha 1995 "Crisis, Restructuring and Urban Poverty in Mexico" Environment and Urbanization, 7(2).

Escobar Latapí, Agustín, Frank D. Bean and Sydney Weintraub 1999 "The Dynamics of Mexican Emigration" in Appleyard, Reginald (Ed.) Emigration Dynamics in Developing Countries Vol. III. London: Ashgate.

Escobar, Agustín, Philip Martin, Katharine Donato and Gustavo López Castro "Factors that influence migration" in Commission for Immigration Reform and Secretaría de Relaciones Exteriores (Eds.) Migration Between Mexico and the United States. Binational Study Austin: Authors.

García, Brígida and Orlandina de Oliveira 1994 Trabajo y vida familiar en México Mexico City: El Colegio de México.

González de la Rocha, Mercedes and Agustín Escobar Latapí 1986 "Crisis y adaptación: hogares de Guadalajara" paper delivered at the 3rd Conference of the Sociedad Mexicana de Demografía Mexico City: El Colegio de México, November.

INEGI (Instituto Nacional de Estadística, Geografía e Informática) 2000 BIE (Banco de Información Económica). On line at WWW.INEGI.GOB.MX



Age groups

Period (quarter)
45 y más

Participation rates




Annual change

1987/01 – 1999/04

1987/01 – 1995/04

1996/01 – 1999/04

Source: INEGI: ENEU, various dates, in BIE (Banco de Información Económica)



Period (quarter)




Annual change

1987 – 1999

1987 – 1995

1996 – 1999

Source: INEGI: ENEU, various dates, in BIE (Banco de Información Económica)



Year and month
Private Services
Community and Social Services

Distribution (%)

Enero 1984

Dic. 1995

Julio 2000

Annual growth rate

1984 –1995

1996 – 2000 (julio)

Source: IMSS statistical office, at BIE (Banco de Información Económica). Excludes electricity and Construction.

a Not related to employment.



Year and quarter
Waged, salaried, com., and piece rate workers
Own account




Absolute change in %

1989 – 1995/03

1995/03 – 2000/02

Relative Change in %

1989 – 1995/03

1995/03 – 2000/02

Source: INEGI: ENEU, various dates, in BIE (Banco de Información Económica)


THE UNDEREMPLOYED: 1987 – 1995 –1999.

(% of total)

Working less than 15 hrs. Per week a/
Working less than 35 hrs. Per week c/
Working in 1 – 5 employee enterprises b/
Wage earners without benefits d/
Earning less than minimum wage




Annual relative change

Crec. Anual ene 87 dic 1995

Crec. Anual ene 96 dic 1999

Source: INEGI: ENEU, various dates, in BIE (Banco de Información Económica)


[1] I have written at length about the variable meanings and the difficulties involved in the use of the formal-informal dichotomy (Escobar 1986ª, 1986b, 1990, 1991ª, 1991b). In this text, however, I will refer to formal jobs as those registered at the Mexican Institute of Social Security (IMSS), which are private sector jobs. Total Social Security job coverage, however, is larger, because government employment is registered at a special public sector security institute, the ISSSTE, comprising close to 3 million employees. Coverage through this institute is growing far less than IMSS coverage.

[2] The various sources conflict with one another on this point. Ministry of labor Statistics coincide with IMSS statistics concerning maquiladora growth, but they show far lower rates of non-maquiladora employment growth than IMSS statistical sources for 1999. This could be due to the fact that Ministry of labor statistics refer only to contracts signed in so-called "federal jurisdiction" enterprises. If this is correct, and formal manufacturing employment is still growing rapidly among non-federal jurisdiction enterprises, then both the assessment and the outlook are better than this paragraph states.

[3] The rate of growth is falling, but absolute growth is still within the range of the last 4 years.

[4] CONAPO foresees little room for further increases in men's participation rates, provided especially that schooling is rising. I agree.

[5] I call it forced demand because these women were forced to join the labor market when their household income dropped. They have few qualifications and are not sought after by employers. During the eighties and early nineties, therefore, they entered mostly informal occupations which entailed long working hours and very low incomes.

[6] In October 2000, President Zedillo announced that the Federal government would cover the ISSSTE's deficit in pension plan payments. This deficit was 800 million dollars. The ISSSTE is in a worse stiuation than the IMSS, because affiliation is not growing, but IMSS faces a similar problem because new workers do not contribute to the old pension fund. Instead, they contribute to their IRAs.

[7] Total additional IMSS enrolment is larger because the "other" column in Table 3 refers to non-employment affiliation and should not be counted as part of this process of job creation.