June 1998
Gender Earnings Differentials in the Microenterprise Sector:
Evidence from Rural and Urban Mexico
Susana M. Sánchez
The World Bank
Latin America Region
Finance, Private Sector and Infrastructure
Washington, DC20433
Tel. (202) 473-1856
Fax (202) 522-2106
Email:
JEL Classification:
Keywords: Microenterprises; Women; Rural; Mexico
______
I thank William F. Maloney, José A. Pagán, and Rodrigo A. Chaves for helpful discussions and Maria Correia, Wendy V. Cunningham, Arianna Legovini, Yira J. Mascaró, and Mark Schreiner for comments. The views and interpretations expressed in this article are solely of the author. They do not represent those of the World Bank or its member countries.
Gender Earnings Differentials in the Microenterprise Sector
Evidence from Rural and Urban Mexico
by Susana M. Sánchez
Abstract
This study examines the sources of male-female earnings differentials in Mexico’s microenterprise sector. Using the 1994 Survey of Rural Entrepreneurs and Financial Services (conducted in the rural regions of Puebla, Guanajato, and Veracruz) and the 1992 National Survey of Urban Microenterprises, I found that female-headed microenterprises in rural and urban areas earn, respectively, 64% and 50% of less than male-headed microenterprises. About 35% (42% in rural areas) of the earnings gap in urban areas can be explained by differences in productive characteristics, whereas 62% (59% in rural areas) can be attributed to structural productive factors, such as household constraints, productivity, and access to education and credit services. Furthermore, a small share of the gender earnings gap is explained by gender differences in sectoral distributions. From a public policy perspective, the results suggest that policies attempting to reduce earnings disparities should focus on addressing gender differences in productive characteristics and structural factors across economic sectors because efforts to reallocate women’s microenterprises to other economic sectors would have a limited impact on increasing their earnings.
1.Introduction
Women are active participants in the microenterprise sector in both developed and less developed countries accounting for about two-thirds of microenterprises in some countries (Rhyne and Holt, 1994; Cohen, 1996; Evans and Leighton, 1989). Despite the role of women in the microenterprise sector, women’s microenterpises earn substantially less than those of men’s, are concentrated in the lower end of the size spectrum, have different sectoral distributions than that of men’s, and are more likely to work from their dwellings (Rhyne and Holt, 1994; Cohen, 1996). These gender differences in entrepreneurial patterns have attracted the attention of policymakers and international organizations, especially because women, who are likely to be affected by poverty, have increased their participation in the microenterprise sector (World bank, 1994a). Although most countries have targeted programs devoted to microenterprise development as well as the promotion of small-firm formation (Wilson and Adams, 1994)[1], only in recent years policymakers have started to incorporate in their microenterprise development agenda that the male and female segment of the microenterprise sector could face different constraints to growth and profitability.
This study examines the sources of earnings differentials between male- and female-headed microenterprises in Mexico’s rural and urban sectors. In particular, it examines the roles of gender differences in size of operations, returns to productive factors, and sectoral distribution. In the case of Mexico’s microenterprise sector, female owners of microenterprises account for 26-44% of microenterprise, but their earnings (profits) are 36-50% of what their male counterparts earn. The importance of the microenterprise sector as a source of household income and employment makes understanding gender disparities in the sector relevant for effective public policy formulation on microenterprise development, poverty, income distribution, and labor market efficiency.
The determinants of gender earnings differentials in the microenterprise sector are examined using the 1994 Survey of Rural Entrepreneurs and Financial Services (SREFS: Encuesta Regional de Servicios Financieros a Unidades de Producción Rural) and the 1992 National Survey of Urban Microenterprises (ENAMIN: Encuesta Nacional de Micronegocios), both conducted by the Mexican National Statistical Institute. The rural survey selected 1,944 households in localities with a population of 1,000 to 20,000 inhabitants in the states of Guanajuato, Puebla and Veracruz. All owners of microenterprises (about 504) were interviewed. The urban survey interviewed 9,036 microenterprise owners with less than six employees (15 in manufacturing) that were selected from the last quarter of the 1991 National Urban Employment Survey. The microenterprise sector includes own-account workers, partners, and employers. Both surveys examine the use of credit services, capital structure, employment patterns, costs, and sales, among other issues.
The main results of this study are that: (a) gender differences in productive characteristics (including owners and businesses) are more important in explaining gender earnings differentials in rural than in urban settings; and (b) efforts to reallocate female-headed microenterprises to other economic sectors could have a limited impact on increasing women’s earnings. Hence, policies should focus on addressing gender differences in productive characteristics and structural factors across economic sectors.
The study is organized as follows. Section 2 presents a profile of microenterprises and examines gender differences in size, choice of economic sector, place of operations, and other gender-specific issues. Section 3 investigates observable and unobservable factors that could determine gender earnings differentials in the microenterprise sector. Section 4 presents the empirical methodology used to evaluate the role of inter- and intra-sectoral factors in explaining gender earnings disparities. Section 5 estimates earnings equations and evaluates the role of differences in endowments. The last section provides some concluding remarks and discusses public policy implications of the results.
2.Profile of Microenterprises in Mexico
In Mexico, female participation in the microenterprise sector has risen dramatically during the first half of the 1990s. According to the Mexican National Employment Surveys, in localities with less than 100,000 inhabitants female representation increased from 36% in 1991 to 44% in 1995, whereas in the rest of the country it only increased from 26% to 29%. The urban and rural surveys show that women own about 24% and 39% of the microenterprises in each area. While women figure prominently in the microenterprise sector, female-headed microenterprises account for 16% of microenterprise earnings in urban areas (27% in rural areas) and their earnings are 50% (36% in rural areas) as much as their male counterparts across all economic sectors. In the context of microenterprises, women’s businesses tend to have lower earnings than that of men as indicated by their smaller share in earnings. Table 1 and Table 2 present descriptive indicators of microenterprise activities for urban and rural microenterprises. The most important gender-related differences are associated to the following variables:
Number of workers. Most microenterprises are operated by the owner in both urban and rural areas (66% and 53%, respectively). In urban areas, the average number of workers (including the owner) in male-headed microenterprises is 1.65 and ranges from a high of 2 workers in manufacturing to a low of 1.47 in the service sector. The average female-headed microenterprise shows little variation in the number of workers across economic sectors (about 1.52). In rural areas, microenterprises are a stronger source of employment generation as both men’s and women’s businesses employ, on average, 2.11 and 1.59 workers, respectively.
Home-based Operations. Most female-owned microenterprises operate from home, although substantial urban-rural differences exist. In the rural areas studied, about 44% of female-headed microenterprises (21% of male-headed) are located at home, while in urban areas 33% (13%) of these businesses are home-based. The business location seems to be closely linked to the sector of economic activity, and thus with earnings. The manufacturing sector, which has the lowest female-to-male earnings ratio (Figure 1), has also a disproportionate share of home-based microenterprises among female owners in both rural and urban areas (88% and 81%, respectively).
Years of Operation. On average, female-headed microenterprises have been in business for 6.7 years in both rural and urban areas. In contrast, male-headed microenterprises are relatively older with 10.8 and 9.2 years of operation in rural and urban areas, respectively. The survival rate of microenterprises—particularly those owned by women— is lower in rural areas. In urban areas, only 36% of female-headed microenterprises (33% in rural areas) have been in business for more than five years, compared to 51% of male-headed microenterprises (50% in rural areas).
Hours Worked. In urban areas, women’s microenterprises operate 7 hours less per week than those of men. However, this gender disparity in the hours of operation greatly differs across economic sectors. In manufacturing, women’s microenterprises operate 20.2 hours less than men’s, but only 2.7 hours less in trade. [2]
Plans to expand. In urban areas, about 67% of microenterprises will not change their business/plant size in the near future. Interestingly, expansion plans appear to be unrelated to the owner’s gender, with 21% (23%) of women’s (men’s) microenterprises having plans to expand their activities by either making new investments, increasing the number of employees, or developing new products. Only 11% of microenterprises reported that they would close their operations in the near future.
Formality Indicators. Overall, microenterprises in both rural and urban areas comply with tax legislation: about 42% of microenterprises are registered with the Ministry of Finance. Registration with tax authorities seems to be related with the sector of economic activity as well as the owner’s gender. For Indeed, gender differences in tax registration is greater in manufacturing reaching 22 and 17 percentage points in rural and urban areas, correspondingly. In this sector, female-headed microenterprises in manufacturing report the lowest registration rates: about 15% and 12% in urban and rural areas. Although compliance with tax legislation seems high, fewer microenterprises are associated to chambers of activity and/or gremiales: 23% and 17% in urban and rural areas, respectively.
Educational Attainment. Gender disparities in educational attainment are greater in the rural areas studied than in urban areas. About 25% (11% in urban areas) of female rural entrepreneurs have not received formal education, compare to 11% (7% in urban areas) of their male counterparts. This indicates that urban-rural differences are substantially greater among female entrepreneurs.
3.Gender Earnings Differentials in the Microenterprise Sector
Earnings differentials between male- and female-headed microenterprises are likely to be influenced by observable and unobservable factors. Within each economic sector, (the log of) microenterprise earnings, I, can be seen as a function of observable productive characteristics, X, (number of employees, capital, level of formality, firm life, human capital variables, market conditions, household demographics, and so on), the market return to each of these characteristics, , and a residual, , capturing unexplained individual-specific effects; namely,
(1)
An examination of (1) suggest that the average earnings of self-employed women and men may differ because of differences in their levels of productive characteristics, X, for example, capital, and/or they may vary due to gender differences in the return for each added unit of capital, . Thus, overall or within each sector, gender differences in earnings can be explained by endowment disparities and/or differences in returns to productive characteristics. The latter will be referred as structural productive factors. The “structural” component encompasses various unobservable variables and/or structural factors that are detailed below.
First, unobservable variables, such as entrepreneurial ability, could be differently distributed between male and female entrepreneurs and would tend to show up statistically as differences in returns to observed characteristics or factors. Variance in entrepreneurial abilities may result in differential returns to human and physical capital among possible types of microenterprises (Maloney and Cunningham, 1997). Moreover, differences in entrepreneurial abilities will determine the success of microenterprises, and thus, their size (Lucas, 1978). If self-employed women have lower levels of entrepreneurial ability, then their business will be smaller and their earnings lower.
Second, labor distortions and/or gender-based household responsibilities may also influence differing returns to measurable characteristics or factors. For example, if women are unable to get salaried work because of employer discrimination or because of household responsibilities, then those who are not entrepreneurially talented might be forced into self-employment and will, thus, depress the mean earnings of self-employed women. Indeed, using the same rural survey in Mexico, Pagán and Sánchez (1998) found that structural supply and demand factors, such as differences in employment preferences, employer discriminatory behavior, or household constraints, induce rural women into self-employment.
Third, the need to balance home and market responsibilities is a major constraint on productivity and earnings (Blau, 1998; World Bank, 1994a). Family responsibilities may have a direct effect on earnings by reducing the amount of energy and effort available to devote to income generating activities (Hersch, 1991). Because household responsibilities generally have predetermined schedules, women’s microenterprises are more likely to be home-based which allow them greater flexibility in hours worked, have lower growth rates and earnings than market-based firms. Moreover, unobservable job characteristics, such as working schedules, may justify lower earnings as a compensating differential for favorable working conditions (to combine family and job responsibilities). This indicates that we should also find differing returns to factors between men’s and women’s microenterprises.
Fourth, credit constraints might be an important impediment to growth among microentrepreneurs (Evans and Jovanovic, 1989; Rhyne and Holt, 1994). The presence of credit constraints would affect microentrepreneurs’ abilities to smooth consumption over time and undertake profitable investments. The investment behavior of microenterprises—and, therefore, their growth potential—could be restrained by limited access to credit services (Tybout, 1983; Nabi, 1989). If female-headed microenterprises face lower access to credit services than male-headed small firms, this would reduce their growth possibilities, and thus affect earnings. However, participation in credit markets is not skewed by gender in urban areas. About 12% of male- and female-headed microenterprises have received a production loan since starting their businesses. Moreover, there is evidence that gender does not influence borrowing activity from formal and informal lenders among urban microenterprises in Mexico (Sánchez, 1998). However, in the rural areas studied, 45% of women have received some type of financing (for consumption and production purposes) during the two years prior to the survey compared with 56% of men.
Fifth, women and men may have different goals with respect to the microenterprise activity, which influences their business strategies and, thus, their earnings. For example, women’s business strategies weight more heavily on their ability to supplement family income and emphasize risk management (Rhyne and Holt, 1994; Rothstein, 1995). Under this hypothesis, women may engage in low-risk, low-return activities to provide a stable source of funds for the family. For example, in Mexico’s urban areas, the main reason for women to enter the microenterprise sector is to supplement family income (57%), while men reported their desire to become independent (47%). Another evidence of differences in risk preferences between male- and female-headed microenterprises can be found in the reported reasons for not requesting a loan. Overall, about 23% and 27% of microenterpreneurs in rural and urban areas, respectively, perceive borrowing as a risky endeavor and prefer to use internally generated funds.
When analyzing the overall gender earnings gap (including all economic sectors), an observable determinant of the gender earnings gap is the difference in the sectoral distributions of female- and male-headed microenterprises. For example, if women’s microenterprises tend to be over-represented in low-returns sectors and under-represented in high-returns sectors, then some of the difference between the average female and male business earnings may result from different sectoral distributions.
Gender disparities in sectoral distributions can be analyzed by using the index of dissimilarity (or Duncan Index), which equals 0.5j|PjF - PjM | (Psacharopoulos and Tzannatos, 1992). PjF and PjM represent the proportion of females and males in the jth economic sector. The index compares the existing sectoral distribution of men and women to the distribution that would exist if “assignment” to sectors was unrelated to gender. If men and women were equally distributed across sectors, the index would equal zero, while if all sectors were completely dominated by one gender it would equal 100%. An estimation of the Duncan Index using a detailed classification of sectors is about 51% for both urban and rural microenterprises.
Male-female differences in sectoral choices are likely to be influenced by several factors. They include individual preferences or by barriers to entry into certain sectors resulting from gender differences in household responsibilities, high levels of required initial investments, lack of business skills, or consumer’s perceptions that the quality of services or products depend on the gender of microentrepreneurs. For example, the absence of women providing construction services may indicate that consumers prefer male providers or that women lack the necessary skills or desire to work in the sector. Moreover, pre-market factors, such as differential treatment by parents, schools, and gender differences in educational access and business experience also channel entrepreneurial women into a narrower range of economic sectors. Female microenterprises are concentrated primarily in economic sectors associated with household activities, such as food processing or
domestic services, often less dynamic sectors or more competitive ones (World Bank, 1994a).
A glance at Figure 1 suggests that across broadly defined sectors, women earn substantially less than men; the female-to-male earnings ratios in the microenterprise sector lay in the ranges of 30% to 58%, and 25% to 55% for urban and rural microenterprises, respectively.[3] Thus, given the above discussion, any meaningful analysis of gender earnings differentials should consider the role of both intra- and inter-sectoral factors in determining the causes for the male-female disparities in earnings. Moreover, the analysis should also incorporate the endogeneity of the sectoral decision to explicitly take into account gender differences in sectoral-specific barriers to entry or preferences.