Why so Few Women in the Labor Market in Turkey?

Abstract

Turkey has one of the widest male–female employment gaps in the world. The post-1950 interplay between economic growth strategies and the male-breadwinner family led to distinct gendered labor market outcomes in the import-substitution versus the export-led growth periods. Examination of aggregate employment data in the 1955–2009 period, as well as regression analyses of household survey data for 1988, 2000, and 2008 and qualitative data from a 1997 field study, show that the lack of a demand-side challenge to the male-breadwinner family resulted in the institutionalization of the gendered labor division and roles as binding constraints on women' labor supply. The prevalence of informal sector employment and absence of paid work–family reconciliation measures magnify these supply-side constraints. Social conservatism is a more limited constraint, while men' unemployment emerges as a counteracting factor. Nevertheless, women' desire for increased autonomy emerges as the primary motivation for entering the labor market.

Keywords:

• Labor force participation
• gender
• Middle East and North Africa (MENA)
• Turkey

JEL Codes:

• J21
• J71
• O53

Acknowledgments

I am grateful to the editors and the referees for their invaluable input toward the improvement of this paper.

Notes

The total female labor force was 6.9 million workers in 2009, of which 2 million were unpaid family workers of whom 91 percent (1.86 million) were unpaid family workers in agriculture (TSI 2009).

In 2009, Jordan, Egypt, Lebanon, Pakistan, Saudi Arabia, Syria, Yemen, Palestine, and Iraq immediately followed Turkish women's 24 percent economic activity rate (UN 2010).

The institutionalization of the patriarchal contract with negative feedback effects for women's employment is not specific to the MENA region. Dante Contreras and Gonzalo Plaza (2010), for instance, show that in Chile the institutionalization of a “machismo culture” shaped around the sexual division of labor creates a disabling institutional environment for women's employment.

The share of agriculture in total employment decreased to 59 percent in 1980, 47 percent in 1990, and 24.6 percent in 2009 (İpek İlkkaracan and İnsan Tunalı 2010).

The term “housewification” is translated from the original German term hausfrauisierung used by Bennholdt-Thomsen (1987).

Using 1980 as baseline for comparison is appropriate because while Turkey was the last of these countries to launch the ELI strategy, the others were in the relatively early phases of their respective liberalization experiences.

Consequently, the women's share of employment in the service sector, which was 12 percent in 1980, increased to 24 percent in 2009; their share of manufacturing increased from 15 to 22 percent in the same period. Electricity, gas and water, mining and quarrying, and construction have also had increasing shares of women, though to a lesser extent.

Note that two studies on export orientation and feminization in Turkey report mixed findings for the early 1980s. Nilüfer Çağatay and Günseli Berik (1994) investigate aggregate Turkish manufacturing data for the first half of the 1980s and conclude that export orientation has not triggered any visible feminization. Şule Özler (2000) uses plant-level data for the same period and finds that export orientation increases the employment of women but increased capital intensity decreases their employment. Looking at the question from a longer time perspective, I find an increasing share of women in manufacturing.

Unlike most MENA countries that do not export oil, Turkey does not have access to oil-related revenues through migrant remittances because Turkey does not share the common language of Arabic. Yet the substantial labor migration of Turkish men to EU countries in the 1950–70 period resulted in a flow of remittances from Europe, which might have contributed to sustenance of the male-breadwinner family. (I am grateful to an anonymous referee for pointing this out.)

In 1980 the rate of unionization was 54 percent according to the Ministry of Labor and 29 percent according to the OECD, depending on whether the rate is expressed as a share of social security-registered employees or all wage employees (İpek İlkkaracan 2005). These estimates are higher than 15 percent for South Korea, 19 percent for Mexico, and 20 percent for Brazil in the 1976–80 period (David G. Blanchflower 2006). More recent estimates for Turkey show a substantial decline in unionization rates to around 8–10 percent for the 2000s (İlkkaracan 2005).

The share of real wages in value added decreased from 40 percent in 1980 to 15 percent by 1990 (see Nurcan Özkaplan [2008]).

See Ahmet Faruk Aysan and Yavuz Selim Hacıhasanoğlu (2007) and Lerzan Özkale and Özgur Kayalıca (2008) for evolution of Turkish export composition from labor-intensive, low value-added to capital-intensive, high value-added sectors.

With each of the three economic crises, married women's participation rate registered an approximately 2 percentage point increase: in 1994, from 11.7 to 13.6 percent (corresponding to a phenomenal 20 percent increase in the number of married women participants from 1993 to 1994); in 2001, from 13.3 to 14.9 percent; and in 2008, from 15.1 to 16.9 percent.

Both discriminatory items were eliminated from the Civil Code in 2003 due to strong and persistent advocacy of the women's movement and pressure for compliance with EU norms, as well as the Convention on the Elimination of All Forms of Discrimination against Women (CEDAW), to which Turkey has been a party since 1985.

The effect of the husband's unemployment status has been examined in previous studies with mixed results on whether women play a role as added or discouraged workers and the relative strength of these two effects (Tansel 2001; Cem Başlevent and Özlem Onaran, 2003). The discouraged worker effect entails those who have stopped looking for a paid job because they lost hope of finding one; hence they are removed from the status of being unemployed to that of nonparticipant. The added worker effect entails entry of household members into the labor market in order to compensate for the loss of income by the primary breadwinner. In doing so, added workers move from the status of being nonparticipant to that of participant (either employed or unemployed).

The decreasing pull of higher education may also have to do with the early retirement scheme that went into effect in the early 1990s. To account for this effect, I have repeated the analysis with prime-age (ages 25–44) working women, but obtained similar decreases in the odds ratio for university and high school graduates, from forty-two and seven times for university and high school graduates respectively in 1988, down to thirty-six and five times in 2008.

The same phenomenon has been noted for the MENA region that increasing gender equality in education has not necessarily translated into decreasing gender gap in labor force participation rates as it has done elsewhere (World Bank 2004; Stephan Klasen and Francesca Lamanna 2009).

Günseli Berik (1997) warns against the uncritical use of survey-generated data in feminist research, pointing out that it deprives researchers of one of the opportunities to understand the subtleties of economic life. She encourages the gathering and use of qualitative data and the use qualitative techniques to generate more reliable quantitative data. Marlene Kim (1997) points out the additional insights provided by feminist participatory field research, particularly one that allows for inputs from women's organizations and research subjects.

The percentages indicating conservatism could have possibly increased over time, given that the period since 1997 in Turkey has been marked by a strong surge of conservative religious politics. Simultaneously, Turkey's EU membership prospects along with a growing women's movement have allowed progressive legal reforms, and there were substantial gains in women's education.

The relation between “three or more children” and women's labor force participation is worthy of mention, given the public statements by the Prime Minister of Turkey in 2010 encouraging women to have three children.

Note that 2008 official data for Turkey also show that the majority of adult women (close to half) have only primary-level education.