75467 State Planning Organization of the Republic of Turkey and World Bank Welfare and Social Policy Analytical Work Program Working Paper Number 5: Determinants of and Trends in Labor Force Participation of Women in Turkey Meltem Dayıoğlu Department of Economics, Middle East Technical University Murat G. Kırdar Department of Economics, Middle East Technical University Ankara, March 2010 Determinants of and Trends in Labor Force Participation of Women in Turkey iii Determinants of and Trends in Labor Force Participation of Women in Turkey Table of Contents Executive Summary.................................................................................................................................................. 1 1. Introduction........................................................................................................................................................... 7 2. Literature Review ................................................................................................................................................ 8 2.1. National Studies . .......................................................................................................................................... 8 2.2. International Studies ..................................................................................................................................... 8 3. Data ...................................................................................................................................................................... 9 4. Legal Framework ................................................................................................................................................. 10 5. Trends in Female Labor Force Participation in Times Series Data ..................................................................... 11 5.1. Trends in Population . ................................................................................................................................... 11 5.2. Time Profiles . ............................................................................................................................................... 11 5.2.1. Female labor force participation rates by age over time ....................................................................... 14 5.2.2. Hours Worked for Employed Females over Time ................................................................................. 16 5.2.3. Occupational Distribution of the Employed by Sex and Year . ............................................................. 18 5.2.4. Employment Status by Sex and Year . ................................................................................................... 19 5.3. Cohort Profiles . ............................................................................................................................................ 21 5.3.1. Female Labor Force Participation Rates by Age for Successive Birth Cohorts . .................................. 21 5.3.1.1 Women in Urban Areas..................................................................................................................... 22 5.3.1.2 Women in Rural Areas ..................................................................................................................... 23 5.3.2. Weekly Hours Worked by Age for Successive Birth Cohorts ............................................................... 25 5.3.2.1 Women in Rural Areas ..................................................................................................................... 25 6. Education, Marital Status, Fertility, Migration and Labor Force Participation ................................................... 27 6.1. Education ...................................................................................................................................................... 27 6.1.1. Schooling attainment by age over time for female population ............................................................. 27 6.1.2. Female labor force participation rates by educational attainment over time ........................................ 30 6.2. Marital Status . .............................................................................................................................................. 32 6.2.1. Marital status of women by age over time ............................................................................................ 32 6.2.2. Female labor force participation rates by marital status and age over time .......................................... 34 6.3. Fertility ......................................................................................................................................................... 36 6.3.1. Age specific fertility rates over time . .................................................................................................... 36 6.3.2. Age specific fertility rates by birth cohorts over time ........................................................................... 37 6.3.3. Female labor force participation rates by number of births and age over time ..................................... 37 6.4. Migration ...................................................................................................................................................... 38 7. Determinants of Female Labor Supply ................................................................................................................ 39 8. A Closer Look at High and Low Skilled Women in Urban Areas and Rural Women ......................................... 43 8.1. Understanding the Declining Participation Rates among the Highly Educated Urban Women .................. 43 8.2. Understanding the Low and Stagnant Participation Rates among the Low-Skilled Urban Women ............ 47 8.3. A further look at the declining participation rates of women in rural areas.................................................. 52 9. Conclusion............................................................................................................................................................ 55 References . .............................................................................................................................................................. 59 Appendix A .............................................................................................................................................................. 61 Appendix B .............................................................................................................................................................. 69 Determinants of and Trends in Labor Force Participation of Women in Turkey 1 Executive Summary areas for women has been declining itself: it has gone down from a level of 50.7 percent in 1988 to 33 1. Female labor force participation rate in Turkey is percent in 2006. Caution must be exercised, though, in quite low by EU and OECD standards: it was 24.9 comparing the numbers before and after 2000 because percent in 2006, compared to 66.1 percent in EU-27 the sampling frame of the Household Labor Survey and 60.8 percent in OECD countries. Moreover, it has used in constructing these numbers changed in 2000. declined from 34.3 percent in 1988 to 24.9 percent However, when these periods are examined separately, in 2006. The purpose of this report is to shed light we still find that the rural labor force participation on factors that determine women’s participation in rate decreased from 50.7 percent to 47.4 percent over the labor market and the reasons behind the observed the 1988 - 1999 period, and from 40.2 percent to 33 trends over the 1988-2006 period. Some of the broad percent over the 2000 - 2006 period. The decline in questions asked in the paper include: rural labor force participation rate for females has been more prominent since 2000. In addition, when • What might be the reasons for the declining labor we examine the change in the participation rates by force participation of women? age groups, we see that the decline has been larger for • What factors affect the labor force participation of younger rural women. women? • In particular, what are the roles of education and 4. Despite the declining trend, the female labor force demographic variables in determining women’s participation rate in rural areas is still higher than that participation in the labor market? in urban areas, which has been more stable over time. In fact, the gender gap in participation rate in urban 2. An important reason for the fall in female areas is much wider. Strikingly, the participation rate participation rate is urbanization. Turkey has witnessed in urban areas has always been lower than 20 percent. high levels migration from rural to urban areas since As can be seen from the right panel of Figure E.1, the 1988. The share of urban population rose from 51.1 percent in 1988 to 63.3 percent in 2006. Since the participation rate has displayed an almost flat profile labor force participation rate in urban areas is much between 1988 and 1999 at a level of approximately 17 lower, at approximately 20 percent in 2006, than that percent. Since 2000, however, it has shown an upward in rural areas, at approximately 33 percent in 2006, the trend: the participation rate of urban women increased increasing share of urban population pulls down the by 2.5 percentage points from a level of 17.4 percent labor force participation rate of women. in 2000 to 19.9 percent in 2006. When we examine the participation rates by age, we find that the increase 3. However, it is not only the decreasing share of in the participation rates for certain age groups in rural population that is pulling down the participation urban areas have been remarkable. For instance, the rate for women. As can be seen from the left panel participation rate for 25- to 29-year-olds increased by of Figure E.1, labor force participation rate in rural 8.5 percentage points from 1988 to 2006 and by 5.1 Figure E.1: Labor Force Participation Rates by Sex and Rural/Urban Status Rural Urban 90 90 80 80 70 70 60 60 50 male 50 male LFPR LFPR 40 female 40 female 30 30 20 20 10 10 0 0 1988 1989 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 1988 1989 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 Source: HLFS web data base, TUIK (http://www.tuik.gov.tr). Note: Covers individuals age 15 plus. Determinants of and Trends in Labor Force Participation of Women in Turkey 2 percentage points from 2000 to 2006. This is important Turkey are unpaid family workers as a result of the as it stands in stark contrast to the gloomy picture large share of agriculture in female employment. frequently given for female labor force participation However, with the decline in agriculture, and family- in Turkey. run establishments, the importance of wage work among working women has been on the rise in both 5. In essence, declining labor force participation rural and urban areas. In fact, wage work is the in rural areas, low and stagnant participation rates dominant form of employment in urban areas where in urban areas, and the downward pressure on the 80 percent of women are found. These imply that labor aggregate participation rate caused by urbanization force participation rate of urban women is rising along emerge as prominent characteristics of female labor with a rising incidence of wage work. force participation rate in Turkey. Yet, it is important, at the same time, to acknowledge the increasing 8. A cross-section analysis of female labor force participation rate among the young urban women in participation rate profiles according to age in urban Turkey. areas in 2006 reveals a hump-shaped age-profile for urban participants, the peak occurring between ages 6. Examining the change in hours worked over 20 and 29. A similar hump-shaped profile is also found time for female employees, we find that it is not the for rural labor force participants; however, its hump is increasing prevalence of part-time work that is behind much weaker – the range between the ages of 20 and the increasing participation of women in urban areas. 59 is relatively flat –. The problem with these cross- To the contrary, part-time work has decreased over section profiles is that the age effects could also stand time. Close to 70 percent of women in Turkey work for cohort and/or calendar year effects. Therefore, for 40 hours or more per week. This figure approaches using pseudo-panel techniques with a series of cross- to 85 percent in urban areas. section data, we decompose age, cohort, and calendar year effects in the participation rate profiles. This 7. Agriculture carries a heavy weight in the lives decomposition is built on the assumption that year of women workers in Turkey. Although declining, effects add up to zero and that they are orthogonal to a a sizeable proportion of women in Turkey, almost linear time trend. 60 percent, still work in agriculture. This implies that changes in agricultural activities will exert a 9. Figure E.2 illustrates the estimated age effects strong influence on the trends in the labor force according to rural/urban status from this decomposition participation of women. Since the agricultural sector analysis. In urban areas, we find a counter-clockwise in Turkey is dominated by small-scale family run rotation in the age-profile of labor force participation establishments, the female labor force in rural areas rates. The peak is now between the ages of 20 and 39, predominantly work as un paid family workers. instead of 20 and 29. The age-profile of labor force In fact, about 40 percent of all working women in participation rates in rural areas changes completely Figure E.2: Age Effects in Female Labor Force Participation Rate according to Rural/Urban Status 0 Rural RURAL 10 Urban URBAN -20 5 ageff ageff -40 0 -60 -5 15 20 25 30 35 40 45 50 55 60 15 20 25 30 35 40 45 50 55 60 Determinants of and Trends in Labor Force Participation of Women in Turkey 3 after the decomposition. We find a monotonic age labor force participation rate to rise and the rural effect: rural women become less likely to participate laborforce participation rate to fall. as they age at all ages. This is dramatically different from the hump-shaped profile we found with cross- 12. We also carried out a similar decomposition sectional analysis. analysis for hours worked in urban areas. As a result of this analysis, we found that younger cohorts of 10. The age effects illustrated in Figure E.2 has women in urban areas are also less likely to work for important implications regarding the labor force shorter hours (less than 40 hours per week) and more participation in Turkey in the future. Since the likely to work for longer hours (more than 50 hours participation rate in rural areas goes down by age, per week) than older cohorts. In other words, younger the aging of the Turkish population will exert a cohorts of women in urban areas are not only more downward pressure in the participation rate in rural likely to participate in the labor market but also more areas. We could expect a similar downward pressure likely to work for longer hours. This implies that the on the urban participation rate as the population ages finding for the whole female urban population – that because participation rates after age 40 are much both participation rates and hours worked are on the lower. However, lower participation rates after age rise – is driven by the different behavior of younger 40 is mostly a result of the retirement insurance cohorts of women entering the labor force. system. As the population ages, the rules of the entry 13. Significant improvements have taken in place in into retirement would be expected to change as well women’s schooling in recent decades in Turkey. For (they were, in fact, revised in early 2000s.) It is also instance, the share of illiterates fell from 33.9 to 19.6 important to note that low participation rates after percent whereas the share of university graduates rose age 40, as a result of early retirement, is an important from 1.8 to 5.8 percent from 1988 to 2006. One of the contributing factor to the low aggregate female labor most salient features of female labor force participa- force participation rate in Turkey. tion in Turkey, displayed in Figure E.4, is that it so much depends on educational attainment: participa- 11. Birth-cohort effects are displayed in Figure E.3 tion rates increase substantially with education. While according to rural/urban status. In urban areas later the participation rates of women in urban areas with- cohorts of women are much likely to participate in the out a primary school diploma is below 10 percent and labor market. In rural areas, on the contrary and perhaps those of women in urban areas without a high school surprisingly, later cohorts of women are less likely to diploma is below 15 percent, those of women with participate in the labor market. These imply that as university degrees is at 70 percent. earlier cohorts leave the population and replaced by forthcoming cohorts–whose labor force participation 14. This figure explains why the labor force partici- rate will presumably be similar to the youngest pation rate in urban areas in Turkey is so low. It is low cohort in the sample–we would expect the urban primarily due to the labor force participation behavior Figure E.3: Birth-Cohort Effects in Female Labor Force Participation Rate according to Rural/Urban Status 80 RURAL Rural 5 URBAN Urban 60 coheff coheff coheff coheff 40 40 00 20 0 -5 86-90 76-80 66-70 56-60 46-50 36-40 26-30 86-90 76-80 66-70 56-60 46-50 36-40 26-30 Determinants of and Trends in Labor Force Participation of Women in Turkey 4 of low-educated women, who still account for a large female population were not high school graduates in share of the urban population: 73.2 percent of the urban 2006. Figure E.4: Labor Force Participation by Education in Urban Areas in 2006 University Vocational High School Women Secondary Men Primary School No Diploma Illiterate 0 10 20 30 40 50 60 70 80 90 Source: 2006 HLFS, http://www.tuik.gov.tr Note: Covers individuals age 15 plus. 15. Another salient feature of women’s labor force of educational attainment have been stagnant. In other participation by educational attainment in urban areas words, conditional on schooling, women’s participation is that participation rates for university graduates as rates in urban areas are either stagnant or falling, yet the well as regular and vocational high school graduates overall participation in urban areas is rising. This is to have fallen over time. This is illustrated in Figure E.5. do with the shift in the composition of the workforce to- The fall for university graduates took place in the mid wards more educated women who have higher participa- 1990s, whereas the fall in high school graduates has tion rates. Had the participation rates of highly educated been more gradual. On the other hand, participation women remained at levels recorded in 90s, women’s par- rates of women in urban areas with lower levels ticipation today would have been much higher. Figure E.5: Change in Labor Force Participation over Time for University Graduates (left-panel) and High School Graduates (right-panel) in Urban Areas University Graduates High School Graduates 100 90 95 80 90 70 85 60 80 High-Men 50 Voc-Men LFPR LFPR 75 Men High-Wom 40 70 Women Voc-Wom 30 65 20 60 55 10 50 0 1988 1989 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 1988 1989 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 Source: HLFS web data base, TUIK (http://www.tuik.gov.tr). Note: Covers individuals age 15 plus 16. Marriage is universal, while divorce is an unlikely by that age. The timing of marriage is early with the event in Turkey: nearly 98 percent of women marry average age at first marriage being 20.7 years. The by age 49 and less than 1 percent of women divorce implication of these demographic factors is that the Determinants of and Trends in Labor Force Participation of Women in Turkey 5 majority of women spend a good part of their life being a lower propensity to participate in the labor market. married. However, the labor force participation rate of This time, perhaps surprisingly, it turns out that married women is lower than that of single women in migrant women originating from cities have a lower Turkey: while the participation rate of single women propensity to participate in the labor market after we is 34.3 percent, the corresponding rate for married account for a number of personal characteristics. women is 23.1 percent. The gap between the two demographic groups widens further in urban areas, 19. The multivariate analyses carried out to see how where the participation of single women increases to various factors are associated with female labor market 35 percent but that of married women drops to 15.5 participation confirm the important role of education. percent. On the bright side, married women’s labor In particular, university education is strongly positively force participation rate in urban areas has increased associated with labor force participation in both rural over time. In addition, the share of single women in and urban areas. As expected, married women are the population –whose participation rates are higher found to have a lower likelihood of participation – is rising. in both areas. Children younger than 15 years were also found to be negatively associated with the 17. Fertility behavior of women is also very important participation probability of urban but not rural women. with regard to their labor force participation decision That the regions were found to be strongly associated as children influence the opportunity cost of market with participation shows that demand side factors are work. This is particularly important in Turkey as also important in determining women’s labor market almost all married women have children. A comparison participation. of fertility rates in 1993 and 2003, using DHS data, reveals falling fertility rates in both rural and urban 20. We also conducted multivariate logit analyses areas. Moreover, the magnitude of this fall is signifi- separately for four educational attainment groups: cant. A decomposition analysis of age, cohort, and time no education, primary, secondary, and higher. One effects shows that cohorts born after the 1970s have interesting finding from this analysis is that children a much lower propensity to have children. When we matters much more for highly educated women. In examine labor force participation rates by motherhood fact, for women with no education, there is no evidence status, we find that women with children have lower at all that children is associated with labor force participation rates, particularly in urban areas. The participation. On the contrary, the negative association lower fertility rates of younger cohorts of women and of household wealth with labor force participation the negative correlation between children and labor strengthens as education level decreases. For women force participation imply a higher participation rate for with higher education, there is no evidence at all for younger women in Turkey. a negative association between household wealth and 18. Given the high internal migration rate in Turkey, labor force participation. We also find that age effects it is also important to understand how the labor force are stronger for higher educational levels. participation behavior of migrant women compares to that of non-migrant women. We find that the 21. One key finding that emerged from our examination participation rate of migrant women in urban areas – of female labor force participation in Turkey is the who changed places within the last five years - is, in declining labor force participation rates in rural areas. fact, slightly higher at 29.6 percent than that of non- Therefore, we examined the potential underlying migrant women in urban areas, which is 27.0 percent. reasons to this fact. First, we find that the geographical Among migrants, we differentiate between those shift in the shares of rural population could partly originating from cities, towns and villages. Migrant account for this fact. There exists substantial variation women who originate from villages have a lower in rural participation rates across geographical regions. participation rate at 23.6 percent than non-migrant The share of rural population residing in the Black Sea women in urban areas, but an appreciable difference regions and Northeastern Anatolia–where participation among others is not observed. However, the lower rates are much higher–is falling. Second, the fact that participation rate of migrant women originating from women in rural areas with a higher propensity to villages can be explained by their different personal participate in the labor market are also more likely to characteristics. Once we account for these variables, migrate could explain the declining participation ra- migrant women originating from villages do not have tes. Comparing the characteristics of migrants leaving Determinants of and Trends in Labor Force Participation of Women in Turkey 6 rural areas with those of the residents of rural areas, multivariate analyses showed that for university we find that migrants, on average, have higher levels graduates household income is not negatively of education and are younger than stayers in rural associated with participation, unlike other educational areas. Given the increasing participation rates with attainment groups. As a last potential explanation, we education, the fact that more educated rural women looked at the change in the composition of university are leaving for other locations would partly account graduates over time but could not identify a change in observable characteristics that could have led to for the falling participation rates. Moreover, that these their declining participation rates. However, this movers are much younger also help us explain why the does not preclude changes occurring in unobserved decline in participation rates in rural areas is especially characteristics that could decrease their participation prominent at younger ages. rates. Indeed, we show that the growth in the share of university graduates has been particularly fast over 22. Another potential explanation to the falling par- the 1994-1999 period, when the biggest decline in the ticipation rates in rural areas is declining agricultural labor force participation of university graduates took wages due to a worsening of agricultural prices. When place. we examine the terms of trade between agriculture and manufacturing, we find that the terms of trade 25. The final interesting finding that requires further for agriculture in fact worsened after 2000. This investigation is the low and stagnant participation also could partly account for the falling participation rates of low skilled women – those with less than rates in rural areas after 2000. In addition, a shift to high school education. Over the 2000-2006 period, less labor-intensive agricultural products could also the participation rate of low skilled women varied explain the decline in participation rates. However, we between 10.9 and 11.8 percent. These are considerably do not observe a significant change in the land allotted lower rates compared to that of low skilled men which, over the same time period, varied between 67.1 and to various agricultural products over time. 68.8 percent. To explain this large gap between the 23. Another important finding with regard to participation rates of low-skilled men and women, agriculture that could help us explain the falling we compared the wages of the two groups and noted an improvement in women’s wages vis-à-vis men. participation rates is the decline in share of households However, we also noted that despite the improvement, in rural areas engaged in agriculture, and in particular women’s wages were still very low, with over 75 own-account agriculture after 2000. A decline in own- percent of women receiving wages that were below account agriculture means a loss of an easily available the minimum wage. These low market wages as well source of work for many women. The above-mentioned as high reservation wages of women stemming from worsening of terms of trade in agriculture could partly the big household sector in Turkey coupled with account for this decline as well. In addition, the re- very long hours of work probably explain why they classification of certain locations as urban could also participate in the labor market in so few numbers. A explain the declining share of agriculture. look at broad segregation indicators such as industrial distribution and firm size has not shown drastically 24. Another key finding of the study is the decline in the different patterns between men and women, though labor force participation of highly skilled women over they are distributed differently across occupations, the 1988-1999 period and their stagnant participations which probably explain some of the gender wage rates (at about 70 percent) after 2000. The absence of gap. wage data has precluded us from investigating the role 26. To have a better understanding of why both low of wages in explaining the declining rates prior to 2000. and high skilled women in urban areas have had The wage data available for the 2002-2006 period stagnant labor force participation patterns in recent indicate a deterioration and a recovery in women’s years, it would have been revealing to look at the wages over the studied period. This is a potential activities of non-employed women as well. For this explanation for why the labor force participation of purpose, we had originally planned to use the 2006 highly skilled women did not record an increase in Time-use Survey of TUIK but at the time of the writing this time period. Another potential explanation for the of this report the data were not released yet. Another declining/stagnant participation rates of highly skilled area that needs further investigation is the family-run women would be an increase in their reservation agricultural establishments, where we see the biggest wages due to higher household incomes. However, our change in female labor supply. Determinants of and Trends in Labor Force Participation of Women in Turkey 7 1. Introduction rate fluctuated around 6.6-9.7 percent for men and 6.3- 10.3 percent for women. 27. One of the salient features of the labor market in Turkey is the distinctly lower participation rates of 29. The low and declining labor market participation women vis-à-vis men. In 2006, while the participation of women has been voiced as a concern in a number rate of men (age 15+) was on the order of 71.5 percent, of policy documents, urging the government to take the corresponding rate for women was 24.9 percent. In action to reverse the observed trend. The government comparison to OECD and EU-27 averages (for ages has recently responded to this rising concern by 15-64), the participation rates of both men and women formulating possible policy actions that could increase in Turkey are lower. However, in the case of men, the women’s participation in the labor market. gap is not more than six percentage points (for ages 15-64). In contrast, the average participation rates 30. The purpose of this report is to give support to of women in EU-27 and OECD countries recorded such efforts by trying to shed light on factors that at 66.1 percent and 60.8 percent, respectively, are determine women’s participation in the labor market well above the Turkish rate. Furthermore, despite the and the reasons behind the observed trends. Some of increasing labor market participation in much of the the broad questions asked in the paper include: EU and OECD countries in the last decades, the labor market participation of women has actually declined • What might be the reasons for the declining labor in Turkey. force participation of women? • What factors affect the labor force participation of 28. Table 1 provides the key labor market indicators for women? men and women for 1988, 2000 and 2006. The decline • In particular, what are the roles of education and in the labor force participation of women is apparent demographic variables in determining women’s from the table. In less than two decades, from 1988 to participation in the labor market? 2006, the participation rate of women went down by about 10 percentage points from 34.3 percent to 24.9 31. The report is organized as follows. Section 2 percent. A similar change in the participation rates has presents a brief literature survey outlining previous also been observed among men where men’s partici- studies done in Turkey on female labor supply and pation went down from 81.2 percent in 1988 to 71.5 summarizes the different lines of research in this area percent in 2006. Over this period, the unemployment in the international literature. Section 3 presents the Table 1: Main Labor Market Indicators Source: HLFS web data base, TUIK (http://www.tuik.gov.tr). Note: Population figures are in thousands. Covers individuals ages 15 plus. Determinants of and Trends in Labor Force Participation of Women in Turkey 8 data sets used in the study. Section 4 looks briefly at Kasnakoğlu and Dayıoğlu (1996), Ercan and Tunalı the legal framework in Turkey as regards to women’s (1998), Tansel (1994, 1999), and Dayıoğlu and Tunalı employment. Section 5 discusses the trends in female (2003) found significant differences in the hourly labor force participation in three main sections. While earnings of men and women. The Oaxaca-Blinder the first sub-section presents the recent population type decomposition exercises employed in many such trends, the second sub-section analyses participation studies indicated that it is not the endowment dif- in time series data and the third sub-section carries ferences that lead to the wage gap but rather the higher out cohort analyses. In Section 6, education, marital market valuation of male traits. Tunalı and Başlevent status, fertility and migration are studied as potential (2001) looking at wage earners and self-employed determinants of the labor force participation of women concluded that wage labor attracts the best women and the observed changes in them. In section workers. Similar findings are noted by Dayıoğlu and 7, multivariate analyses are carried out to determine Başlevent (2006), drawing attention to the rather the factors affecting female labor force participation. heterogeneous nature of the female workforce. Section 8 presents more focused analyses of urban low and high skilled women and rural women. Section 9 35. Another line of research investigated the effect of concludes the report. structural adjustment policies and changing industrial structure on women’s employment (see for instance 2. Literature Review Çağatay and Berik, 1990; Çağatay and Özler, 1995; Özler, 2000; Onaran and Başlevent, 2004). These 2.1. National Studies studies found evidence for the feminization of the labor force following the 1980 structural adjustment 32. The low labor market participation of women has program that adopted export-oriented growth strategy. been the subject of various studies. Some of the earlier However, plant-level data indicated that the effect of work on the topic relied on published data and tried technological innovation was not always favorable to link macro variables with the falling participation to female labor (Ansal, 1997). There were also few rates of women (see for instance World Bank, 1993; studies that investigated the effect of business cycles on Özar, 1994; Özbay, 1994; TUSİAD, 2000, 2008). women’s participation. Başlevent and Onaran (2002) Urbanization, migration from rural to urban areas, and and Özler (2000), for instance, noted that it is more the low levels of female education were noted as the likely for women to become added than discouraged probable causes for the low and declining participation workers during economic downturns. rates of women. The General Directorate of Women’s Status and Problems also contributed to the early 2.2. International Studies efforts by commissioning research projects on various aspects of women’s work lives. Most of these studies 36. Our study builds on the existing knowledge were based on small-scale surveys on specific sectors. and aims to throw light on recent developments in women’s participation in Turkey by drawing on both 33. As large scale micro-level data became publicly the national international experiences. available, studies on various aspects of female labor supply flourished. One line of research investigated 37. Killingworth and Heckman (1985), in their survey the determinants of labor force participation of women paper, review the theoretical and empirical work on using cross-sectional data (for instance Dayıoğlu, female labor supply with a focus on Western economies. 2000; Dayıoğlu and Kasnakoğlu, 1997; Ercan and The 1985 special issue of Journal of Labor Economics Tunalı, 1998; Tunalı, 1997). These studies found years includes a collection of papers examining female labor of schooling, age, marital status and number of children supply in a number of developed countries. Many of to be important determinants of the participation these studies highlight the rising participation rate decision of women. In an interesting study, Çınar for women, in particular that for married women. (1994) found that ‘safety’ concerns affect women’s Gustafsson and Jacobsson (1985) draw attention to employment decision, inducing them to prefer home- the decreased gender wage differential in Sweden based work over formal work arrangements. in explaining the rising female participation rate. Similarly, Shimada and Higuchi (1985) report that 34. Studies investigating the gender earnings gap improvements in female education and wages led were a natural extension of participation studies. to a substantial increase in paid female employment Determinants of and Trends in Labor Force Participation of Women in Turkey 9 in Japan. Greenwood, Seshadri, and Yörükoğlu (2005) it was increased to 15,000 households in 1994 (but draw attention to the importance of consumer durable allowed for non-response rather than substitution) goods revolution in the rise of married female labor- and further to 7,800 households per month or 23,000 force participation. The substantial increase in female households per quarter in 2000. The sample size in labor force participation in the U.S. during the 1970s 2004 was expanded further to 13,000 households and 80s came to a slowdown in the 1990s and early per month to provide NUTS2 level estimates on key 2000s according to Juhn and Potter (2006). variables. Besides the changes in sample size, the sampling methodology also changed in 2000 so that 38. Many studies examine the relationship between the same households were included in the survey for labor force participation and demographic factors four times over a period of 18 months. It is not clear like fertility and marital status. Michael (1985) finds how this methodological change affected the key evidence for correlation between the participation rate variables. However, our analysis shows a break in of females and many demographic variables using the proportion of women employed in agriculture: a bivariate autoregressive structures. Recent studies significant fall in the labor market participation and on the relationship between fertility and labor force employment of women in agriculture was observed in participation in OECD countries tend to find a positive a single year going from 1999 to 2000. Since we do trend, contrary to the findings until 1985. However, not have information on how the redesign could have Kögel (2004) contests this finding and argues that there affected the key variables we do not adjust the data in has only been a decrease in the negative relationship any way. Another change to the HLFS questionnaire but it has not turned positive. –but not the sampling frame–occurred in 2005. In collaboration with EUROSTAT the questionnaire 39. Psacharopoulos and Tzannatos (1993) examine expanded to include an additional 12 questions. Some female labor force participation in 15 Latin American of the additional questions were aimed at increasing the countries. They find increasing participation despite quality of the data collected, while others were included adverse economic conditions. They also report a to be in line with the questionnaires of EUROSTAT. negative association between participation and Going from 2004 to 2005, in a single year, significant marriage as well as fertility. falls in the participation rates of rural women were 40. With regard to the impact of business cycles again observed. While it is hard to know whether the on female employment, Lim (2000) examines the changes made to the questionnaire has anything to do differential impact of East Asian economic crisis on with the unusually big drop, in a report prepared for the employment of women compared to men. TUSIAD, Yükseler and Türkan (2008) argues that this is a possibility that should not be brushed aside. 3. Data 43. In terms of the information collected, the HLFS 41. The main source for labor force statistics in Turkey carries the common features of a regular labor force are the Household Labor Force Surveys (HLFS) of the survey with the exception that wage data are only Turkish Statistical Institute (TUIK), which have been available in recent applications. The definitions of key conducted regularly since 1988. Earlier data on labor variables have remained constant over the 1988-2006 force and employment are rather scanty and do not applications so that the HLFS data are comparable necessarily rely on similar definitions or methodologies across time. as HLFS. Therefore, we primarily rely on the HLFS data of TUIK from 1988 onwards to analyze the main 44. Although the HLFS includes basic information on trends in the labor force participation of women. demographics, such as the composition of the household and its size, information on fertility is missing. This 42. The HLFS were originally conducted biannually. is crucial in the analysis of the effect of children on Starting with the 2000 survey, data have been released women’s labor market participation. Through HLFS, quarterly and since 2005 on a monthly basis using a one can only determine the number of children currently moving average of three months. Annual data are also residing in the household, though in households where released and we primarily rely on annual data to depict multiple families are found it is not always possible the participation profiles. While during the 1988-1994 to link mothers with their children. Information on applications the sample size was 11,160 households, reproductive behavior, however, can be obtained Determinants of and Trends in Labor Force Participation of Women in Turkey 10 from the Demographic and Health Survey (DHS) that increase the cost of women workers vis-à-vis of Hacettepe University. DHS have been conducted male workers. For instance, the Labor Law still views in 5-year intervals since 1968. However, it is only child care as the responsibility of the mother, requiring in recent surveys, i.e. in 1998 and 2003 rounds, that workplaces employing more than 100 female workers information of women’s labor market participation is to set up nursing rooms, and those with 150 women collected. workers or more to provide day care. That the establishment of day care depends on the number of 45. In trying to understand why women do not women workers and not on the total number of workers participate in the labor market, one has to account for creates an asymmetry in the cost of employing female the fact that the household sector in Turkey is rather and male workers. big. Women who do not participate in the labor market are actually engaged in various productive (but not 49. The most significant change that has occurred in necessarily ‘economic’) activities at home ranging the Labor Law concerns the extension of maternity from child and elderly care to preparing, serving and leave from 12 weeks to 16 weeks, with the flexibility preserving food for household’s own consumption. of using all but three weeks after the birth of the child. While in Western countries most of these activities The law also allows women to have 6 months of unpaid are replaced by their market substitutes, such is not leave following the end of 16 weeks of paid maternity the case in Turkey. In an effort to understand the leave. Although the new maternity provision enables activities of women who do not participate in the labor women to maintain their job attachment, it may also market and the characteristics of their households, lead firms to shy away from hiring women workers. the Turkish Statistical Institute has conducted a Time There are no provisions in the law for paternity leave. Use Survey in 2006. The Survey is a household-based cross-section and is representative of urban and rural 50. The new Labor Law also includes provisions that households in Turkey. However, this survey has not are geared toward making the workplace more women been made public yet and, therefore, could not be used friendly by explicitly recognizing and outlawing in this study. sexual harassment at workplace and shifting the burden of proof to the party that has been accused of 46. None of the above surveys carry a panel feature. harassment. However, multiple cross-sections of the HLFS and DHS allow the construction of synthetic panels, which 51. There are also provisions in other legal documents are used in disentangling the cohort effects from age that may negatively affect the labor market partici- and calendar year effects. pation of women or their attachment to it. For instance, should a female worker quit her job due to marriage she 4. Legal Framework is entitled to receive a severance payment. (A similar measure exists for men leaving their jobs to do their 47. Before embarking on an analysis of the trends military service.) Although this measure may seem to in the labor market participation of women and benefit women, it may also work to decrease women’s discussing the possible economic and demographic work attachment. There were also provisions in the factors behind the observed trends, we need to answer social security system that allowed unmarried females the following question: could the legal framework in who did not work to receive survivor’s benefits. This Turkey be limiting women’s participation in the labor provision was recently changed such that an upper market? age limit (that already existed for males) was set to identify the potential beneficiaries. 48. Outright discrimination against women in hiring, promotion and wage policies is illegal in Turkey. 52. Retirement age and the required contributions for However, ‘protective’ provisions play a role limiting retirement with full benefits were the other areas where women’s work and earnings opportunities. For women received preferential treatment. Under the old instance, until recently, women (with some exceptions) social security law, women could retire with full benefits could not take-up night work. The new Labor Act that with 20 years of service. (The requirement for men was adopted in May 2002 abolished this provision. was 25 years.) Amendments made to the law increased However, others remained: For instance, women the number of contribution years and established a cannot work in coal mines, underground quarries or in minimum age for retirement that was lower than that dangerous jobs (Süral, 2007). There are also provisions for men. The most recent amendment aims to equalize Determinants of and Trends in Labor Force Participation of Women in Turkey 11 the pension ages for men and women gradually at population above the age of 15 has increased by 2.5 65 years by 2048 (Süral, 2007). The early retirement million people from 16.5 million to 19 million during opportunity for women with full benefits (that is being an 18-year span from 1988 to 2006, whereas during gradually phased out) helps explain why women, the same time frame the urban population increased by especially in urban areas, retire earlier than men. roughly 15.5 million people from 17.2 to 32.7 million people. In other words, the share of urban population 53. The new Labor Law also introduced part-time rose from 51.1 percent in 1988 to 63.3 percent in and atypical work and allowed the establishment of 2006.1 temporary employment agencies. As to what extent these atypical work arrangements will facilitate 56. Figures 2 and 3 display the age structure for the women’s entry into the labor market remains to be population aged 15 and above in rural and urban areas, seen. respectively. What is most striking is that the share of young population, aged 15 to 24, has fallen and the 54. The government has also passed a new employment share of 25-49 age group has risen considerably in rural package in May 2008 with an aim to increase the areas. In urban areas, the age composition has been participation rates of women and the youth in the labor more stable. However, the share of young population market. In accordance with the amendment made to the has been falling in urban areas as well since the late Labor Law, the employer’s share of the social security 1990s. contributions for women (above the age of 18) and 5.2. Time Profiles the youth (younger than 29 years) hired within a year following the enactment of the law will be subsidized 57. Changes in the labor force participation of men by the Treasury for a period of five years. The subsidy and women in Turkey over time are depicted in Figure scheme is such that the government will pay 100 4 (See also Appendix Table A1). A gradual decline in percent of the social security premiums for women the participation rates for both men and women are and the youth in the first year. In subsequent years, clearly visible from the figure. While the labor force the subsidy will be cut by 20 points, so that in the final participation rate of men in 1988 was 81 percent, by year, the Treasury will pay 20 percent of the social 2006 this rate was down to 71.5 percent. Similarly, security premiums on behalf of the employers. The while the participation rate of women was 34.3 percent government plans to use the Unemployment Insurance in 1988, it was down to 24.9 percent in 2006. Fund to finance the program. The employment package is expected to positively affect the employment of 58. Part of the reason for the observed decline in women since the cost of hiring female vis-à-vis male the participation rates is to do with the declining importance of agriculture in employment, where par- workers is reduced. The package is also expected to ticipation rates have been traditionally higher than increase registered employment. other sectors due for instance to the prevalence of 5. Trends in Female Labor Force small-scale family-run farms.2 Different from many of the OECD and EU countries, agricultural sector is still Participation in Times Series Data sizeable in Turkey, though its share in employment 5.1. Trends in Population is on the decline. While in 1988, 46.5 percent of employed individuals were in agriculture, in less than 55. A very important feature that characterizes two decades this figure decreased to 27.3 percent. The the population of Turkey in the last decades is the move out of agriculture and rural areas is expected to high migration rates from rural to urban areas. reduce the overall participation rates for the reasons Accordingly, the urban population of Turkey has that the participation rates in urban areas (shown increased much faster than the rural population. As below), especially for women, are lower and a smaller can be seen in Figure 1, the non-institutional rural part of the population now live in rural areas. 1 The re-classification of urban/rural areas over time – notably, the re-classification of certain rural areas as part of metropolitan areas - could also explain the declining share of rural places. 2 Statistically, it is also easier to be classified as a labor market participant in rural areas, where market and domestic spheres often overlap. Determinants of and Trends in Labor Force Participation of Women in Turkey 12 Figure 1: Non-institutional Rural and Urban Population 35000 Population (in thousands) 30000 25000 20000 Rural 15000 Urban 10000 5000 0 1988 1989 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 Source: HLFS web data base, TUIK (http://www.tuik.gov.tr). Note: Covers individuals age 15 plus. Figure 2: Age Structure of the Rural Population 60 50 40 15-24 25-49 30 % 50-64 20 65+ 10 0 1988 1989 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 Source: HLFS web data base, TUIK (http://www.tuik.gov.tr). Note: Covers individuals age 15 plus. Figure 3: Age Structure of the Urban Population 60 50 40 15-24 25-49 30 % 50-64 20 65+ 10 0 1988 1989 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 Source: HLFS web data base, TUIK (http://www.tuik.gov.tr). Note: Covers individuals age 15 plus. Determinants of and Trends in Labor Force Participation of Women in Turkey 13 Figure 4: Labor Force Participation by Sex 90 80 70 60 50 LFPR 40 male 30 female 20 10 0 1988 1989 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 Source: HLFS web data base, TUIK (http://www.tuik.gov.tr). Note: Covers individuals age 15 plus. For data see Appendix Table A1. Figure 5: Labor Force Participation by Sex in Rural Areas 90 80 70 60 50 male LFPR 40 female 30 20 10 0 1988 1989 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 Source: HLFS web data base, TUIK (http://www.tuik.gov.tr). Note: Covers individuals age 15 plus. For data see Appendix Table A1. 59. Figures 5 and 6 show the participation rates in 60. A declining trend is also observed among men in rural and urban areas over time. As would be expected, urban areas: in comparison to the late 80s, participation there is a general fall in the participation rates in rural rates were down by almost 10 percentage points in the areas, which are defined as settlements with population early 2000s (Figure 6). In contrast to men, women’s less than 20,001. The fall in the participation rates are participation in urban areas was stagnant in much of particularly sharp in the case of women. While in 1988, the 1990s, but recorded a slight increase in the last few 50.7 percent of rural women entered the labor market, years, with the result that the gender participation gap this rate was down to 33 percent in 2006 (Figure 2). has somewhat narrowed (Figure 6). Despite the recent Corresponding rates for males were 84.7 percent and improvement, women’s participation remained below 72.7 percent, respectively. 20 percent. Determinants of and Trends in Labor Force Participation of Women in Turkey 14 Figure 6: Labor Force Participation by Sex in Urban Areas 90 80 70 60 50 male LFPR 40 female 30 20 10 0 1988 1989 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 Source: HLFS web data base, TUIK (http://www.tuik.gov.tr). Note: Covers individuals age 15 plus. For data see Appendix Table A1. 5.2.1. Female labor force participation rates by age thereon. The labor force participation rate of men is over time highest during their prime age years exceeding 90 61. The age-participation profiles of men and women percent (Figure 7). For women, participation is also are hump-shaped: participation is low at young ages, highest during prime age years but reaches rates of increases at prime age years (24-45) and declines only 30 percent. Figure 7: Labor Force Participation in 2006 by Age Groups 100 90 80 70 60 Men lfpr 50 Women 40 30 20 10 0 15-19 20-24 25-29 30-34 35-39 40-44 45-49 50-54 55-59 60-64 65+ Source: HLFS web data base, TUIK (http://www.tuik.gov.tr). Note: Covers individuals age 15 plus. 62. In urban areas, the hump-shape participation -reaching rates of about 30 percent. Participation rates profile of men strengthens as a result of faster exits of urban women decline as they reach their 30s. Even from the labor market at older ages and delayed faster declines occur as they reach their 40s when entry at younger ages owing to higher average many women make their transition to retirement schooling levels in urban areas. In comparison to (Figure 8). In the 45-49 age group, only 15.2 percent their male counterparts, urban women’s participation of women enter the labor market, which is half the peaks rather early-when they are in their twenties figure observed for the 20-29-year-old women.3 3 The rather generous retirement scheme in Turkey – which is being gradually phased out (see Section 4) - is among the possible explanations for the drastic drop in participation as early as mid 40s. Appendix Figures B1 and B2, compare age-participation profiles of men and women in Turkey to their counterparts in the US, OECD, EU member countries. That the drop is observed for both men and women starting around mid 40s, but much later in other countries, gives support to the conjecture that early retirement explains part of the reason for the relatively lower participation rates in Turkey. Determinants of and Trends in Labor Force Participation of Women in Turkey 15 Figure 8: Labor Force Participation in 2006 in Urban Areas by Age Groups 100 90 80 70 60 Men LFPR 50 Women 40 30 20 10 0 15-19 20-24 25-29 30-34 35-39 40-44 45-49 50-54 55-59 60-64 65+ Source: HLFS web data base, TUIK (http://www.tuik.gov.tr). Note: Covers individuals age 15 plus. For data on women see Appendix Table A2. Figure 9: Labor Force Participation in Urban Areas by Age Groups in 1988 and 2006 100 90 80 70 60 Men-06 Women-06 LFPR 50 Men-88 40 Women-88 30 20 10 0 15-19 20-24 25-29 30-34 35-39 40-44 45-49 50-54 55-59 60-64 65+ Source: HLFS web data base, TUIK (http://www.tuik.gov.tr). Note: Covers individuals age 15 plus. For data on women see Appendix Table A2. 63. Figure 9 compares the change in the age- on the order of 5.1 percentage points between 1988 participation profiles of men and women between 1988 and 2006. Increasing participation among prime age and 2006 for urban areas. For all age groups a declining women is consistent with the observed improvements trend is observed for men, which is particularly sharp in educational outcomes and demographic changes at younger and older ages. In the case of women, the such as delayed marriage and reduced fertility. average participation is up by 2.2 percentage points. The increase is particularly sharp among those in 64. A closer look at urban women’s participation over time reveals that the observed increase is not their prime age years. For instance, in comparison a consequence of the years chosen for comparison to 1988 participation is up by almost 9 percentage purposes but that the labor market participation of points among the 25-29-year-olds. However, a sharp women has been on the increase among the 20-54 age decline (8 percentage points) in participation is group for some time (Figure 10). It is also interesting observed among the youngest age group (15-19 years). to note that except for the year 2006, in all other years Falling participation rates among younger and older under consideration, the peak in women’s participation women pull the average participation rate down. Re- occurred in the 20-24 age group followed by a sharp estimating participation among 20-49 year-old women decline in the 25-29 age group. The profile for 2006 indicates an improvement in average participation shows a delayed peak and a postponed fall to the 30- Determinants of and Trends in Labor Force Participation of Women in Turkey 16 34 age group. Whether this will be the new trend depends on the size of time and cohort effects. It remains to be seen. However, in none of the profiles do might very well be that the patterns observed in age- we see a second significant upsurge in participation. participation profiles tell a different story when they 65. To what extent the participation profiles obtained are corrected for time and cohort effects. We will return from cross-section analysis shows how an average to this discussion in Section 5.3 where we construct woman behaves in the urban labor market as she ages synthetic panels. Figure 10: LFPR of Women in Urban Areas by Age Groups in 1988, 1994, 1999 and 2006 35 30 25 1988 20 1994 LFPR 1999 15 2006 10 5 0 15-19 20-24 25-29 30-34 35-39 40-44 45-49 50-54 55-59 60-64 65+ Source: HLFS web data base, TUIK (http://www.tuik.gov.tr). Note: Covers individuals age 15 plus. For data on women see Appendix Table A2 For data see Appendix Table A2. Figure 11: LFPR of Women in Rural Areas by Age Groups in 1988, 1994, 1999 and 2006 100 90 80 70 60 Men88 Men06 LFPR 50 Wom88 40 Wom06 30 20 10 0 15-19 20-24 25-29 30-34 35-39 40-44 45-49 50-54 55-59 60-64 65+ Source: HLFS web data base, TUIK (http://www.tuik.gov.tr). Note: Covers individuals age 15 plus. For data on women see Appendix Table A3. 66. The picture is quite different for women in rural 5.2.2. Hours Worked for Employed Females over areas: labor market participation is on the decline in all Time age groups with particularly sharp declines observed among younger women (Figure 11). While men’s 67. Long hours of work characterize the labor market in participation rates in rural areas have also declined Turkey. Figure 12 shows the distribution of employed over time, the decline observed for women has been individuals by hours of work per week. Nine out of 10 particularly sharp. employed men work 40 or more hours per week. Full- Determinants of and Trends in Labor Force Participation of Women in Turkey 17 time work is also prevalent among women, 67 percent an increase in the hours of work is observed. While of whom work 40 or more hours per week. in 1988, 22.5 percent of women were employed for fewer than 40 hours per week, this proportion dropped 68. An interesting question in the context of hours to 17.3 percent in 2006. Furthermore, for both men of work is whether the increase in the participation and women the proportions who work very long hours rates observed among urban women has occurred through an increase in the proportion of women (60 hours per week or more) increased dramatically. in part-time jobs. Figure 13, which depicts the These finding indicate that if we were to weigh the distribution of hours of work among employed male participation rate of women by hours worked, the and female populations in urban areas in 1988 and total labor input would, in no doubt, show a stronger 2006 indicate the opposite. For both men and women, increase than the participation rate itself. Figure 12: Hours of Work by Sex in 2006 50 45 40 35 % of employed 30 Men 25 Women 20 15 10 5 0 1-39 40-59 60+ Hours of work per week Source: HLFS web data base, TUIK (http://www.tuik.gov.tr). Note: Covers individuals age 15 plus. Figure 13: Hours of Work in Urban Areas by Sex in 1988 and 2006 Source: HLFS web data base, TUIK (http://www.tuik.gov.tr). Note: Covers individuals age 15 plus. Determinants of and Trends in Labor Force Participation of Women in Turkey 18 5.2.3. Occupational Distribution of the Employed To get a better sense of the distribution outside by Sex and Year4 agriculture we turn to urban areas.5 Table 3 shows that even in urban areas, the proportion of women 69. The occupations held by working women are engaged in agriculture is non-negligible, though its shown in Table 2. About 60 percent of working women share is on the decline. While in 1988, 14.4 percent are found in agricultural occupations. This result stems of urban working women were in agriculture, in 2003, this figure dropped to 9.7 percent. Besides agriculture, from the dominance of agricultural activities in rural women are over-represented among professionals/ areas, which is evident from the observation that 24.7 technical workers and clerical workers: excluding percent of men are also in agriculture, and the fact that agricultural workers, about half of the working women rural women make up six out of 10 working women in in 2003 were found in one of these two occupations as Turkey. opposed to 18 percent of men. When the distribution is Table 2: Occupational Distribution by Sex in 2003 (%) ISCO-68 Men Women Professional, Technical and Related Workers 7.85 10.70 Administrative and Managerial Workers 3.93 0.96 Clerical and Related Workers 5.11 8.80 Sales Workers 14.01 5.25 Service Workers 12.47 5.93 Agricultural, Animal Husbandry and Forestry Workers, 24.66 59.05 Fishermen and Hunters Production and Related Workers, Transport Equipment 31.96 9.31 Operators and Laborers Source: HLFS web data base, TUIK (http://www.tuik.gov.tr). Note: Covers individuals age 15 plus. 2003 is the final year in which ISCO-68 was used. Table 3: Occupational Distribution in Urban Areas by Sex (%) Men Women ISCO-68 1988 2000 2003 1988 2000 2003 Professional, Technical and Related Workers 8.80 10.02 10.48 19.74 25.24 23.89 Administrative and Managerial Workers 3.76 4.14 5.27 1.30 2.10 2.30 Clerical and Related Workers 6.72 7.44 6.99 21.32 22.50 20.19 Sales Workers 15.81 18.72 18.89 6.33 10.32 11.25 Service Workers 12.85 14.03 14.93 10.06 12.28 13.78 Agricultural, Animal Husbandry and 4.43 2.66 3.10 14.43 8.71 9.67 Forestry Workers, Fishermen and Hunters Production and Related Workers, Transport 47.62 42.97 40.33 26.82 18.88 18.93 Equipment Operators and Laborers Source: HLFS web data base, TUIK (http://www.tuik.gov.tr). Note: Covers individuals age 15 plus. 2003 is the final year in which ISCO-68 was used. 4 The occupational classification used by TUIK from 1988 to 2003 followed ISCO-68. From 2004 onwards, occupations were categorized in line with ISCO-88. To ease comparison across time we report 2003, the final year that ISCO-68 was used. 5 This is partly due to the definition of urban places which relies on the size of the settlement rather than other characteristics such as economic activity or services rendered to the population. Determinants of and Trends in Labor Force Participation of Women in Turkey 19 examined over time, a fall in the share of agricultural the overwhelming majority of such women work, and production workers are observed. In contrast, meant a fall in unpaid family work as well. Figure 14 the proportion of men and women engaged in other shows the declining trend in unpaid family work for occupations (with a single exception) showed an women, which is occurring in parallel to the decline in increase. In the case of women, the sharpest increases self-employment among men, depicted in Figure 15. occurred among professionals, sales and services 71. For the first time in 2005, the share of unpaid workers. Quite interestingly, the proportion of women family workers among working women fell short of engaged in clerical jobs did not record an increase and wage workers. Figure 14 shows that the decline in in fact, slightly fell. unpaid family work and increase in wage work has 5.2.4. Employment Status by Sex and Year been happening, pretty much uninterrupted, since 1988 so that we could expect this trend to continue 70. Another important stylized fact about women’s em- in the coming years as well. In the case of men, we ployment in Turkey is that a sizeable proportion works also observe that an increasing proportion is turning to as unpaid family workers. In 2006, such women consti- wage employment. In 2006, wage earners constituted tuted 39 percent of all working women. However, the 60 percent of all working men, while own account declining share of agriculture in employment, where workers 34 percent. Figure 14: Employment Status of Women by Year 80 70 60 % of employed 50 Wagework 40 Own account Unpaid 30 20 10 0 1988 1989 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 Source: HLFS web data base, TUIK (http://www.tuik.gov.tr). Note: Covers individuals age 15 plus. For data see Appendix Table A4. Figure 15: Employment Status of Men by Year 70 60 50 % of employed 40 Wagework Own account 30 Unpaid 20 10 0 1988 1989 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 Source: HLFS web data base, TUIK (http://www.tuik.gov.tr). Note: Covers individuals age 15 plus. For data see Appendix Table A4 Determinants of and Trends in Labor Force Participation of Women in Turkey 20 72. In urban areas, wage employment has been the among whom we also see a rising share of wage major form of employment for women (and for men) workers (see Figure 16). It is also interesting to note and there is indication that more and more women from Figure 16 that the two years in which significant are seeking wage work. Indeed, Figure 16 shows declines in wage employment among women that while in 1988, 74.8 percent of women were occurred (in 1994 and 1999) were both crisis years.6 employed as wage earners, this figure increased Interestingly, during these two years, the proportion of to 80.3 percent in 2006. Similar developments men employed as wage earners has slightly increased are taking place in the employment status of men, (Figure 17). Figure 16: Employment Status of Women in Urban Areas by Year 90 80 70 60 % of employed Wagework 50 Own account 40 Unpaid 30 20 10 0 1988 1989 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 Source: HLFS web data base, TUIK (http://www.tuik.gov.tr). Note: Covers individuals age 15 plus. For data see Appendix Table A5. Figure 17: Employment Status of Men in Urban Areas by Year 80 70 60 % of employed 50 Wagework 40 Own account Unpaid 30 20 10 0 1988 1989 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 Source: HLFS web data base, TUIK (http://www.tuik.gov.tr). Note: Covers individuals age 15 plus. For data see Appendix Table A5. 73. To complete the picture, in Figure 18, we also work, on the other hand, has been on the increase. show the changes in the status of employment among 74. It is also worth noting that wage work often takes rural women over time. As noted earlier, the proportion the form of regular employment (see Appendix Table of women employed as unpaid family workers A7). For instance, in 2006, 92 percent of women has been continuously declining since 1988. Wage wage workers in urban areas and 71.6 percent in rural 6 However, it must be noted that a similar decline was not observed during the 2001 crisis, which was a deeper recession that those in 1994 and 1999. Determinants of and Trends in Labor Force Participation of Women in Turkey 21 areas were regular workers. The corresponding seems to be on the rise (see Appendix Table A7). figures for men were 90.1 percent and 78.2 percent, While during the 1988-1994 period, the proportion respectively. While the proportion of women wage of wage workers in regular employment fluctuated workers employed regularly in urban areas has around 55-65 percent, after registering a dip in 1995 fluctuated around 90 percent over the 1988-2006 at 47.6 percent started to increase and reached 72.8 period, in the case of rural women, this proportion percent in 2000. Figure 18: Employment Status of Women in Rural Areas by Year 100 90 80 70 % of employed 60 Wagework 50 Own account 40 Unpaid 30 20 10 0 1988 1989 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 Source: HLFS web data base, TUIK (http://www.tuik.gov.tr). Note: Covers individuals age 15 plus. For data see Appendix Table A6. 5.3. Cohort Profiles to generate these profiles, these age effects could also stand for cohort and calendar time effects as 5.3.1. Female Labor Force Participation Rates by data points for different ages correspond to different Age for Successive Birth Cohorts birth-cohorts and calendar times as well. To capture age effects, we would ideally need panel data where 75. As discussed earlier, we see a hump-shape in the we can follow the same birth-cohorts over time. age-profile of labor force participation of women both However, even when panel data are not available, it in rural and urban areas – much weaker in rural areas, is possible to construct a synthetic panel using a series though–. However, since a single-cross section is used of cross-section data.7 We construct these synthetic Table 4: Cohort Definitions Cohort Year of Birth Age in 1990 Age in 1995 Age in 2000 Age in 2005 1 1986-1990 15 2 1981-1985 15 20 3 1976-1980 15 20 25 4 1971-1975 15 20 25 30 5 1966-1970 20 25 30 35 6 1961-1965 25 30 35 40 7 1956-1960 30 35 40 45 8 1951-1955 35 40 45 50 9 1946-1950 40 45 50 55 10 1941-1945 45 50 55 60 11 1936-1940 50 55 60 12 1931-1935 55 60 13 1926-1930 60 7 See Browning et al. (1985) and Attanasio (1998). Determinants of and Trends in Labor Force Participation of Women in Turkey 22 cohorts based on four rounds of HLFS: 1990, 1995, the average participation rates would mask the recent 2000 and 2005. The cohort definitions are given in developments by averaging participation rates over Table 4. For instance, individuals from the 1971-1975 younger and older birth cohorts. birth-cohort are 15 to 19-year-old in the 1990 data, 20 to 24-year-old in the 1995 data, 25 to 29-year-old in 5.3.1.1 Women in Urban Areas the 2000 data, and 30 to 34-year-old in the 2005 data. 77. Indeed, Figure 19 which depicts age-participation 76. Although in these surveys, the same people are profiles by birth cohorts for urban women shows that not followed over time we can nevertheless learn except for the youngest age-group (15-19 years), in all about the changing behavior by looking at successive other age groups, labor force participation of women cohorts and their participation rates since individuals tends to be higher for younger than older cohorts. For of the same cohort would share, for instance, similar instance, while the participation rate among the 25-29 educational opportunities, birth control technologies, year-old women in the 1960-1965 cohort was on the and exhibit similar attitudes towards birth control, order of 20.7 percent, this rate was recorded at 22.3 marriage and schooling than individuals of different percent in the 1965-1970 birth cohort, 24.9 percent in cohorts. For instance, we would expect women aged the 1970-1975 birth cohort and finally at 29.3 percent 15-19 from the 1986-1990 birth cohort to have the in the 1975-1980 birth cohort (see also Appendix lowest labor force participation rate in 2006, since Table A8). The observed changes would reflect both this is the only cohort that has been affected from the changing behavior (changing attitudes towards the extension of compulsory schooling from five participation, higher levels of schooling, etc.) – cohort to eight years in 1997. To the extent that younger effects, as well as changing economic circumstances generations participate more in the labor market, that affect respective cohorts –time effects. Figure 19: Female Labor Force Participation in Urban Areas by Birth Cohort 35 30 25 1985-1990 1980-1985 20 1975-1980 LFPR 1970-1975 15 1965-1970 1960-1965 10 1955-1960 1950-1955 5 1945-1950 0 15-19 20-24 25-29 30-34 35-39 40-44 45-49 50-54 55-59 Source: HLFS web data base, TUIK (http://www.tuik.gov.tr). Note: For data see Appendix Table A8. 78. Comparing Figures 9 and 19, we can deduce is distinctly different from that would obtain from a that age-participation profiles that are not corrected regular cross-section. Figure 6 indicates the following for cohort and time effects somewhat exaggerate the participation rates for the same age groups: 30 percent, decline in participation by age at older age categories. 26.3 percent, 25.9 percent and 21.6 percent. Although participation is clearly higher during prime age years, significant declines in participation seem 79. A regression analysis on the synthetic panel used to occur only beyond age 40. Consider for instance to construct Figure 19 confirms that participation the 1960-1965 birth cohort: participation rates for this is higher among younger than older cohorts (Figure birth cohort at ages 20-24, 25-29, 30-34 and 35-39 20). Controlling for age, we find, for instance, that the were on the order of 20.7 percent, 21.8 percent, 20.5 average participation rate of women from the 1976- percent and 20.2 percent, respectively. This profile 1980 cohort is roughly five and a half percentage Determinants of and Trends in Labor Force Participation of Women in Turkey 23 points (p<0.01) higher than their counterparts from the Figure 22: Year Effects (in percentage points) on Labor 1961-1965 cohort. The regression results also indicate Force Participation Rate in Urban Areas age to be an important determinant of participation. 1 However, controlling for cohort effects, the difference in the participation rates among women aged 20- .5 39 is not more than two percentage points (p<0.01 for all categories). In contrast, there is more than 10 yreff 0 0 yreff percentage point difference between the participation rates of 15-19 and 20-24 year-old women. Moreover, -.5 even bigger differences are observed between women of prime age years and those older. For instance, the -1 difference in the participation rates between women 1990 1995 2000 2005 aged 25-29 and 55-59 is on the order of 15 percentage points. The predicted changes in the participation Note: Year effects add up to zero by assumption. rates of women of different ages, corrected for cohort and time effects, are shown in Figure 21. Any effort 80. Finally, the year effects, displayed in Figure 22, to increase the participation rate of women will not indicate that labor force participation of women in only need to increase the participation rates of prime- urban areas were more likely in 1990 and 2005. age women but reduce the sharp decline that comes at age 40 as well, which is primarily a result of the 5.3.1.2 Women in Rural Areas availability of retirement at these early ages. 81. To better asses the falling labor market participation among rural women, we carry out a cohort analysis Figure 20: Cohort Effects (in percentage points) on Labor Force Participation Rate in Urban Areas (Baseline Cohort = using a synthetic panel. The age-participation profiles 1986-90) depicted for various cohorts in Figure 23 look very different from the urban profile (see Figure 19). 5 While for urban women we arrived at the conclusion that younger cohorts are more likely to participate in the labor market, for rural women the opposite observation is made: for all age categories under consideration, women from younger generations have coheff coheff 0 0 lower likelihood of participating in the labor market than their counterparts from older generations and furthermore, the fall in participation rates from one birth cohort to the next especially among the recent -5 86-90 76-80 66-70 56-60 46-50 36-40 26-30 cohorts is very drastic. To give an example, consider the participation rates for 25-29 year-old women Figure 21: Age Effects (in percentage points) on Labor from four successive birth cohorts: 54.5 percent for Force Participation in Urban Areas (Baseline Age Group = the 1960-1965 cohort, 52.9 percent for the 1965-1970 15-19) cohort, 44.4 percent for the 1970-1975 cohort and 35.6 percent for the 1975-1980 cohort (see also Appendix 10 Table A9). In the face of rising education levels among rural women and declining fertility rates, illustrated later in the report, these findings are surprising. 5 82. The age-participation profiles also show a very ageff strong age effect: holding the birth cohort constant, 0 participation drops as women age. To give an example, consider the 1970-1975 birth cohort. For this birth co- hort, participation is highest for women aged 20-24 at -5 15 20 25 30 35 40 45 50 55 60 58 percent. Participation then drops to 54.2 percent for the 25-29 year age group, further to 44.4 percent for Determinants of and Trends in Labor Force Participation of Women in Turkey 24 the 30-34 year age group and finally, to 38.5 percent sectional analyses underestimate the age effect on par- for the 35-39 year-age group. These figures indicate ticipation due to the higher likelihood of participation that the age-participation profiles obtained from cross- among older generations at all ages. Figure 23: Female Labor Force Participation in Rural Areas by Birth Cohort 70 60 1985-1990 50 1980-1985 1975-1980 40 1970-1975 LFPR 1965-1970 30 1960-1965 1955-1960 20 1950-1955 1945-1950 10 0 15-19 20-24 25-29 30-34 35-39 40-44 45-49 50-54 55-59 Source: HLFS web data base, TUIK (http://www.tuik.gov.tr). Note: For data see Appendix Table A9. Figure 24: Age Effects (in percentage points) on Labor Figure 26: Year Effects (in percentage points) on Labor Force Participation in Rural Areas (Baseline Age Group = Force Participation Rate in Rural Areas 15-19) 0 2 -20 1 ageff yreff -40 0 -60 -1 15 20 25 30 35 40 45 50 55 60 1990 1995 2000 2005 Note: Year effects add up to zero by assumption. Figure 25: Cohort Effects (in percentage points) on Labor Force Participation, Rural Areas (Baseline Cohort = 1986- 90) 83. Age effects from the decomposition of age, cohort, and time effects in the labor force participation decision 80 of rural women are presented in Figure 24. Unlike the age effects in urban areas – where it displayed a hump 60 shape with a flat region between the ages of 20 and 40 – the age effects in rural areas display a monotonically decreasing profile. In rural areas, women become less coheff coheff 40 40 likely to participate in the labor market as they age. 20 84. Cohort effects for women in rural areas are illus- trated in Figure 25. According to this, younger cohorts 0 86-90 76-80 66-70 56-60 46-50 36-40 26-30 of women in rural areas are less likely to participate in the labor market. This is contrary to the finding in Determinants of and Trends in Labor Force Participation of Women in Turkey 25 urban areas where younger cohorts are more likely to but not the mean hours of work. In our analysis of participate. the change in the weekly work hours by age and birth cohorts, we, therefore, look at the distribution 85. Finally, time effects from the decomposition of women across different work hour categories. For analysis are presented in Figure 26. Women in rural this purpose, we consider the proportion of women areas had a higher propensity to participate in the labor working less than 40 hours and those working for 50 market in 1995 than in the other three years. hours or more. 5.3.2. Weekly Hours Worked by Age for Successive 87. Figure 27 depicts the proportion of women who work Birth Cohorts for less than 40 hours per week by age and birth cohort 5.3.2.1. Urban Women in urban areas. The proportion working for less than 40 hours per week increases in age. However, it is also the 86. The TUIK data base, from where we derive the case that, among older cohort, the proportion of women cohort profiles, provides the distribution of work hours working for less than 40 hours per week is higher. Figure 27: Proportion of Women Working for Less than 40 hours per Week by Age and Birth Cohort in Urban Areas 45 40 35 1985-1990 1980-1985 30 1975-1980 25 1970-1975 1965-1970 % 20 1960-1965 1955-1960 15 1950-1955 10 1945-1950 5 0 15-19 20-24 25-29 30-34 35-39 40-44 45-49 50-54 55-59 Source: HLFS web data base, TUIK (http://www.tuik.gov.tr). 88. Figure 28 displays the cohort effects from likely to participate in the labor market. Here, we the decomposition of age, cohort, and time further show that these younger cohorts of effects in working less than 40 hours per week urban women are also less likely to have a short for urban women. Previously, we showed that work week compared to older cohorts. In other younger cohorts of urban women were more words, the higher propensity of labor force Figure 28: Cohort Effects (in percentage points) on Working Figure 29: Age Effects (in percentage points) on Working for Less than 40 Hours per Week, Urban Areas (Baseline for Less than 40 Hours per Week, Urban Areas (Baseline Cohort = 1986-90) Age Group = 15-19) 30 15 20 10 coheff ageff 10 5 0 0 86-90 76-80 66-70 56-60 46-50 36-40 26-30 15 20 25 30 35 40 45 50 55 60 Determinants of and Trends in Labor Force Participation of Women in Turkey 26 participation of younger cohorts of urban women in The cohort effects, displayed in Figure 31, reveal Turkey is not driven from their part-time work. that younger cohorts are more likely to work for longer hours. This is also consistent with the 89. Age effects in working for less than 40 hours previous finding that these later cohorts were also per week are given in Figure 29. Overall, as age less likely to work for less than 40 hours per week. increases, working for less than 40 hours per week Therefore, we can claim that younger generations of becomes more likely. The rise in shorter working women are not only more likely to participate in the hours after age 50 is more prominent. labor market but also they are more likely to work 90. Next, we decompose the cohort, age, and time for longer hours. effects in working for long hours, in particular for longer than 50 hours per week. Figure 30 plots the 91. Age effects in working for more than 50 hours data points used in the decomposition analysis, per week exhibit a surprising trend. As age increases, and Figures 31 and 32 display the cohort effects the probability for working more than 50 hours per and age effects, respectively, obtained from this week increases. This is also apparent from Figure decomposition analysis. 32. Figure 30: Proportion of Women Working for 50 Hours or More per Week by Age and Birth Cohort in Urban Areas 70 60 1985-1990 50 1980-1985 1975-1980 40 1970-1975 % 1965-1970 30 1960-1965 1955-1960 20 1950-1955 1945-1950 10 0 15-19 20-24 25-29 30-34 35-39 40-44 45-49 50-54 55-59 Source: HLFS web data base, TUIK (http://www.tuik.gov.tr). Figure 31: Cohort Effects (in percentage points) on Working Figure 32: Age Effects (in percentage points) on Working for More than 50 Hours per Week, Urban Areas (Baseline for More than 50 Hours per Week, Urban Areas (Baseline Cohort = 1986-90) Age Group = 15-19) 0 20 15 -20 10 coheff ageff 5 -40 0 -60 -5 86-90 76-80 66-70 56-60 46-50 36-40 26-30 15 20 25 30 35 40 45 50 55 60 Determinants of and Trends in Labor Force Participation of Women in Turkey 27 6. Education, Marital Status, percent to 19.6 percent, the proportions of men and Fertility, Migration and Labor Force women with just primary school education became Participation equal and the proportion of women with more than primary school education more than doubled (Table 5). 6.1. Education 93. The urban-rural divide in educational attainment 6.1.1. Schooling attainment by age over time for among women is as drastic as the gender education female population gap. In 2006, while the proportion of illiterate women in urban areas was on the order of 14.8 percent, this 92. The majority of women in Turkey do not have figure was 27.7 percent among rural women (Table more than primary school education. This situation, no 6). Likewise, while the proportion of women with doubt, negatively affects their participation and puts more than primary school education was near 40 percent in urban areas, it was half this figure among them at a disadvantage in comparison to men, who in women in rural areas. It also interesting to note that terms of schooling fare somewhat better than women. the improvement in the educational attainment of In 2006, one out of five women was illiterate in women with no or limited education (no diploma) has comparison to only 4 percent of men. The proportion been faster in urban than rural areas with the result with more than primary school education, on the that the education gap among this group of women other hand, was limited to one third of the female has actually increased over time. However, it is also population, as opposed to half the male population.8 the case that the proportion of women with more than However, the schooling levels of both men and primary school education has grown at a faster rate in women are improving. In less than two decades, the rural than urban areas, helping somewhat to close the proportion of illiterate women dropped from 33.9 gap among urban and rural women. Table 5: Distribution of Male and Female Population by Education (%) Source: 1988, 2000 and 2006 HLFS, http://www.tuik.gov.tr Note: Covers individuals age 15 plus. 94. The urban-rural divide is not unique to women but differentiated for men as it was for women. Similar exits for men as well (Table 7). To give an example, to women, faster improvements occurred in the while in 2006, 6.2 percent of rural men were illiterate proportion of men with more than primary school this was the case for 2.8 percent of urban men. education in rural areas, helping to close the urban to rural gap. In the face of continuing migration from Similarly, while almost 60 percent of urban men had rural to urban areas, the urban-rural divide deserves above primary school education, the corresponding special attention especially from the perspective of rate among rural men was limited to 40 percent. As women’s participation in the labor market since, as observed for women, relatively faster improvements will be demonstrated shortly, the participation rate occurred in educational attainment at the lower end of women with less than primary school education is in urban than rural areas, though the rate was not as extremely low in urban areas. 8 In 1997, compulsory schooling was extended from five to eight years and primary and junior-high schools were merged under Basic Education Schools. To ease comparison across time, we have created the ‘secondary school’ category that refers to the graduates of junior-high schools and more recently, of Basic Education. Determinants of and Trends in Labor Force Participation of Women in Turkey 28 Table 6: Distribution of Urban and Rural Female Population by Education (%) Source: 1988, 2000 and 2006 HLFS, http://www.tuik.gov.tr Note: Covers individuals age 15 plus. Table 7: Distribution of Urban and Rural Male Population by Education (%) Source: 1988, 2000 and 2006 HLFS, http://www.tuik.gov.tr Note: Covers individuals age 15 plus. 95. To get a better sense of the extent of recent educa- schooling of prime age women still lagged behind the tional developments, we look at changes in average average schooling levels of their male counterparts in education levels of different age groups over time. 1988. Judging from the gender schooling gap among Figure 33 shows that younger women and men have the 20-24 year-olds, the education gap between the considerably higher levels of schooling than older prime age men and women is not likely to improve individuals and that for all age groups, the average significantly in the near future.9 years of schooling is on the rise. For instance, the average schooling of 20-24 year-old men and women 96. Turning to the rising labor force participation – those who would have completed their schooling rate among urban women, it might be instructive to and entered the labor market - has improved by 2.2 look at improvements in average education levels of years from 1988 to 2006. However, it is striking to urban residents over time. Is the rising labor force note that even after eight years, the average years of participation of urban women consistent with changes 9 The fact that gender education gap exists and is not likely to close in the very near future, has important implications for the labor market. From the perspective of women’s labor market entry, it puts them at a disadvantage compared to men. From the perspective of labor market attachment, by reducing women’s earnings and other promotion opportunities vis-à-vis men, makes them more likely to exit the labor market. Determinants of and Trends in Labor Force Participation of Women in Turkey 29 in their levels of schooling? Indeed, Table 8 shows that composition of the labor force did not change as there has been improvement in the schooling levels of drastically as the composition of the urban population. urban women. For instance, while in 1988, university For instance, while in 1988, 12.9 percent of the labor graduates constituted only 2.8 percent of the urban force was made up of university graduates, the figure female population (15+), in 2006 this figure became in 2006 was about twice this rate recorded at 28.5 8.1 percent (Table 8). Similar improvements took place percent. While the changing composition of the urban in all educational categories above primary school. work force towards more educated women would In contrast, the proportion of illiterates, functional certainly explain the rising participation rates, it is of literates and primary school graduates dropped. interest to also look at the changes in the participation rates of women in different schooling levels over 97. However, it is interesting to note that the education time. Figure 33: Average Years of Schooling in 1988 and 2006 by Age and Sex 10 9 8 7 average years of schooling 6 Men-06 Women-06 5 Men-88 4 Women-88 3 2 1 0 15-19 20-24 25-29 30-34 35-39 40-44 45-49 50-54 55-59 60-64 65+ age Source: 1988 and 2006 HLFS, http://www.tuik.gov.tr Table 8: Distribution of Female Population and Labor Force by Education -Urban Areas Source: 1988, 2000 and 2006 HLFS, http://www.tuik.gov.tr Note: Covers individuals age 15 plus. Determinants of and Trends in Labor Force Participation of Women in Turkey 30 6.1.2. Female labor force participation rates by 99. The role education plays in the labor force educational attainment over time participation of women becomes more apparent when we consider urban areas. Participation rates shown in 98. Labor force participation increases with education Figure 35 indicate that urban labor market is practically for both men and women. Figure 34 illustrates the closed for women who have no education. The labor participation rates at different schooling levels. For force participation rate among illiterate women in urban both men and women, the highest participation rates Turkey in 2006 was 5.6 percent. The corresponding are recorded for university graduates. Although men’s rate for men was on the order of 36.4 percent. The participation rates change with schooling, the observed participation rates gradually improve for women as changes between different categories of schooling are they become more educated, ending with a drastic not as drastic as those observed for women. What jump for those with a higher education diploma. While matters for men’s labor market participation seems to the labor force participation for women with vocational be a primary school diploma, while for women this training was on the order of 35.6 percent in 2006, it does not open many doors. jumps to 69.8 percent for university graduates. Figure 34: Labor Force Participation by Education in 2006 University Vocational High School Women Secondary Men Primary School No Diploma Illiterate 0 10 20 30 40 50 60 70 80 90 Source: 2006 HLFS, http://www.tuik.gov.tr Note: Covers individuals age 15 plus. For data on women see Appendix Table A10. Figure 35: Labor Force Participation by Education in Urban Areas in 2006 University Vocational High School Women Secondary Men Primary School No Diploma Illiterate 0 10 20 30 40 50 60 70 80 90 Source: 2006 HLFS, http://www.tuik.gov.tr Note: Covers individuals age 15 plus. For data on women see Appendix Table A10. Determinants of and Trends in Labor Force Participation of Women in Turkey 31 100. Figure 36 shows the changes in the participation percent, this rate declined to 27.6 percent in 2006. rates of men and women university graduates in urban Similar changes occurred for the male graduates of areas over time. Participation rates of university general high schools as well. In the case of vocational graduate women, although quite high in all years, school graduates, while men’s rates have remained have recorded a decline over time. In comparison to rather stable, women’s participation rates showed a the early 90s during which participation rates among decline over time. university graduate women were 80 percent, this rate declined and stabilized around 70 percent in the 102. Finally, Figures 38 and 39 illustrate the 2000s. It is again interesting to note that a similar trend participation rates for those with secondary school although at a smaller scale has been observed among education and less. Among women, in none of the university educated men as well. categories do we see significant changes worth mentioning. However, in the case of men a fall in 101. Figure 37 illustrates the participation rates for participation is observed in all categories except for general and vocational high school graduates over time. secondary school graduates for whom participation For both groups of women, a declining trend is ob- has been roughly constant and functional literates served. While in 1988, the participation rate of general whose participation rates after rather sharp decreases high school graduate women was on the order of 44.3 over several years looks on the rise. Figure 36: LFPR of University Graduates in Urban Areas 100 95 90 85 80 LFPR 75 Men 70 Women 65 60 55 50 1988 1989 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 Source: HLFS web data base, TUIK (http://www.tuik.gov.tr). Note: Covers individuals age 15 plus. For data on women see Appendix Table A11. Figure 37: LFPR of High School Graduates in Urban Areas 90 80 70 60 High-Men 50 Voc-Men LFPR High-Wom 40 Voc-Wom 30 20 10 0 1988 1989 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 Source: HLFS web data base, TUIK (http://www.tuik.gov.tr). Note: Covers individuals age 15 plus. For data on women see Appendix Table A11. Determinants of and Trends in Labor Force Participation of Women in Turkey 32 Figure 38: LFPR of Primary and Secondary School Graduates in Urban Areas 100 90 80 70 60 Primary-M Sec-M LFPR 50 Primary-W 40 Sec-W 30 20 10 0 1988 1989 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 Source: HLFS web data base, TUIK (http://www.tuik.gov.tr). Note: Covers individuals age 15 plus. For data on women see Appendix Table A11. Figure 39: LFPR of Illiterates and Functional Literates in Urban Areas 80 70 60 50 Illit-Men 40 N/Dip-Men Illit-Wom 30 N/Dip-Wom 20 10 0 1988 1989 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 Source: HLFS web data base, TUIK (http://www.tuik.gov.tr). Note: Covers individuals age 15 plus. For data on women see Appendix Table A11. 103. The above findings indicate that had the were more of select group. As the numbers increase, participation rates of high school and university the group becomes more heterogeneous and so is the graduates remained at levels recorded in the 1990s, participation behavior. We investigate this conjecture women’s current labor force participation rates would and other possible reasons for the falling participation have been considerably higher. It is interesting to note among highly educated women later in the report (see that conditional on schooling, the participation rates Section 8). are either stagnant or falling for women yet, the overall 6.2. Marital Status participation in urban areas are increasing, albeit rather slowly. This is to do with the shift in the composition 6.2.1. Marital status of women by age over time of the workforce towards more educated women who have higher participation rates. 105. Marriage is almost universal in Turkey, which makes married women’s labor market participation 104. On the other hand, the declining participation all the more important and requires monitoring the rates among highly educated women must be to do changes occurring in married women’s participation with the composition of the university graduates. We over time. The DHS data indicate that nearly 98 percent conjecture that when university graduates constituted a of women marry by age 49. In contrast, divorce is an much smaller proportion of the urban population, they unlikely event, with its rate estimated at less than 1 Determinants of and Trends in Labor Force Participation of Women in Turkey 33 percent among 15-49 year-old women. Furthermore, almost universal and takes place when women are marriage occurs early on in life; the average age at first young, Figure 40, which shows the proportion of never marriage is 20.7 years among women aged 15-49. married women over time, indicates that it is being gradually postponed to later years. The proportion 106. The HLFS data conf irm the near universal of never married women has registered the sharpest marriage rates. In 2006, only 2.1 percent of women increase among the 20-24-year-old women, followed aged 50-54 had never been married. Though marriage is by the 25-29-year age group. Figure 40: Proportions of Never Married Women by Age over Time 100 90 80 proportion never married (%) 70 15-19 60 20-24 25-29 50 30-34 40 35-39 40-44 30 20 10 0 1988 1989 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 Source: HLFS web data base, TUIK (http://www.tuik.gov.tr). 107. Marriage, as an institution, remains strong in 89 percent of urban and 89.6 percent of rural women urban areas as well. Despite the considerably different had been married at least once. education profiles of urban and rural women illustrated earlier, the incidence of marriage turns out to be very 108. Figure 41 shows the changes in the proportion similar among rural and urban women: by age 50-54, of never married women by age groups in urban ar- 97.7 percent of urban and 98.2 percent of rural women eas. As noted for the country at large, there is a ten- had been married at least once. However, the timing of dency for delaying marriage to later ages. Again the marriage somewhat differs, with rural women marrying sharpest increases have been observed among the younger. Even then, the marriage behaviors of urban and 20-24 and 25-29-year age groups. These develop- rural women converge relatively early; by age 30-34, ments are consistent with the increasing participation Figure 41: Proportions of Never Married Women in Urban Areas by Age over Time 100 90 80 proportion nver married (%) 70 15-19 60 20-24 25-29 50 30-34 40 35-39 40-44 30 20 10 0 1988 1989 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 Source: HLFS web data base, TUIK (http://www.tuik.gov.tr). Determinants of and Trends in Labor Force Participation of Women in Turkey 34 of women in urban areas since, as will be discussed whom the above rates were 86.4 percent for married, later in the report marital status is an important pre- 81.1 percent for divorced and 71.8 percent for single dictor of labor market participation in urban areas. men. The ‘age effect’ – the fact that single men tends to be younger, less experienced and more likely to at- 6.2.2. Female labor force participation rates by tend school – explain the difference in the participa- marital status and age over time tion rates between married and single men. In the case of women, the fact that married women participate less 109. Table 9 shows the drastically different labor than their single counterparts indicate that in addition force participation behavior of men and women by to the age effect, there is a ‘marriage effect’ that works marital status in Turkey. While in the case of men, the in the opposite direction to reduce the labor market highest participation rate is recorded for married men, participation of women. for married women, we observe the second lowest rate after widowed women, who are much older than their 110. Table 9 also shows that for both men and wom- married counterparts. In 2006, while the participation en, a general decline in the participation rates has been rate of married women was 23.1 percent, for single and observed in all categories over time, which is consist- divorcee women significantly higher rates were ob- ent with the overall decline in the participation rates served at 34.3 percent and 42.1 percent, respectively. of men and women. Divorcee women constitute the This pattern in distinctly different from that of men for only exception to this general pattern for whom we Table 9: Labor Force Participation by Marital Status % % 1988 2000 2006 Change Change (in percentages) Male Female Male Female Male Female Male Female 1988-2006 1988-2006 Single 71.8 47.8 59.4 35 58 34.3 -19.2 -28.2 Married 86.4 32 81.1 25.2 78.2 23.1 -9.5 -27.8 Divorced 81.1 41.5 72.5 41 68.2 42.1 -15.9 1.4 Widowed 30.1 16 29.7 11.5 22.4 10 -25.6 -37.5 Overall 81.2 34.3 73.7 26.6 71.5 24.9 -11.9 -27.4 Source: 1988, 2000 and 2006 HLFS, http://www.tuik.gov.tr Note: Covers individuals age 15 plus. Table 10: Labor Force Participation in Urban Areas by Marital Status % Change % Change Male Female 1988- 1988- 2006 2006 Source: 1988 and 2006 HLFS, http://www.tuik.gov.tr Note: Covers individuals age 15 plus. Determinants of and Trends in Labor Force Participation of Women in Turkey 35 observe a slight improvement in participation. This single and married women, the increase has been lim- deviation from the general behavior might be to do ited to less than 2.5 percentage points, there has been with the changing composition of the divorcees as a huge jump (on the order of almost 10 percentage societal attitudes towards divorce changes. points) in the participation rates of divorcee women. However, since this group constitutes a very small 111. Table 10 repeats the same exercise for urban portion of the female workforce, the average partici- areas showing participation rates and changes in them pation rate has not been affected to that degree. over time. Owing to the wider schooling gap and the tendency to marry younger in rural areas, single men’s 113. Given the low rates of participation among participation is lower in urban areas. However, this married women and their share in potential workforce, gap seems to have closed over time so that except for if women’s participation rate is to be increased, this married and widowed women, participation rates in would heavily depend on the change in the parti- urban areas for men and women are very similar to cipation behavior of married women. It is therefore, the overall rates obtained for the country at large. The important to understand the reasons behind the low lower participation rates of married women in urban participation rates of married women. Plausible areas indicate that the “marriage effect� is stronger in explanations include child bearing, societal norms urban than rural areas. The lower participation rates of divorcee women, on the other hand, indicate that the that see women’s proper role as motherhood, lack of urban labor market is closed for older women. affordable child care facilities, reliance on informal arrangements for child care, and female seclusion. 112. Looking at changes over time, we note the While quantifying the role of tradition and societal general fall in the participation rates of men irre- values in determining women’s participation rate is spective of their marital status. However, with the hard, the results of multivariate analysis undertaken exception of widowed women, in all other catego- and discussed later in the report shows that the number ries women’s participation has increased. While for of children affects women’s participation negatively. Figure 42: LFPR of Women by Marital Status and Age in Urban Areas 70 60 50 40 LFPR 30 20 10 0 15-19 20-24 25-29 30-34 35-39 40-44 45-49 50-54 55-59 60-64 married Age not married Source: 2006 HLFS, TUIK. Figure 43: LFPR of Women by Marital Status and Age in Rural Areas 50 45 40 35 30 LFPR 25 20 15 10 5 0 15-19 20-24 25-29 30-34 35-39 40-44 45-49 50-54 55-59 60-64 married Age not married Source: 2006 HLFS, TUIK. Determinants of and Trends in Labor Force Participation of Women in Turkey 36 114. Finally, we examine the age-participation pro- first-birth is 22.5 years. An average woman then expects files of women by marital status. Figure 42 displays to be married with one child before she reaches 25 this comparison in urban areas and Figure 43 for rural years and a second child before she reaches age 30 areas. In urban areas, we find that married women are (DHS, 2003). Perhaps, more importantly, by the end much less likely to participate in the labor market at of their productive years, less than 2 percent of women all ages. For instance, for 25- to 29-year-old women, fail to have a child. Notwithstanding these general while the participation rate for married women is be- patterns, the fertility rate is closely associated women low 20 percent, it is above 60 percent for non-married women’s education level. As given in Appendix Table women. This gap diminishes in later ages once wom- B1, while an illiterate woman expects to have 3.7 en with high propensity to work also get married. In children before the end of her reproductive years, the rural areas, between the ages of 20 and 35, married corresponding figure among women with high school women are also less likely to participate; however, the education or more is 1.4 children. gap is much narrower than that in urban areas. Moreo- ver, there is a reversal after age 40. Nonetheless, it is 117. Given the importance of children in the lives of hard to interpret this reversal because very few women the majority of women in Turkey, in this subsection, above this age are non-married in rural areas. we examine the fertility behavior of women according to rural/urban residence and how it has changed over 6.3. Fertility time. We, then, analyze how the fertility behavior of successive birth cohorts compares. Finally, we 115. Total fertility rate has fallen substantially in compare the labor force participation rates according Turkey in the last 30 years. As can be seen in Figure to motherhood. B3 in Appendix B, it was recorded at a level of 2.2 children per woman in 2003, down from a level of 5.7 6.3.1. Age specific fertility rates over time children per woman in 1968 and 3 children per woman in 1988. 118. Figure 44 compares the fertility rates for various age groups in urban areas in 1993 and 2003 using the 116. There is a rigid sequencing of demographic DHS data for both years. Fertility rates until age 35 events in Turkey. In less than two years following are considerably lower in 2003 than in 1988. There is marriage, women give birth to their first child. also a counter-clockwise rotation in this profile in the Although over the years, the age at marriage has sense that fertility moves to later ages in urban areas. increased, the lapse of time between marriage and While the fertility rate at ages 20-24 is higher than that first-birth has remained constant at around 1.8 years.10 at age 25-29 in 1993, they are at about the same level According to 2003 DHS results, the average age at in 2003. Figure 44: Age Specific Fertility Rates (per 1000 Women) in Urban Areas 180 160 140 120 100 1993 80 2003 60 40 20 0 15-19 20-24 25-29 30-34 35-39 40-44 45-49 Source: 1993 and 2003 DHS. 10 Giving birth out-of-wedlock is a very rare event. Determinants of and Trends in Labor Force Participation of Women in Turkey 37 Figure 45: Age Specific Fertility Rates (per 1000 Women) in Rural Areas 250 200 150 1993 2003 100 50 0 15-19 20-24 25-29 30-34 35-39 40-44 45-49 Source: 1993 and 2003 DHS. 119. Figure 45 displays the age-specific fertility rates 6.3.3. Female labor force participation rates by in rural areas in 1993 and 2003. Even though the levels number of births and age over time of fertility are higher in rural areas than those in urban areas, the patterns of change from 1993 to 2003 are 121. Figures 47 and 48 display how labor force similar. Fertility rates between the ages of 20 and 34 participation rates of women with children and without are much lower in 2003. Moreover, there is a counter- children compare in urban areas and rural areas, clockwise rotation in the profile. respectively. In urban areas, women with children are much less likely to participate until age 40. In rural areas, too, mothers are less likely to participate. 6.3.2. Age specific fertility rates by birth cohorts However, in this case, the evidence exists until age over time 35. Moreover, the gap is not as acute as that in urban areas.11 Therefore, we can conclude that there is a 120. In order to examine differences in fertility negative association between children and labor force behavior over successive birth cohorts, we also carry participation of women, in particular for those in urban out a decomposition analysis of cohort, age, and areas. calendar time effects using three cross-section waves of the DHS (1993, 1998, 2003). Figure 46 displays 122. The findings that there is a negative association the cohort effects according to this decomposition. between children and labor force participation of Women born after 1974 have a lower propensity to women and that fertility rates are falling in Turkey as have children compared to the earlier cohorts. well as that later cohorts of women in Turkey have Figure 46: Coefficient Estimates of Cohort Effects on Figure 47: Labor Force Participation Rate by Motherhood Fertility Rates in Urban Areas Status in Urban Areas 40 .6 30 .4 coheff coheff 20 20 .2 10 0 15-19 20-24 25-29 30-34 35-39 40-44 45-49 0 age age 84-88 79-83 74-78 69-73 64-68 59-63 54-58 49-53 44-48 no child child Source: 2003 DHS. 11 The sample size for women who are older than 35 and do not have children is very small in rural areas. Determinants of and Trends in Labor Force Participation of Women in Turkey 38 a lower propensity to have children imply that labor contribute to the downward trend. Given the high force participation of women in Turkey is likely to internal migration from agricultural rural areas to increase in the near future due to the changes that have urban areas in Turkey, an important question is how taken place in fertility behavior. migrants fare in the labor markets of urban areas. Could poor economic integration of migrant women Figure 48: Labor Force Participation Rate by Motherhood explain part of the low labor force participation rates Status in Rural Areas in urban areas? .5 124. Table 11 presents the share of employment in agriculture from 1988 to 2006, which exhibits a .4 significant decline. While 46.5 percent of employment .3 was in agriculture in 1988, this share went down all the way to 27.3 percent in 2006. .2 125. According to 2003 wave of the Demographic .1 15-19 20-24 25-29 30-34 35-39 40-44 45-49 and Health Survey, migrants – those who changed age age no child child residence in the last five years- constitute 15.1 percent of the working-age female population in urban areas. Source: 2003 DHS. The survey identifies migrants originating from cities, towns and villages. Migrant women from villages 6.4. Migration constitute 4.3 percent of the working-age female population in urban areas. 123. The decline in the agricultural sector may also affect the participation rates in urban areas 126. While 29.6 percent of migrant women in urban through migration. The move out of agriculture in areas are in the labor force, among non-migrants this rural areas and the resulting migration to urban areas rate is 27.0 percent. In other words, migrant women would change the urban participation rates if the are in fact slightly more likely to be in the labor force. participation propensities of rural migrants differed However, when we examine the migrant women who from those of urban residents. The difference in the originate from villages, we see that their labor force dependency ratios between migrant and non-migrant participation rate is lower at 23.6 percent. Therefore, households, increasing school enrollment among we can conclude that while migrants in general are not younger men and the change in the age-composition pulling the labor force participation rate in urban areas of the urban residents over time would also to lower levels, migrants from villages are. Table 11: Share of Agriculture in Employment (%) Source: HLFS web data base, TUIK (http://www.tuik.gov.tr). Determinants of and Trends in Labor Force Participation of Women in Turkey 39 127. The fact that migrants originating from villages age as well as regional variation in residence, there is are less likely to participate in the labor market is no evidence that migration from a village within the resulting from their personal characteristics. Table last five years is associated with a lower labor force 12 displays the effects of migrant status according participation probability. However, there is evidence to origin destination on the odds of labor force that migration from a city center is associated with participation. Once we account for the schooling lower odds of labor force participation in the new attainment, household wealth, marital status, destination. Table 12: Effect of Migrant Status According to Origin Destination Type on Labor Force Status in Urban Areas Note: Other controls include dummies for schooling attainment, household wealth, marital status, age, city size, and region. ** significant at 5%. 128. We also defined migrant status based on whether in determining women’s participation. We base our the current province of residence matches the province analyses on two data sets: 2006 HLFS and 2003 DHS. of birth. Based on this definition, we find that while As noted earlier, we primarily rely on DHS to see the 25.8 percent of the non-migrants in urban areas are role of children in determining women’s participation. in the labor force, 29.4 percent of the migrants in An added advantage of DHS data is that it includes urban areas are. Furthermore, we checked this for wealth indicators, which are constructed on the basis low educational groups because migrants with low of household durables and facilities, which we use educational attainment could be more likely to stay out to asses the role of permanent household income on of the labor force. However, again migrants are more women’s participation. Both surveys lack demand likely to be in the labor force: while 14.9 percent of side factors. To proxy for these, we use the place of the non-migrants with no education in urban areas are residence information provided at NUTS2 level. in the labor force, 18.6 percent of the migrants with no education are in the labor force. Therefore, with 130. The logistic regression on women’s participation this definition of migration, there is no evidence that shown in Table 13 confirms our earlier findings: wo- migrants are pulling the labor force participation of men with more schooling are more likely to participate women to lower levels. in the labor market (with the exception of primary 7. Determinants of Female Labor and secondary school graduates who surprisingly Force Participation have lower participation rates than illiterate women); participation increases in age, reaches a peak around 129. The analyses in Section 6 indicate that school- 35-39 years and then declines; it is more likely for ing, age, marital status and the number of children are single women to participate in the labor market; the important determinants of women’s labor force par- number of children in the household (ages 0-14) ticipation. Besides supply side factors, demand side decreases the participation of women. Women living factors are also important in determining the participa- in rural areas are more likely to participate than their tion of women. As also shown in Section 6, employed women are clustered into a small number of occupa- urban counterparts. Regions are also found to impact tions so that economic growth would benefit women’s on the participation probability of women differently. participation to the extent that job opportunities ex- To get a better sense of the impact of both supply and pand in these occupations. In this section of the report, demand side factors we run separate regressions for we carry out multivariate analyses to ascertain the rural and urban areas, which are also given in Table role of demand and supply side factors noted above 13. Determinants of and Trends in Labor Force Participation of Women in Turkey 40 Table 13: Logit Regression Results for the Likelihood of Labor Market Participation (based on 2006 HLFS data) Determinants of and Trends in Labor Force Participation of Women in Turkey 41 Notes: Robust standard errors in brackets. Covers women aged 15-64 years. * significant at 10%; ** significant at 5%; *** significant at 1%. 131. The role of education in determining women’s entering the labor market. The effect is especially participation becomes stronger when we consider strong for university graduates. women in urban areas only. Compared to illiterate women, those with higher levels of schooling have 132. As depicted earlier, the age-participation profiles progressively higher participation rates that peak for are hump-shaped in both urban and rural areas, though those with higher education. Among rural women, age is a stronger correlate of participation in urban illiterates and primary school graduates are not any areas. However, as discussed earlier, these profiles more likely to participate than illiterate women. mask cohort and time effects. Surprisingly, those with secondary and general high school education have a lower likelihood of 133. Being married is negatively associated with entering the labor market as compared to illiterate participation in both urban and rural areas, with a women, which may stem from demand side factors: particularly large effect in urban areas. Separated and the unavailability of “socially appropriate� jobs for divorced women are also less likely to participate in them and the changing economic structure of rural rural areas but not in urban areas. In both places, it households with the proportion engaged in agriculture seems less likely for widowed women to enter the declining. The last point is taken up in Section 8. labor market. The number of children in the household However, vocational high school graduates and those are also negatively associated with the participation with university degrees have a higher likelihood of probability of women in urban but not in rural areas. Determinants of and Trends in Labor Force Participation of Women in Turkey 42 134. Finally, regions are found to be strongly to these estimates, Figure 49 displays the effect associated with the participation of urban and rural of number of children on labor force participation women. In seven regions out of 12, the effect of regions according to education. What we see from this on women’s participation is in the same direction figure is that the number of children matters more in urban and rural areas. However, while residence for more educated women. In fact, for women with in the regions of West, Central, North East and Mid no education, there is no evidence at all that number East Anatolia as opposed to Istanbul decreases the of children matters in the labor force participation likelihood of women’s participation in urban areas, it decision. This is perhaps because the majority of such increases it in rural areas. women are in rural areas where they work as unpaid 135. Table B2 in Appendix B presents the odds ratios family workers. The increasing negative association from a logit estimation of labor force participation between labor force participation and children as the equations according to four highest educational number of children increases is especially prominent attainment groups using 2003 DHS data: no education, for women with higher education who are more likely primary, secondary, and higher education. According to be in urban areas working as wage earners. Figure 49: Coefficient Estimates of Effect of Number of Children on Labor Force Status by Highest Educational Attainment (Baseline: No Children) 1.00 0.50 0.00 1 2 3 4+ -0.50 No education Primary -1.00 Secondary -1.50 Higher -2.00 -2.50 -3.00 Number of children Figure 50: Coefficient Estimates of Effect of Wealth Quintile on Labor Force Status by Highest Educational Attainment (Baseline: Lowest Quintile) 1.00 0.50 0.00 No education 1 2 3 4 Primary -0.50 Secondary Higher -1.00 -1.50 -2.00 Income Quintile 136. Also according to the estimates in Table B2, wealth of the family matters more for less educated Figure 50 displays the effect of household wealth on women, in particular those with no education or pri- the labor force participation decision of Turkish wom- mary education. For those with secondary education, en by highest educational attainment. In this case, the only those in the two highest quintiles are less likely Determinants of and Trends in Labor Force Participation of Women in Turkey 43 to participate than those in the lowest quintile. For that the fall in the labor force participation among the women with higher education, there is no evidence higher-educated urban women in Turkey is not likely at all that household wealth is negatively associated to be caused by their increasing non-labor income (like with labor force participation. This last finding implies spouse’s income). Figure 51: Coefficient Estimates of Effect of Age on Labor Force Status by Highest Educational Attainment (Baseline: 15-19 age group) 5.00 4.00 3.00 No education Primary 2.00 Secondary Higher 1.00 0.00 20-24 25-29 30-34 35-39 40-44 45-49 -1.00 age 137. Finally, Figure 51 illustrates age effects once 8. A Closer Look at High and Low we hold all other variables in the logit regression Skilled Women in Urban Areas and constant. As it was displayed before, for the two Rural Women lower educational groups, age does not matter much in labor force participation status. For secondary 139. The analysis carried out in previous sections school graduates, all other age groups are more likely on the labor force participation of women has shown to participate than the youngest, 15-19, age group; three interesting trends: the labor participation of whereas for the highest educational group, there is highly educated women in urban areas is rather more of a hump-shape where labor force participation high, while those of less educated women very low. is more likely between the ages of 25 and 44 compared Furthermore, since 2000, the participation rates of to earlier as well as later ages. both groups have been stagnant. On the other hand, the labor force participation of women in rural areas 138. Few other points in Table B2 are worth is substantially higher than the low skilled women in mentioning. For women with no education, labor urban areas but their participation rates have recorded force participation is much less likely in large cities a drastic fall since 2000. The purpose of this section is compared to towns. On the contrary, for women with to understand the possible factors behind these trends. secondary school education, labor force participation To ease the discussion in this section, we define low is more likely in urban areas than in towns. Regional skilled workers as those with less than high school differences matter more for less educated women: education. labor force participation of women residing in the eastern regions lag behind that of women residing in 8.1. Understanding the Declining Participation Istanbul more for less educated women. There is no Rates among the Highly Educated Urban Women obvious pattern with regard to how the association between marital status and labor force participation 140. As illustrated earlier, labor force participation of changes with educational attainment, except that high skilled women – those with high school education marital status does not seem to matter at all for the and above - showed a declining trend from 1988 to highly educated women. 2000. For university graduates, the labor force parti- Determinants of and Trends in Labor Force Participation of Women in Turkey 44 cipation rate stabilized after 2000 at around 70 to 2006 period. The monthly earnings of these women percent. For high school graduates, the decline in somewhat deteriorated from 2002 to 2004, and picked the labor force participation slowed down beyond up again only to reach slightly higher figures in 2006 2000, but did not quite stabilize. Potential reasons as compared to 2002. A similar pattern was observed for falling/non-increasing participation rates include for men as well. In term of hourly wages, the drop declining market wages making it less likely for over the 2002-2004 is rather sharp, but so is the them to enter the labor market, increasing reservation recovery after 2004. In 2006, the level of the hourly wages due for instance to higher household incomes wages of women was close to its 2002 level but and changes in the composition of the highly somewhat higher than that of their male counterparts. educated women in Turkey. The fact that hourly wages of highly skilled women (and for that matter for the entire group of highly 141. Figures 52 and 53 show the changes in the month- skilled workers including men) in 2000s did not ly and hourly wages of highly skilled women (those record an appreciable increase is a potential factor with high school education and more) over the 2002 that can explain their stagnant participation rates. Figure 52: Monthly Real Earnings of Highly Educated Women in Urban Areas 600 500 Monthly real earnings (YTL) 400 male 300 female 200 100 0 2002 2003 2004 2005 2006 Notes: Covers individuals age 15 plus. Includes wage earners only. Source: 2002-2006 HLFS. Figure 53: Hourly Real Earnings of Highly Educated Women in Urban Areas 3 2.9 Hourly real earnings (YTL) 2.8 male 2.7 female 2.6 2.5 2.4 2002 2003 2004 2005 2006 Notes: Covers individuals age 15 plus. Includes wage earners only. Source: 2002-2006 HLFS. Determinants of and Trends in Labor Force Participation of Women in Turkey 45 142. Another potential factor can be the rising household probabilities of highly educated women. Furthermore, incomes of highly educated women, making it less likely given that the monthly average earnings of highly for them to participate in the labor market. However, the skilled men followed similar patterns as that of highly regression results on the basis of different educational skilled women and a higher likelihood of highly skilled categories reported earlier had shown household wealth women to come from households with highly skilled to be an insignificant determinant of the participation men, this does not stand out as a likely explanation. Table 14: Comparison of Individual Characteristics of Women According to Educational Attainment in 1993 and 2003 (based on DHS data) Figure 54: LFPR of Women in Urban Areas by Age and Education in 2006 90 80 70 University 60 High school 50 LFPR Secondary 40 Primary 30 Less primary 20 10 0 15-19 20-24 25-29 30-34 35-39 40-44 45-49 50-54 55-59 60-64 Age Source: 2006 HLFS. Determinants of and Trends in Labor Force Participation of Women in Turkey 46 143. The last explanation can be the changing participation rate. A key difference in the characteristics composition of the highly educated women and of women with higher education in 2003 compared therefore, their changing propensities to enter the labor to 1993 is their age distribution. The highly educated market. Table 14 presents the distribution of certain women in 2003 tend to be younger. The shares in the characteristics of women aged 15 and above according 20-24 and 25-29 age brackets are much higher in 2003. to their educational attainment (the highest schooling However, the participation rates in these ages are close level attended) using DHS data. The education cate- to the peak values at mid 30s. (This is illustrated below gorizations used in DHS is slightly different from in Figure 54.) In addition, when we examine how the that of HLFS such that secondary education includes labor force participation rates by age changed over time, grades 6 through 11. (This corresponds to secondary which is displayed in Figure 55, we see that compared education and high school education in the HLFS.) to 1994, labor force participation rates in 2006 were The question we ask here is: Could the lower labor lower for women with higher education regardless of force participation rates of highly educated women age. Therefore, the change in the age composition of after 2000 compared to early 1990’s be resulting from university graduates is not likely to explain the declining the changing characteristics of these women due to the trend in their participation rates, either. substantial increase in educational attainment within this time frame? 145. Another important difference in the characteristics of women with higher education in 2003 is with regard 144. According to Table 14, women with higher to their household head status. They are more likely to education are more likely to be never married in 2003. be household heads in 2003. However, since we would Given the higher probability of never-married women expect household heads to be more likely to participate to participate in the labor market, the increasing share in the labor market, this does not help us explain the non- of never-married among the highly educated could increasing trend in the participation rates of educated not explain the decreasing trend in their labor force women. Figure 55: LFPR of Women in Urban Areas with Higher Education in 1994 and 2006 100 90 80 70 60 1994 LFPR 50 2006 40 30 20 10 0 20-24 25-29 30-34 35-39 40-44 45-49 50-54 55-59 Source: 1994 and 2006 HLFS. 146. In 2003, a higher share of women with higher with higher education in the mid 1990s. Obviously, education was residing in the Northern and Eastern women with higher education in 2003 could still be regions of the country, where labor force participation different from the women with higher education in rates are lower. However, labor force participation 1993 in other ways that we cannot observe, and these rates for women with higher education exhibit little unobserved characteristics could explain the fall in regional variation according to 2004 HLFS. LFPR. Given the significant expansion in the pool of college graduates in this time interval, it is very 147. In essence, the pool of women with higher edu- possible that certain other characteristics of college cation in 2003 are in fact different in certain ways from graduates changed over time. the corresponding pool in 1993 in the ways that they are more likely to be never married, younger, heads of 148. Finally, it is important to note that the LFPR their households, and reside in Northern and Eastern of women with higher education decreased primarily regions of the country. However, none of these chang- between 1994 and 2000. When we examine the increase es would help us explain the falling LFPR for women in the fraction of women with higher education in the Determinants of and Trends in Labor Force Participation of Women in Turkey 47 total population, we see that the increase was faster lower than those of their male counterparts. Lower in this period than it was before or after. While the wages (for the same skill level) may stem from women fraction of women with higher education grew by pushed into segments of the labor market where lower 0.22 percent points on average per year between 1988 wages prevail or their limited access to better paying and 1994, and by 0.26 percent points on average per jobs, or simply due to wage discrimination. year between 2000 and 2006, it grew by 0.40 percent points on average per year between 1994 and 2000. 150. The wage data available from HLFS 2002 on- 8.2. Understanding the Low and Stagnant wards show that low skilled women do indeed receive Participation Rates among the Low-Skilled Urban lower wages compared to their male counterparts. Women Measured on the basis of monthly earnings from main job, women’s earnings (wage workers only) were equal 149. The reasons behind the low participation of to 71 percent of men’s earnings in 2006. When corrected low skilled women are likely to be their relatively for hours of work, women’s earnings were equal to 80.6 high reservation wages due to the large household percent of men’s earnings. These figures were up from sector in Turkey and low market wages for this 65.6 percent and 71.4 percent, respectively, in 2002. group of women. Given that domestic work is largely looked upon as women’s work, the reservation Table 15 shows the gradual closing of the discrepancy wages of women are likely to be higher than that of in the earnings of low skilled men and women in urban men, making it less likely for them to join the labor areas. market. Time-use surveys show the importance of women’s household production for the livelihood of 151. Female workers are somewhat younger than their low income households (Dayıoğlu and Kasnakoğlu, male counterparts, which may partly explain their lower 2000). Joining the labor force becomes even less earnings. Over a quarter of low skilled female wage earners likely if low skilled women are offered wages living in urban areas are younger than 25 years (Table 16). Table 15: Earnings of Low Skilled Women Wage Earners Compared to Low Skilled Male Wage Earners in Urban Areas Year Share of monthly Share of hourly earnings LFPR LFPR earnings of men of men men women 2002 65.6 71.4 68.2 11.4 2003 67.2 74.7 67.1 10.9 2004 70.4 79.1 68.0 10.9 2005 70.1 82.2 68.8 11.6 2006 71.0 80.6 67.8 11.8 Notes: Covers individuals age 15 plus. Includes wage earners only. Source: 2002-2006 HLFS, TUIK. Table 16: Distribution of low skilled urban men and women by age categories Source: 2006 HLFS, TUIK. Determinants of and Trends in Labor Force Participation of Women in Turkey 48 152. Women’s earnings have not only improved vis-à- skilled women remained below the (net) minimum vis men but also in real terms. Figure 56 shows that in wage, and the gap grew over time. While in 2002, the comparison to 2002, monthly earnings of low skilled average earnings of women were 93 percent of the urban women were up by about 20 percent in 2006. minimum wage, this figure dropped to 85 percent in Hourly earnings of women have also shown similar 2006. That the average earnings of women are below improvements (Figure 57). Notwithstanding these the minimum wage indicates that an average low favorable developments, the average earnings of low skilled woman living in urban is not covered by the Figure 56: Minimum wages and monthly earnings of low skilled women in urban areas 300 250 Monthly real earnings (YTL) 200 Earnings 150 MinWage 100 50 0 2002 2003 2004 2005 2006 Notes: Covers individuals age 15 plus. Wage earners only. 2002=100. Source: 2002-2006 HLFS, TUIK and MLSS. Figure 57: Hourly earnings of low skilled women in urban areas 1.2 1 Hourly real earnings (YTL) 0.8 0.6 0.4 0.2 0 2002 2003 2004 2005 2006 Notes: Covers individuals age 15 plus. Wage earners only. 2002=100 Source: 2002-2006 HLFS, TUIK. Table 17: Proportion of low skilled workers earning less Table 18: Employment status of low skilled workers in than the minimum wage in urban areas urban areas - 2006 Notes: Covers individuals age 15 plus. Wage earners only. Note: Covers individuals age 15 plus. Source: 2002-2006 HLFS, TUIK. Source: 2006 HLFS, TUIK. Determinants of and Trends in Labor Force Participation of Women in Turkey 49 minimum wage law, though improvements in mini- low-skilled wage workers across firms of different mum wages favorably reflect on her wages. Indeed, sizes are quite similar (Table 19). About 60 percent of in 2006, 77 percent of low skilled women as opposed both male and female wage earners work in establish- to 47 percent of men had monthly earnings that were ments employing less than 25 workers. The proportion below the minimum wage (see Table 17). The low fe- of those in very large establishments (250+) is also male earnings – despite the recent improvements – and quite similar: 10.1 percent of men and 8.7 percent of their relatively higher reservation wages probably ex- women worked in large establishments in 2006. plain why women’s participation remained around 11 percent as opposed to almost 70 percent of men. 155. However, in terms of social security registration a sizeable gender gap emerges. Considering the whole 153. The wage information presented above is for group of low skilled workers, while 55.2 percent of wage earners who made up about 70 percent of all men were registered with a social security institution low skilled male and female workers in 2006 (Table in 2006, the corresponding figure for women was 30.9 18). The employment status of the remaining 30 per- percent. The larger proportion of unpaid family work- cent varied between men and women. While the over- ers among the latter no doubt contributes to this differ- whelming majority of men worked on their own ac- ence. However, a large gap is also observed among the count (either as employers or self-employed), over 16 self-employed. In the case of men, while 42.2 percent percent (over half this group) of women worked as un- of the self-employed were registered, the correspond- paid family workers. As the figures in Table 18 show, ing figure among women was a meager 12.2 percent. a distinctly smaller proportion of low skilled women Registration improves for both men and women wage are able to join the labor market working on their own- workers. In fact, the improvement occurs on a larger account. scale for women. While 58.7 percent of wage workers in 2006 were registered with a social security institu- 154. Judging on the basis of the size of their tion, the corresponding figure was 40.8 percent among workplace, the majority of low skilled wage workers women. Therefore, although a larger proportion of and those who work on their own-account are women wage workers work without coverage, the concentrated in small establishments. However, quite gender gap in coverage is smaller for them than those interestingly, the distributions of male and female who are not wage workers. Table 19: Establishment size of low skilled workers in urban areas- 2006 Note: Covers individuals age 15 plus. Source: 2006 HLFS, TUIK. 156. In terms of sector of economic activity, low group at large. For instance, while a fifth of low skilled skilled female wage workers are found in manufac- women work in the agricultural sector, this proportion turing (44.6 percent), community services (24.2 per- is limited to 7.3 percent of female low skilled wage cent), wholesale and retail trade (16.5 percent) and in earners. Notwithstanding these differences, it is still agriculture (7.3 percent) (Table 20). The distribution the case that manufacturing, community services and for male low skilled wage workers is somewhat dif- wholesale and retail trade employ the majority of low ferent with a sizeable proportion found in construction skilled female workers. (12.5 percent) and in communications (7.2 percent). The proportion in wholesale and retail trade is also 157. The average earnings prevailing in various sec- higher among men. However, smaller proportions are tors for low skilled workers are also given in Table 20. found in manufacturing, community services and in Highest wages are observed in mining and electricity agriculture. It is also interesting to note that the dis- gas and water sectors where very few men and almost tribution of low skilled wage earners across sectors no low skilled women are found. Lowest wages, on the of economic activity differs considerably from their other hand, are observed in agriculture where a larger Determinants of and Trends in Labor Force Participation of Women in Turkey 50 fraction of low skilled women work. Looking at the sectoral distributions of men and women as weights, three sectors where low skilled women are mostly the average hourly wages of low skilled men and wom- found - manufacturing, community services and whole- en are found to be 2.3 YTL for both groups. Hence, sale and retail trade – reveal that community services in terms of the sectoral distribution of women, low offer relatively higher wages than the rest. Using the skilled women do not seem to be at a disadvantage. Table 20: Distribution of Low Skilled Men and Women across Economic Sectors in Urban Areas in 2006 Notes: Covers individuals age 15 plus. Average monthly wages are based on wage earners only. Hourly wages are calculated by dividing monthly wages to weekly hours worked. Standard deviations are in parentheses. Source: 2006 HLFS, TUIK. Table 21: Distribution of Low Skilled Men and Women across Occupations Note: Covers individuals age 15 plus. Source: 2006 HLFS, TUIK. Determinants of and Trends in Labor Force Participation of Women in Turkey 51 158. The occupations held by male and female wage earner counterparts. Although the mean hours low skilled workers in urban areas show significant worked by self-employed men is higher than male variations. Male wage earners, for instance, are more wage earners at 58.2 hours per week, the difference is likely to work as craft and trades workers, and plant rather small. and machine operators and assemblers, while women 160. Figure 59 illustrates the distribution of hours are more likely to work in elementary occupations worked by low-skilled self-employed men and women (Table 21). It is also interesting to note that, unlike the in 2006. The distribution is drastically different from sectoral distribution, the occupational distribution of that of female wage earners: 56.4 percent of women low skilled female wage workers and the rest (own work less than 40 hours per week as opposed to only account workers and unpaid family workers) are quite 14.5 percent of female wage earners. In the case of similar. men, while it is the case that a larger proportion of self- employed men work less than 40 hours (13 percent 159. Although the statuary work week in Turkey is as opposed to 4.2 percent), the proportion working 45 hours, low skilled wage earners work well beyond for 60 hours or more is also higher (57.2 percent as these hours. The mean hours worked in the reference opposed to 46.3 percent). Low skilled women who week was 56.7 hours among men and 49.8 hours chose/are able to enter the labor market as own- among women in 2006. The distribution of hours account workers are probably a select group, who may illustrated in Figure 58 shows a very small proportion desire to work limited number of hours. Considered in of women working less than 40 hours. The mean hours this perspective, the limited availability of part-time worked among the self-employed women at 37.3 jobs may be limiting the participation of low skilled hours per week is substantially lower than that of their workers. Figure 58: Distribution of Low Skilled Wage Earners in Urban Areas by Hours Worked in 2006 60 50 % of employed 40 Men 30 Women 20 10 0 1-39 40-59 60+ Hrs Source: 2006 HLFS, TUIK. Note: Covers individuals age 15 plus. Figure 59: Distribution of Low Skilled Self-Employed Men and Women in Urban Areas by Hours Worked in 2006 70 60 50 % of employed 40 Men 30 Women 20 10 0 1-39 40-59 60+ Hrs Source: 2006 HLFS, TUIK. Note: Covers individuals age 15 plus. Determinants of and Trends in Labor Force Participation of Women in Turkey 52 161. The above analyses have shown that the low 8.3. A further look at the declining participation participation of low skilled women in the labor market rates of women in rural areas and the non-increasing trend is unlikely to stem from employment discrimination against women. 162. Potential reasons for the decline in participation The distribution of women across industries shows rates in rural areas include a geographical shift – that apart from construction and communication due to the regional differences in out-migration and sectors, women are able to enter sectors where fertility rates – in the distribution of rural population low skilled men are also found. Neither is there to areas where female labor force participation is low, an unexpected concentration of women in firms of migration of rural women with higher propensity certain size. However, the occupational distribution of labor force participation to urban areas, a fall in agricultural wages and earnings, which could arise as does show variations among low-skilled men and a result of falling prices for agricultural products for women, indicating that women take up work of instance, production shifts from more to less labor- different nature within these firms/industries. The intensive crops, and a fall in the fraction of households plausible explanations then for the low participation engaged in self-account agriculture in rural areas. of low skilled women are low market wages and high reservation wages. The conditions of work, as measured 163. First, we examine how the geographical distri- by hours of work, also indicate heavy work hours, bution of rural women changed over time, which is which possibly conflict with the domestic demand on displayed in Table 22 for the years 1990 and 2000. women’s time. It is not surprising then that the majority The last column of Table 22 displays the variation in of low skilled women workers are relatively young. LFPR in these geographical areas in 2004. (We took Table 22: Geographical Distribution of the Population of Rural Women and Participation Rate of Rural Women by Geographical Region (HLFS data) Share of Rural Population (%) LFPR in Rural Areas (%) 1990 2000 2004 İstanbul 1.8 3.9 18.5 Western Marmara 5.8 5.4 39.1 Agean 14.2 14.5 41.3 Eastern Marmara 7.4 7.9 32.0 Western Anatolia 5.7 6.2 26.3 Mediterranean 13.0 14.7 28.3 Central Anatolia 8.3 7.7 24.8 Western Black Sea 12.6 10.4 50.5 Eastern Black Sea 7.4 6.7 69.0 Northeastern Anatolia 6.1 5.1 50.0 Central Eastern Anatolia 7.6 7.2 22.8 Southeastern Anatolia 10.0 10.4 22.9 LFPR values in 2004 as the previous Household Labor population is the Mediterranean region, where this Force Surveys do not have this regional information.) share rose from 13.0 to 14.7 percent. However, the We observe a fall in the share of rural population in LFPR of rural women in this region is also relatively regions where the labor force participation rates of rural low at 28.3 percent. women are the highest like Western Black Sea, Eastern 164. Another potential explanation to the fall in Black Sea and Northeastern Anatolia. The share of participation rates in rural areas is that women who rural women in the Western Black Sea region fell from choose to migrate from rural areas to urban areas 12.6 percent to 10.4 percent, and the share in Eastern could be different from the ones who choose to stay. Black Sea Region from 7.4 percent to 6.7 percent, and For this purpose, using the 2003 DHS, we compare the share in Northeastern Anatolia from 6.1 percent to the certain characteristics of “stayers� in rural areas 5.1 percent. The LFPR in rural areas of these regions – who are living in rural areas now and were living are above fifty percent, it is actually 69 percent in the there five years ago as well – with the characteristics Eastern Black Sea region. On the other hand, the share of “movers�, who were living in rural areas five years of rural population in Istanbul increased from 1.8 to ago but live elsewhere now. The first striking difference 3.9 percent, where few of these rural women would be between stayers and movers is in their age distribution: living on family farms where jobs are easily available. Movers are much younger compared to stayers. Three In fact, the LFPR in this region is the lowest at 18.5 quarters of movers are under the age of 30, whereas percent. Another region with increasing share of rural less than 45 percent of the stayers are. Movers are Determinants of and Trends in Labor Force Participation of Women in Turkey 53 also different from stayers in the ways that they are terms of educational attainment will be stronger among wealthier, they are more likely to married (because a the younger women, and, as a result, the fall in their husband facilitates migration), and they have smaller LFPR is more prominent. household sizes. 168. Next, we investigate how the terms of trade 165. The key difference between stayers and movers between agricultural and manufacturing products have is in their educational attainment. Movers are more changed over time. The prices of agricultural products educated compared to stayers. Compared to 14.8 not only determine the wage rate in agriculture but also percent of movers, 27.3 percent of stayers have no the earnings of households in the agricultural sector. As education at all. Primary school completion rate of can be seen from Figure 60, the terms of trade between movers is roughly 12 percentage points higher for agriculture and manufacturing, in fact, improved from movers. Similarly, secondary school education as the 1994 to 1999. However, labor force participation rates highest educational level (incomplete or complete) is of females in rural areas decreased from 48.9 percent also higher for movers. Furthermore, secondary school to 47.4 percent in the same period. The terms of trade completion rate is roughly 3 percentage points higher between agriculture and manufacturing decreased for movers. from a level of roughly 1.2 in 2000 to a level of 1.0 in 2006. In this period of declining terms of trade for 166. Given the fact that movers include a higher fraction of primary and secondary school graduates agriculture, the labor force participation rate for rural at the expense of women with no diploma and that females fell from 40.2 to 33 percent. In other words, primary and secondary school graduates have much the fall in the labor force participation was higher in higher participation rates compared to those with no a period of declining terms of trade than in a period diploma could in part account for why labor force of improving terms of trade. Therefore, there is some participation rate in rural areas is falling. indication that the fall in the terms of trade in agriculture could have contributed to the fall in the participation 167. Moreover, our argument that out-migration of rate in rural areas for females. However, the fact that more educated women accounts for part of the decline the participation rate fell even in a period of improving in rural LFPR is consistent with our previous finding terms of trade between 1994 to 1999 points out that that the LFPR has fallen much more for younger there are some other factors contributing to the decline women than older women in rural areas. In Table in the participation rates in rural areas. The terms of 23, we show that migrants are much more likely trade in agriculture is just a contributing factor. (Note to be younger women. Therefore, the selection in that the participation rate in rural areas for females fell Table 23: Comparison of Characteristics of Emigrants from Rural Areas (Movers) and Residents of Rural Areas (Stayers) - based on 2003 DHS data Stayers Movers 15-19 0.187 0.202 20-24 0.148 0.341 25-29 0.100 0.195 30-34 0.112 0.087 35-39 0.109 0.069 40-44 0.098 0.039 45-49 0.086 0.027 50-54 0.085 0.028 55-59 0.075 0.011 First Wealth Quintile 0.337 0.257 Second Wealth Quintile 0.263 0.260 Third Wealth Quintile 0.187 0.248 Fourth Wealth Quintile 0.139 0.135 Fifth Wealth Quintile 0.075 0.100 No Education 0.273 0.148 Incomplete Primary 0.086 0.083 Complete Primary 0.475 0.593 Incomplete Secondary 0.104 0.085 Secondary 0.042 0.071 Higher Education 0.021 0.020 Never Married 0.304 0.155 Married 0.653 0.822 Widowed 0.033 0.015 Divorced 0.007 0.005 Not living together 0.004 0.003 Number household members 6.134 5.533 Number of children under 5 0.654 0.863 Determinants of and Trends in Labor Force Participation of Women in Turkey 54 Figure 60: Changes in internal terms of trade between agriculture and manufacturing 1.80 1.60 1.40 1.20 1.00 TOT 0.80 0.60 0.40 0.20 0.00 1994_10 1997_10 2000_10 2003_10 2006_10 1994_1 1995_7 1996_4 1997_1 1998_7 1999_4 2000_1 2001_7 2002_4 2003_1 2004_7 2005_4 2006_1 2007_7 2008_4 Note: Terms of trade shows the ratio of the value-added in agriculture to manufacturing. Source: Producer Prices Indices, TUIK. from 47.4 percent in 1999 to 40.2 percent in 2000. As decline in such a short period of time. Similarly, the discussed earlier, the change in the sampling frame in fraction of household heads in rural areas engaged in 2000 can be a factor in this fall. However, the falling agriculture, but not necessarily on their own-account – LFPR is observed both before and after 2000.) which would also include wage-earners – also display a similar remarkable fall. Therefore, the fall in the 169. We also analyze the area devoted to various importance of self-employed agriculture is not due to a crops to see whether there has been any shift toward transition to wage-work in agriculture but a transition products that are less labor-intensive. Table 24 to other sectors. (This would in part be due to the compares the shares of total agricultural area devoted change in the rural/urban classification of certain areas to major agricultural products in 1995 and 2004. The over time. For instance, some rural areas may become crop patterns have not shown drastic changes over part of metropolitan areas.) time. 171. As a robustness check, we also look at the frac- 170. Finally, the fall in LFPR could arise due to the declining share of agriculture in rural areas, in tion of households with at least one member working particular due to the decline in the number of family- in agriculture. We find similar results. The proportion run establishments that kept the participation rates of households that have at least one member working of rural women at levels higher than their urban on own-account in the agricultural sector – we identify counterparts. On the basis of a number of measures, these households as having an agricultural establish- Table 25 provides the change in the importance of ment - has declined from 18 percent in 2002 to 14.5 agriculture and family-run agricultural establishments percent in 2006. A similar exercise that considers only in rural areas. The first column lists the fraction of the rural areas shows a more drastic drop in agricul- household heads in rural areas engaged in agriculture tural establishments. While in 2002, 48.4 percent of on own-account. This share drops from 41.3 percent to rural households had an agricultural establishment, 30.5 percent from 2000 to 2006. This is a remarkable this figure drops to 36.8 percent by 2006. Table 24: Changes in the area contributed to the cultivation of various crops (%) Source: 2004 figures are from Çakmak and Eruygur (2008), Table 9. Table excludes fallow land. Determinants of and Trends in Labor Force Participation of Women in Turkey 55 Table 25: Declining Share of Agriculture in Rural Areas Source: 2002-2006 HLFS, TUIK. 9. Conclusion However, when we examine these periods separately, we still find that the rural labor force participation 172. In this study, we describe the main features rate decreased from 50.7 percent to 47.4 percent over of women’s labor force participation and determine the 1988 - 1999 period, and from 40.2 percent to 33 the associations of a number of demographic percent over the 2000 - 2006 period. The decline in and economic factors with women’s labor force rural labor force participation rate for females has participation. In addition, we search for answers for been more prominent since 2000. In addition, when a number of interesting findings about women’s labor we examine the change in the participation rates by force participation in Turkey including the declining age groups, we see that the decline has been larger for labor force participation rates in rural areas as well as younger rural women. for highly educated women in urban areas and the low and stagnant participation rates for urban low-educated 175. Unlike the declining labor force participation rate women. in rural areas, participation rates in urban areas have been more stable. It displayed an almost flat profile 173. Female labor force participation rate in Turkey between 1988 and 1999 at a level of approximately 17 is low by OECD standards. Moreover, it has decreased percent. In fact, since 2000, it has shown an upward from 34 percent in 1988 to 27 percent in 2006. An trend: the participation rate of urban women increased important reason for this fall in participation rate by 2.5 percentage points from a level of 17.4 percent is urbanization. Turkey has witnessed high levels in 2000 to 19.9 percent in 2006. When we examine migration from rural to urban areas since 1988. The the participation rates by age, we find that the increase share of urban population rose from 51.1 percent in in the participation rates for certain age groups in 1988 to 63.3 percent in 2006. Since the labor force urban areas have been remarkable. For instance, the participation rate in urban areas is much lower, at participation rate for 25- to 29-year-olds increased by approximately 20 percent in 2006, than that in rural 8.5 percentage points from 1988 to 2006 and by 5.1 areas, at approximately 33 percent in 2006, the percentage points from 2000 to 2006. increasing share of urban population pulls down the labor force participation rate of women. 176. We also find that it is not the increasing prevalence of part-time work that is behind the 174. However, it is not only the decreasing share of increasing participation of women in urban areas. To rural population that is pulling down the participation the contrary, part-time work has decreased over time. rate for women. Labor force participation rate in Close to 70 percent of women in Turkey work for 40 rural areas for women has been declining itself: it has hours or more per week. This figure approaches to 85 gone down from a level of 50.7 percent in 1988 to percent in urban areas. 33 percent in 2006. We need to be careful, though, in comparing the numbers before and after 2000 because 177. Agriculture carries a heavy weight in the lives the sampling frame of the Household Labor Survey of women workers in Turkey. Although declining, a used in constructing these numbers changed in 2000. sizeable proportion of women, 60 percent, still work Determinants of and Trends in Labor Force Participation of Women in Turkey 56 in agriculture. This implies that changes in agricultural work for longer hours (more than 50 hours per week) activities will exert a strong influence on the trends than older cohorts. In other words, younger cohorts in the labor force participation of women. Since the of women in urban areas are not only more likely to agricultural sector in Turkey is dominated by small- participate in the labor market but also more likely to scale family run establishments, the female labor force work for longer hours. This implies that the finding in rural areas predominantly work as unpaid family for the whole female urban population – that both workers. In fact, about 40 percent of all working participation rates and hours worked were on the rise – women are unpaid family workers as a result of the is driven by the different behavior of younger cohorts large share of agriculture in female employment. of women entering the labor force. However, with the decline in agriculture, and family- run establishments, the importance of wage work 181. Significant improvements have taken in place among working women has been on the rise in both in women’s schooling in recent decades in Turkey. rural and urban areas. In fact, wage work is the For instance, the share of illiterates fell from 33.9 to dominant form of employment in urban areas where 19.6 percent whereas the share of university graduates 80 percent of women are found. These imply that labor rose from 1.8 to 5.8 percent from 1988 to 2006. One force participation rate of urban women is rising along of the most salient features of female labor force with a rising incidence of wage work. participation in Turkey is that it so much depends on educational attainment: participation rates increase 178. A cross-section analysis of the age-participation substantially with education. While the participation profiles of female labor market participants in urban rates of women in urban areas without a primary school and rural areas in 2006 reveals a hump-shaped age- diploma is below 10 percent and those of women in profile for urban participants, the peak occurring urban areas without a high school diploma is below 15 between ages 20 and 29. A similar hump-shaped percent, those of women with university degrees is at profile is also found for rural labor force participants; 70 percent. however, its hump is much weaker– the range between the ages of 20 and 59 is relatively flat –. The problem 182. Another salient feature of women’s labor force with these cross-section profiles is that the age effects participation by educational attainment in urban areas is could also stand for cohort and/or calendar year effects. that participation rates for university graduates as well Therefore, using pseudo-panel techniques with a series as regular and vocational high school graduates have of cross-section data, we decompose age, cohort, and fallen over time. The fall for university graduates took calendar year effects in the participation rate profiles. place in the mid 1990s, whereas the fall in high school graduates has been more gradual. On the other hand, 179. As a result of this decomposition analysis, we participation rates of women in urban areas with lower find a counter-clockwise rotation in the age-profile of levels of educational attainment have been stagnant. labor force participation rates in urban areas. The peak In other words, conditional on schooling, women’s is now between the ages of 20 and 39, instead of 20 participation rates in urban areas are either stagnant and 29. The age-profile of labor force participation or falling, yet the overall participation in urban areas rates in rural areas completely changes as a result of is rising. This is to do with the shift in the composition the decomposition. We find a monotonic age effect: of the workforce towards more educated women who rural women become less likely to participate as they have higher participation rates. Had the participation age at all ages. In terms of cohort effects, in urban rates of highly educated women remained at levels areas we find that younger cohorts of women are much recorded in 90s, women’s participation today would likely to participate in the labor market. In rural areas, have been much higher. on the contrary and perhaps surprisingly, younger cohorts of women are less likely to participate in the 183. Marriage is universal, while divorce is an labor market. unlikely event in Turkey: nearly 98 percent of women marry by age 49 and less than 1 percent of women 180. We also carried out a similar decomposition divorce by that age. The timing of marriage is early analysis for hours worked in urban areas. As a result of with the average age at first marriage being 20.7 years. this analysis, we found that younger cohorts of women The implication of these demographic factors is that the in urban areas are also less likely to work for shorter majority of women spend a good part of their life being hours (less than 40 hours per week) and more likely to married. However, the labor force participation rate of Determinants of and Trends in Labor Force Participation of Women in Turkey 57 married women is lower than that of single women in 186. The multivariate analyses carried out to see how Turkey: while the participation rate of single women various factors are associated with female labor market is 34.3 percent, the corresponding rate for married participation confirm the important role of education. women is 23.1 percent. The gap between the two In particular, university education is strongly positively demographic groups widens further in urban areas, associated with labor force participation in both rural where the participation of single women increases to and urban areas. As expected, married women are 35 percent but that of married women drops to 15.5 found to have a lower likelihood of participation percent. On the bright side, married women’s labor in both areas. Children younger than 15 years were force participation rate in urban areas has increased also found to be negatively associated with the over time. In addition, the share of single women in participation probability of urban but not rural women. the population –whose participation rates are higher That the regions were found to be strongly associated – is rising. with participation shows that demand side factors are also important in determining women’s labor market 184. Fertility behavior of women is also very participation. important with regard to their labor force participation decision as children influence the opportunity cost 187. We also conducted multivariate logit analyses of market work. This is particularly important in separately for four educational attainment groups: Turkey as almost all married women have children. A no education, primary, secondary, and higher. One comparison of fertility rates in 1993 and 2003, using interesting finding from this analysis is that children DHS data, reveals falling fertility rates in both rural matters much more for highly educated women. In and urban areas. Moreover, the magnitude of this fall fact, for women with no education, there is no evidence is significant. A decomposition analysis of age, cohort, at all that children is associated with labor force and time effects shows that cohorts born after the participation. On the contrary, the negative association 1970s have a much lower propensity to have children. of household wealth with labor force participation When we examine labor force participation rates by strengthens as education level decreases. For women motherhood status, we find that women with children with higher education, there is no evidence at all for have lower participation rates, particularly in urban a negative association between household wealth and areas. The lower fertility rates of younger cohorts labor force participation. We also find that age effects of women and the negative correlation between are stronger for higher educational levels. children and labor force participation imply a higher participation rate for younger women in Turkey. 188. One key finding that emerged from our examination of female labor force participation in 185. Given the high internal migration rate in Turkey, Turkey is the declining labor force participation it is also important to understand how the labor force rates in rural areas. Therefore, we examined the participation behavior of migrant women compares potential underlying reasons to this fact. First, we to that of non-migrant women. We find that the find that the geographical shift in the shares of rural participation rate of migrant women in urban areas population could partly account for this fact. There – who changed places in the last five years – is, in exists substantial variation in rural participation fact, slightly higher at 29.6 percent than that of non- rates across geographical regions. The share of rural migrant women in urban areas, which is 27.0 percent. population residing in the Black Sea regions and On the other hand, migrant women who originate from Northeastern Anatolia – where participation rates villages have a lower participation rate at 23.6 percent are much higher – is falling. Second, the fact that than non-migrant women in urban areas. However, the women in rural areas with a higher propensity to lower participation rate of migrant women originating participate in the labor market are also more likely to from villages can be explained by their different migrate could explain the declining participation rates. personal characteristics. Once we account for these Comparing the characteristics of migrants leaving variables, migrant women originating from villages rural areas with those of the residents of rural areas, do not have a lower propensity to participate in the we find that migrants, on average, have higher levels labor market. This time, perhaps surprisingly, it turns of education and are younger than stayers in rural out that migrant women originating from cities have a areas. Given the increasing participation rates with lower propensity to participate in the labor market after education, the fact that more educated rural women we account for a number of personal characteristics. are leaving for other locations would partly account Determinants of and Trends in Labor Force Participation of Women in Turkey 58 for the falling participation rates. Moreover, that these looked at the change in the composition of university movers are much younger also help us explain why the graduates over time but could not identify a change decline in participation rates in rural areas is especially in observable characteristics that could have led to prominent at younger ages. their declining participation rates. However, this does not preclude changes occurring in unobserved 189. Another potential explanation to the falling characteristics that could decrease their participation participation rates in rural areas is declining agricul- rates. Indeed, we show that the growth in the share tural wages due to a worsening of agricultural of university graduates has been particularly fast over prices. When we examine the terms of trade between the 1994-1999 period, when the biggest decline in the agriculture and manufacturing, we find that the terms of labor force participation of university graduates took trade for agriculture in fact worsened after 2000. This place. also could partly account for the falling participation rates in rural areas after 2000. In addition, a shift to 192. The final interesting finding that requires further less labor-intensive agricultural products could also investigation is the low and stagnant participation explain the decline in participation rates. However, we rates of low skilled women – those with less than do not observe a significant change in the land allotted high school education. Over the 2000-2006 period, to various agricultural products over time. the participation rate of low skilled women varied between 10.9 and 11.8 percent. These are considerably 190. Another important finding with regard to agri- lower rates compared to that of low skilled men which, culture that could help us explain the falling participa- over the same time period, varied between 67.1 and tion rates is the decline in share of households in rural 68.8 percent. To explain this large gap between the areas engaged in agriculture, and in particular own-ac- participation rates of low-skilled men and women, count agriculture after 2000. A decline in own-account we compared the wages of the two groups and noted agriculture means a loss of an easily available source an improvement in women’s wages vis-à-vis men. of work for many women. The above-mentioned However, we also noted that despite the improvement, worsening of terms of trade in agriculture could partly women’s wages were still very low, with over 75 account for this decline as well. In addition, the redefi- percent of women receiving wages that were below nition of rural/urban locations over time and certain the minimum wage. These low market wages as well rural areas becoming part of major metropolitan areas as high reservation wages of women stemming from could also explain the declining share of agriculture. the big household sector in Turkey coupled with very long hours of work probably explain why they 191. Another key finding of the study is the decline in participate in the labor market in so few numbers. A the labor force participation of highly skilled women look at broad segregation indicators such as industrial over the 1988-1999 period and their stagnant partici- distribution and firm size has not shown drastically pations rates (at about 70 percent) after 2000. The ab- different patterns between men and women, though sence of wage data has precluded us from investigat- they are distributed differently across occupations, ing the role of wages in explaining the declining rates which probably explain some of the gender wage prior to 2000. The wage data available for the 2002- gap. 2006 period indicate a deterioration and a recovery in women’s wages over the studied period. This is a po- 193. To have a better understanding of why both low tential explanation for why the labor force participa- and high skilled women in urban areas have had stag- tion of highly skilled women did not record an increase nant labor force participation patterns in recent years, in this time period. Another potential explanation for it would have been revealing to look at the activities the declining/stagnant participation rates of highly of non-employed women as well. For this purpose, skilled women would be an increase in their reserva- we had originally planned to use the 2006 Time-use tion wages due to higher household incomes. How- Survey of TUIK but at the time of the writing of this ever, our multivariate analyses showed that for uni- report the data were not released yet. Another area that versity graduates household income is not negatively needs further investigation is the family-run agricul- associated with participation, unlike other educational tural establishments, where we see the biggest change attainment groups. As a last potential explanation, we in female labor supply. Determinants of and Trends in Labor Force Participation of Women in Turkey 59 References Ansal, H. (1997), “Ekonomik Yeniden Yapılanma Dayıoğlu, M. and Başlevent, C. (2006) “Female Sürecinde Kadın Emeği� (Women workers during Employment, Earnings Inequality and Household the economic re-structuring), 1997 TÜRK-İŞ Well-being: The Case of Urban Turkey�, Yıllığı, Ankara. Proceedings of the 10th Annual Conference of the Economic Research Forum. Attanasio, Orazio P. (1998), “Cohort Analysis of Saving Behavior by U.S. Households.� Journal of Dayıoğlu, M. and Kasnakoğlu, Z. (1997), “Kentsel Human Resources, Vol. 33 (3), pp. 575-609. Kesimde Kadın ve Erkeklerin İşgücüne Katılımları ve Kazanç Farklılıkları� (Labor Force Participation Başlevent, C. and Onaran. O (2003) “Are Married of Men and Women and Gender Earnings Gap in Women in Turkey More Likely to Become Added Urban Turkey) METU Studies in Development, or Discouraged Workers?� Labour, Vol.27 (3), pp. Vol. 24 (3), pp. 329-361. 439-458. Dayıoğlu, M. and Tunalı, I. (2003), “Falling Behind Başlevent, C. and Onaran. O (2002) “The Effect of While Catching Up: Changes in the Female-Male Export-Oriented Growth on Female Labor Market Wage Differential in Urban Turkey, 1988 to 1994�. Outcomes in Turkey.� World Development, Vol Mimeo. 32. pp. 1375-1393. Demographic Health Survey (DHS) (2003), Hacettepe Browning, Martin, Angus Deaton ve Margaret Irish University Institute of Population Studies. (1985), “A Profitable Approach to Labor Supply Ercan, H. and Tunalı, I. (1998), “Labor Market and Commodity Demands over the Life-Cycle�, Segmentation in Turkey�, in Main Characteristics Econometrica, Vol 53(3), pp. 503-43. and Trends in the Turkish Labor Market, T. Bulutay Çağatay, N. and Berik, G. (1990), “Transition to Export- (ed.), Ankara: SIS. Led Growth in Turkey: Is There a Feminization Greenwood J., Seshadri A. and Yörükoğlu M. (2005) of Employment?�, Review of Radical Political “Engines of Liberation�, Review of Economic Economics, Vol 22 (1), pp. 115-134. Studies, Vol. 72(1), pp. 109-133. Çağatay, N. and Özler, Ş. (1995) “Feminization of Gustafsson S. and Jacobsson R. (1985), “Trends in the Labor Force: The Effects of Long-Term Female Labor-Force Participation in Sweden�, Development and Structural Adjustment.� World Journal of Labor Economics 3(1), pp. S256-274. Development, Vol. 23, pp. 1883-1894. Juhn C. and Potter S. “Changes in Labor Force Çakmak, E. and Eruygur, O. (2008) “Food, Rural, Participation in the United States�, Journal of Agricultural and Fisheries Policies in Turkey,� Economic Perspectives, Vol. 20(3), pp. 27-46. CIHEAM Country Profile. Kasnakoğlu, Z. and Dayıoğlu, M. (1997), “Female Çınar, M. (1994) “Unskilled Urban Migrant Women Labor Force Participation and Earnings and Disguised Employment: Home-Working Differentials between Genders in Turkey� in Women in Istanbul, Turkey.� World Development, Economic Dimensions of Gender Inequality: A Vol. 22 (3), pp.369-380. Global Perspective, J. M. Rives and M. Yousefi (eds.), Westport, CT: Praeger. Dayıoğlu, M. (2000), “Labor Force Participation of Women in Turkey� in Gender and Identity Kasnakoglu, Z. and Dayioglu, M. (2002), “Measuring Construction: Women of Central Asia, the the Value of Home Production in Turkey� in New Caucasus and Turkey, F. Acar and A. Gunes-Ayata Developments in National Accounts, T. Bulutay (eds.), Leiden: Brill. (ed.), Ankara: SIS, pp. 73-97 Determinants of and Trends in Labor Force Participation of Women in Turkey 60 Killingsworth, M. and Heckman J. (1985), “Female Japan�, Journal of Labor Economics Vol. 3(1), pp. Labor Supply: A Survey� in Handbook of Labor S355-S374. Economics, O. Ashenfelter and R. Layard (eds.), North-Holland: Amsterdam. Süral, N. (2007) “Legal Framework for Gender Equality at Work in Turkey�, Middle Eastern Kögel T. (2004), “Did the association between fertility Studies, Vol. 43 (5), pp. 811-824. and female employment within OECD countries really change its sign?�, Journal of Population Tansel, A. (1994) “Wage-employment, Earnings and Economics, Vol. 17(1), pp. 45-65. Returns to Schooling for Men and Women in Turkey.� Economics of Education Review, Vol Lim, J.Y. (2000), “The Effects of the East Asian Crisis 13(4), pp.305-320. on the Employment of Women and Men: The Philippine case�, World Development, Vol. 28(7), Tansel, A. (1999) “Public-Private Employment pp. 1285-1306. Choice, Wage Differentials and Gender in Turkey.� Economic Growth Center Discussion Paper No. Michael R.T. (1985), “Consequences of the Rise in Female Labor-Force Participation Rates 797, Yale University. – Questions and Probes�, Journal of Labor Tunalı, I. (1997), “To Work or Not to Work: An Economics, 3(1), pp. S117-146. Examination of Female Labor Force Participation Özar S. (1994), “Some Observations on the Position of in Urban Turkey�, Proceedings of the 5th Annual Women in the Labor Market in the Development Conference of the Economic Research Forum. Process of Turkey�, Boğaziçi Journal, Vol. 8 (1/2), pp.21-43. Tunalı, I. and Başlevent, C. (2001), “Married Women’s Participation Choices and Productivity Özbay, F. (1994), “Women’s Labor in Rural and Urban Differentials: Evidence from Urban Turkey�, Koç Setting�, Boğaziçi Journal, Vol. 8, ( 1/2), pp.5-19. University working paper. Özler, S. (2000), “Export Orientation and Female TÜSİAD (2000), Kadın-Erkek Eşitliğine Doğru Employment�, World Development, Vol. 28 (7), Yürüyüş: Eğitim, Çalışma Yaşamı ve Siyaset pp. 1239-1248. (In the path to gender equality: Education, labor market and politics), Istanbul: Lebib Publishing Pencavel, J. (1985), “Labor Supply of Men: A Survey�, House. in Handbook of Labor Economics, O. Ashenfelter and R. Layard (eds.), North-Holland: Amsterdam. TÜSİAD (2008), Türkiye’de Hanehalkı: İşgücü, Psacharopoulos G. and Tzannatos Z. (1993) “Economic Gelir, Harcama ve Yoksulluk Açısıdan Analizi and Demographic Effects on Working Women in (Households in Turkey: Analysis from the Latin America�, Journal of Population Economics, Perspective of Labor Force, Income, Expenditure Vol. 6(4), pp. 293-315. and Poverty), Istanbul. Shimada H. and Higuchi Y. (1985), “An Analysis of World Bank (1993), Turkey: Women in Development, Trends in Female Labor-Force Participation in Washington D.C. APPENDIX A Determinants of and Trends in Labor Force Participation of Women in Turkey 63 Table A1: Male and Female LFPR by Urban/Rural Residence Source: HLFS data base, TUIK. Note: Covers individuals aged 15 and above. Table A2: LFPR of Women by Age in Urban Areas Source: HLFS data base, TUIK. Note: Covers individuals aged 15 and above. Determinants of and Trends in Labor Force Participation of Women in Turkey 64 Table A3: LFPR of Women by Age in Rural Areas Source: HLFS data base, TUIK. Note: Covers individuals aged 15 and above. Table A4: Employment Status of Men and Women Source: HLFS data base, TUIK. Note: Covers individuals aged 15 and above. Determinants of and Trends in Labor Force Participation of Women in Turkey 65 Table A5: Employment Status of Men and Women in Urban Areas Source: HLFS data base, TUIK. Note: Covers individuals aged 15 and above. Table A6: Employment Status of Men and Women in Rural Areas Source: HLFS data base, TUIK. Note: Covers individuals aged 15 and above. Determinants of and Trends in Labor Force Participation of Women in Turkey 66 Table A7: Regular Work as percent of Wage Work Source: HLFS data base, TUIK. Note: Covers individuals aged 15 and above. Table A8: Female LFPR by Age and Birth Cohort in Urban Areas Source: HLFS data base, TUIK. Table A9: Female LFPR by Age and Birth Cohort in Rural Areas Source: HLFS data base, TUIK. Determinants of and Trends in Labor Force Participation of Women in Turkey 67 Table A10: LFPR of Women by Schooling Level Source: HLFS data base, TUIK. Note: Covers individuals aged 15 and above. Table A11: LFPR of Urban Women by Schooling Level Source: HLFS data base, TUIK. Note: Covers individuals aged 15 and above. APPENDIX B Determinants of and Trends in Labor Force Participation of Women in Turkey 71 Figure B1: Age-Participation Profiles of Women in Turkey, US, EU and OECD 90 80 70 60 T urkey 50 US LFPR 40 EU-19 OECD 30 20 10 0 15-19 20-24 25-34 35-39 40-44 45-49 50-54 55-59 60-64 Age Source: OECD Stat Extracts. Figure B2: Age-Participation Profiles of Men in Turkey, US, EU and OECD 100 90 80 70 60 T urkey US LFPR 50 EU-19 40 OECD 30 20 10 0 15-19 20-24 25-34 35-39 40-44 45-49 50-54 55-59 60-64 Age Source: OECD Stat Extracts. Determinants of and Trends in Labor Force Participation of Women in Turkey 72 Figure B3: Total fertility rate (TFR), mean age at marriage and mean age at first-birth 6.0 25.0 5.0 20.0 Mean age at marriage and first-birth 4.0 TFR 15.0 3.0 10.0 2.0 TFR Mean age at marriage 5.0 1.0 Mean age at first-birth 0.0 0.0 1968 1973 1978 1983 1988 1993 1998 2003 Source: Institute of Population Studies, Hacettepe University Table B1: Total Fertility Rate and Number of Births per Woman by Education in 2003 Source: DHS 2003, Table 4.2 Institute of Population Studies, Hacettepe University. Determinants of and Trends in Labor Force Participation of Women in Turkey 73 Table B2: Odds Ratios from a Logistic Regression of Labor Force Participation Status by Highest Educational Attainment in Urban Areas (based on 2003 DHS data) No Education Primary Secondary Higher Marital Status (reference = single) Married 0.525 0.262*** 0.440*** 0.685 (0.211) (0.0491) (0.0874) (0.195) Widowed 0.978 0.574 2.217 0.760 (0.578) (0.213) (1.107) (0.767) Divorced 3.804** 1.514 1.997* 0.855 (2.567) (0.572) (0.814) (0.491) Not living together 1.561 1.344 4.098* (1.289) (0.655) (3.424) Children (reference = no child) One child 1.072 0.747 0.441*** 0.361*** (0.493) (0.145) (0.0888) (0.114) Two children 1.673 0.684** 0.473*** 0.383*** (0.614) (0.116) (0.0949) (0.135) Three children 1.745 0.671** 0.431*** 0.238** (0.627) (0.127) (0.118) (0.135) Four or more children 0.947 0.639** 0.272*** 0.0677* (0.305) (0.133) (0.122) (0.105) Wealth Second quintile 0.664* 0.632** 0.641 1.919 (0.158) (0.117) (0.215) (0.890) Third quintile 0.608* 0.559*** 0.739 1.385 (0.177) (0.0977) (0.236) (0.569) Fourth quintile 0.512** 0.545*** 0.569* 1.466 (0.162) (0.0987) (0.178) (0.542) Fifth quintile 0.219** 0.420*** 0.341*** 1.906* (0.132) (0.0831) (0.106) (0.663) Type of location of residence (reference = town) Large city 0.432** 1.372 1.729** 0.847 (0.185) (0.266) (0.376) (0.266) Small city 0.723 0.979 0.855 0.715 (0.195) (0.148) (0.139) (0.180) Region (reference = Istanbul) West Marmara 0.287* 0.789 1.599* 0.426** (0.184) (0.210) (0.454) (0.164) Aegean 0.861 0.785 0.883 0.564** (0.383) (0.170) (0.195) (0.158) East Marmara 0.302* 0.938 0.928 0.996 (0.189) (0.188) (0.203) (0.338) West Anatolia 0.586 0.403*** 0.644** 0.403*** (0.354) (0.0875) (0.128) (0.110) Mediterranean 0.449** 0.861 0.806 0.754 (0.183) (0.152) (0.157) (0.247) Central Anatolia 0.232** 0.553** 0.660 0.900 (0.151) (0.142) (0.184) (0.396) West Black Sea 0.0628*** 0.714 1.337 0.907 (0.0544) (0.179) (0.340) (0.345) East Black Sea 0.175*** 0.757 1.064 1.203 (0.109) (0.201) (0.275) (0.462) Northeast Anatolia 0.537 0.407*** 0.678 0.751 (0.263) (0.111) (0.189) (0.304) Central East Anatolia 0.143*** 0.362*** 0.653 0.444* (0.0737) (0.103) (0.181) (0.214) Southeast Anatolia 0.188*** 0.311*** 0.599* 0.481 (0.0867) (0.0754) (0.161) (0.221) Age (reference = 15-19) 20-24 0.538 0.909 5.398*** 4.629*** (0.221) (0.160) (0.889) (1.364) 25-29 0.665 1.107 7.221*** 28.40*** (0.285) (0.219) (1.497) (10.10) 30-34 0.605 1.512* 6.434*** 43.49*** (0.273) (0.331) (1.520) (20.80) 35-39 1.184 1.835*** 8.894*** 50.11*** (0.497) (0.411) (2.168) (24.76) 40-44 1.062 1.692** 6.483*** 53.30*** (0.473) (0.387) (1.826) (27.30) 45-49 1.120 1.082 7.397*** 11.67*** (0.507) (0.268) (2.324) (5.943) Constant 2.061 1.481 0.621 0.214*** (1.105) (0.440) (0.242) (0.111) Observations 1246 3939 3013 1193 Pseudo R2 0.1077 0.1027 0.1513 0.1693 Robust standard errors in parentheses; *** p<0.01, ** p<0.05, * p<0.1 Notes: Notes: Notes: