Turkish and Kurdish Fertility in Turkey: New Statistical Evidence for Convergence

Word Count: 3,528
Number of Tables: 5
Number of Figures: 4
Running Head: Turkish and Kurdish Fertility in Turkey

Abstract

The purpose of this quantitative study was to present and discuss statistical evidence for the convergence of Turkish and Kurdish fertility rates in the Republic of Turkey. Linear regression and other statistical methods were utilized in order to illustrate convergence, which was theorized as (a) resulting from the delayed exposure of Turkey’s Kurds to the dynamics of modernization and (b) providing compelling evidence against the claim that Turks and Kurds belong to fundamentally different demographic regimes. One important form of divergence between these populations—bride’s age at first marriage—was acknowledged and related to a potential change in sexual economics among Turkish Kurds.

Keywords: Turkish fertility; Kurdish fertility; Turkish fertility transition

Turkish and Kurdish Fertility in Turkey: New Statistical Evidence for Convergence
Introduction
Although the overall fertility rate for Turkey has declined precipitously as part of the country’s demographic transition, perhaps the earliest in the Muslim Middle East (Angin & Shorter, 1998), there is a well-documented disparity between the fertility rates of ethnic Turkish and ethnic Kurdish women in Turkey (Işık & Pınarcıoğlu, 2006; Yüceşahin & Özgür, 2008). The following conclusion has been drawn from the data: “…strong demographic differences exist between Turkish and Kurdish-speaking populations…the convergence of the two groups does not appear to be a process under way. Turks and Kurds do indeed appear to be actors of different demographic regimes, at different stages of demographic and health transition processes” (Koç, Hancıoğlu, & Cavlin, 2008, p. 447). The topic of ethnic Turkish and Kurdish fertility in Turkey is not widely studied among scholars, but remains of immense political and social interest. There has been journalistic speculation that Kurds will come to constitute the dominant ethnic population within Turkey (Ghosh, 2012). Turkish Prime Minister Recep Tayyip Erdoğan has made fertility an important policy theme, repeatedly asking women in Turkey to have more children (Altıok, 2013). Erdoğan’s outreach to Turkish Kurds (Taheri, 2014) is likely to be motivated by the demographic strength of ethnic Kurds in the Southeast of Turkey. For several years, Turkey’s Kurds have demanded greater autonomy (Gunter, 2013), an autonomy that they are can plausibly obtain through greater numbers. Clearly, then, much depends on obtaining accurate insight into the fertility of Turkish Kurds. For some time, the scholarly (Işık & Pınarcıoğlu, 2006; Koç et al., 2008) and popular (Ghosh, 2012) narrative has been that, in Koç et al.’s phrase, there is no convergence underway between the fertility of Turks and Kurds in Turkey. Admittedly, Koç et al. utilized a concept of convergence that included not only fertility but also several other demographic variables, including demography. Nevertheless, on the basis of fertility alone, there are good reasons to examine the claim of non-convergence more carefully, especially given that (a) Koç et al. only employed data from two surveys (the 2003 and 2008 Turkey Demographic and Health Surveys carried out every five years by researchers affiliated with Hacettepe University in Turkey); and (b) Koç et al. extrapolated a great deal from the Turkey Demographic and Health Surveys, including the ethnicity of populations within Turkey, in a manner unsupported by the data. My analysis of data from the 1978 to 2008 Turkey Demographic and Health Surveys, triangulated with an analysis of Turkish provincial-level demographic data provided by the Turkish Statistical Institute (TurkStat, 2014), offers some statistical support for the claim that Turkish and Kurdish fertility is in fact converging. The significance of this finding is threefold. First, the idea that Turkish Kurds constitute a special population vis-à-vis fertility has to be discarded; Turkish Kurds are obeying the same regional and indeed global forces driving lower fertility. Second, various theories that have drawn on the observed difference between Turkish and Kurdish fertility to make claims about hard social distinctions between Turks and Kurds might have to be reconsidered. Third, the theme of Kurdish fertility as it functions in all spheres of Turkish politics ought to be understood as specter reflecting the fears and ambitions of ethno-politicians more than as the reflection of any demographic reality.
Findings
The Hacettepe University-sponsored Turkish Demographic and Health Surveys from 1978 to 2008 (Hacettepe, 1978, 1983, 1988, 1993, 1998, 2003, 2008) trace fertility in five regions of Turkey, the north, south, east, west, and center. In order to measure convergence, I began by gathering the fertility rates for each region and placing them into Stata for analysis. Before proceeding to a discussion of the model, it should be noted that Hacettepe University’s decision to divide Turkey into cardinal directions and a center is not likely to yield the most accurate insights into the comparative fertility of different ethnic populations, because there is no neat overlap between this division of the country and the distribution of Turkish and Kurdish populations. Research is greatly complicated by the fact that the Turkish Statistical Institute and other official Turkish government bodies do not recognize the demographic existence of Turks; thus, while the Turkish Statistical Institute provides populations for every province and region of Turkey, these figures are not categorized by ethnicity. In the two most recent Turkish Demographic and Health Surveys, fertility has been reported for additional regions, including the Southeast, the part of the country with the highest concentration of ethnic Kurds. While this approach is an improvement, there are innate limits in measuring Kurdish facility solely by tracking fertility in the southeast. To begin with, vast numbers of Kurds live outside the southeast; in particular, many ethnic Kurds live in Istanbul (Laciner & Bal, 2004), which has one of the lowest fertility levels measured in the recent Turkish Demographic and Health Surveys. To offer more accurate insight into the difference between ethnic Turkish and ethnic Kurdish populations, it would be necessary to perform analysis on a dataset that tracked ethnicity. In the absence of such data, all extant analyses on the topic of fertility and ethnicity, including my own, are severely limited. Another limitation is that the historical data gathered in the Turkish Demographic and Health Surveys do not track fertility by the officially recognized regions of Turkey (of which there are currently 12, not, as in the Turkish Demographic and Health Surveys, 5), making it impossible to place the Turkish Demographic and Health Surveys data into inferential models with other regional data. The assumption that researchers have to make is that the fertility levels measured for the east represent Kurdish fertility whereas fertility levels in the other 4 regions in the Turkish Demographic and Health Surveys are assumed to represent Turkish fertility. With these limitations noted, a simple line graph reveals two points of note. First, the graph indicates that the fertility rates for the western, central, southern, and northern regions of Turkey have very nearly converged on 2. Second, the graph indicates that, while fertility rates for eastern Turkey remain ahead of fertility rates from the other four regions, they, too, are converging. In fact, the year of convergence can be predicted through linear regression. I regressed year on fertility level for each of the five regions, then used the Beta coefficients and constants to make predictions about fertility at several upcoming five-year intervals.
[Insert Table 1 here]
[Insert Figure 1 here]
The five regression equations obtained from the data were as follows:
Fertility Rate in West = (Year)(-.0611429) + 124.3906
Note: p = .0284, R2 = .6506, F = 9.31

Fertility Rate in South = (Year)(-.0865) + 175.6031
Note: p = .0250, R2 = .6667, F = 10.00

Fertility Rate in South = (Year)(-.0978571) + 198.2079
Note: p = .0076, R2 = .7880, F = 18.59

Fertility Rate in North = (Year)(-.1113571) + 225.2748
Note: p = .0004, R2 = .9332, F = 69.83

Fertility Rate in East = (Year)(-.0972143) + 198.3638
Note: p < .0001, R2 = .9795, F = 238.76

Interestingly, the highest R2, lowest p, and highest F value were all obtained for the east. These statistics indicate that fertility in the east is in strongly linear decline. If fertility is unlikely to decline below a set point, which for Turkey appears to be somewhere under 2, linear regression should not be used to predict the future values of the western, northern, southern, and central areas of Turkey, which are already close to 2. Rather, it seems more plausible to use linear regression to predict when the fertility rate for the east would enter the vicinity of 2, which, according to the calculations above, would be in 2018. Of course, this prediction is a statistical fiction; it could well be that, for whatever reasons, the set point of eastern fertility is going to remain significantly higher than the set point of fertility in every other region of Turkey. The near-perfect pattern of linear decline suggests otherwise, however, at least until eastern fertility gets closer to 2. Another way to support the idea of convergence without appealing to predictions based on regression is to construct a simple correlation matrix.
[Insert Table 2 here] The strong and significant positive correlations between the east and each of the other regions suggest that the regions are subject to the same basic force acting in the same direction over time, another argument supporting eventual convergence of fertility rates among ethnic Kurds and ethnic Turks in Turkey. Koç et al. (2008) conceptualized convergence in a complex manner, appealing to not only fertility but also other indicators of demography. Analysis of data regional provided by the Turkish Statistical Institute demonstrates that, even over a relatively brief period of time (2007-2013), there is a significant convergence between ethnic Turkish and ethnic Kurdish demographic indicators other than fertility. For example, the crude marriage rates of Istanbul and Southeast Anatolia are converging:
[Insert Table 3 here]
The convergence trend in Table 3 becomes clear when placed in a line graph:
[Insert Figure 2 here] Although the Turkish Statistical Institute does not provide year-by-year fertility rate tracking by regions, some inferences can be drawn from crude marriage rates to fertility given that the incidence of childbirth out of wedlock is relatively low in Turkey, for both ethnic Turks and ethnic Kurds (Kırdar, Dayıoğlu, & Koç, 2012). If so, then the convergence in crude marriage rates can be used as indirect evidence for convergence in fertility as well. Another point of convergence is the number of births that have taken place in Southeast Anatolia versus Istanbul. Analysis of these data would support convergence more strongly if the number of births in each year could be divided by the provincial population each year, or, for that matter, if parents’ ethnicities were tracked, but, in the absence of data, it can once again be assumed that Southeast Anatolia represents Kurdish fertility while Istanbul reflects Turkish fertility.
[Insert Table 4 here]
The convergence in births is clear in Figure 3 below:
[Insert Figure 3 here] Again, these methods of calculating convergence are crude, in that drawing inferences from regional patterns to ethnic differences is fraught with methodological difficulties and data gaps of the kind discussed earlier in this study. However, Southeastern Anatolia has been predominantly Kurdish for generations (Taşpınar, 2005) and therefore serves as a legitimate point of reference Kurdish demographics in Turkey. Moreover, even when regions other than Istanbul are added to the analysis, the underlying dynamics over 2007 to 2013 do not change. Turkish Kurds’ crude marriage rates are in decline, converging with crude marriage rates elsewhere in the country. These empirical findings strongly suggest that Turks and Kurds either belong, or will shortly belong, to the same fertility regime, if not necessarily the same demographic regime.
Support for Some Kinds of Demographic Divergence My analyses of fertility, the total number of births, and crude marriage rates also suggest the convergence of Turkish and Kurdish patterns. However, there appear to be some measures that support the case for divergence. Consider the marked divergence that has taken place between the average age at first marriage for brides in Istanbul and Southeast Anatolia. In 2004, these two regions had very similar bridal ages at first marriage, but, over the next several years, the average age of brides at first marriage climbed in Istanbul and descended in Southeast Anatolia.
[Insert Table 5 here]
[Insert Figure 4 here] A paired t-test disclosed that the mean difference in age between the bride’s first marriage age in Istanbul and the bride’s first marriage age in Southeast Anatolia was 1.06 years (SD = .883), and there was a statistically significant difference between these two ages (p = .0043, t = 3.7943). This divergence raises some fascinating possibilities. If marriage is conceptualized as a transaction in which women exchange the provision of sex for the receipt of security, a form of transaction for which repeated evidence has been found in the developing world (Westerhaus, Finnegan, Zabulon, & Mukherjee, 2008), then the earlier entry of Southeast Anatolian women into the marriage market might represent a means of compensating for lower fertility by providing earlier sexual access. On the other hand, if lower fertility is not a female-driven decision, but a collaborative decision between men and women, then it is unlikely that women would need to provide this kind of compensation to men in the marriage market. There are more possibilities, of course, but the idea of a trade-off between age of marriage and reduced fertility is also an invitation to examine the hitherto understudied sexual / marital marketplace in Southeast Anatolia Doubtlessly, there are other demographic variables illustrating a divergence between ethnic Turks and ethnic Kurds are in Turkey. Without a comprehensive review of the dozens of demographic variables tracked by the Turkish Statistical Institute and Hacettepe University, no sweeping conclusion about general demographic similarities and dissimilarities between ethnic Turks and Kurds in Turkey can be reached.
Discussion of Literature The fertility of Turkish and Kurdish women fits the frame of historical sociology as discussed by Abrams: Historical sociology is thus not some special kind of sociology; rather, it is the essence of the discipline. All varieties stress the so-called ‘two-sidedness’ of the social world, presenting it as a world of which we are both the creators and the creatures, both makers and prisoners; a world which our actions construct and a world that powerfully constrains us. The distinctive quality of the social world for the sociologist is, accordingly, its facticity—the way in which society is experienced by individuals as a fact-like system, external, given, coercive, even when individuals are busy making and re-making it through their own imagination, communication, and action (Abrams, 1982, p. 2).
As a phenomenon, fertility represents an inextricable intertwining of external and imaginative forces, if indeed it is granted that some theoretical or methodological distinction can be made between these kinds of forces. On the one hand, demographic transitions impose themselves on the subjects of fertility, women, with the changing facts of a new world-system of lower childhood mortality, non-farm economies, improved educational and professional opportunities for women, and increased costs (whether economic or emotional) of child-rearing. At the same time, fertility represents a key component of many imagined communities, whether at the level of nation, tribe, or family; at this level, fertility is not so much an external force as a prism for individual and collective expression and experience. A full understanding of the historical sociology of fertility requires an equal appreciation of fertility as an externally imposed fact-system and an internally generated cultural construct. Indeed, using historical sociology as a theoretical point of departure, the significance of my findings about the relative fertility of Turks and Kurds in Turkey touched on two larger themes: (a) The relative importance of external (environmental) facts versus internal (cultural) constructions as determinants of fertility and other demographic behaviors; and (b) the need to reconsider theories of fertility in light of what is concluded about the explanatory power of the imposed external fact-system versus the internally determined cultural system. If the fertility levels of ethnic Turkish and ethnic Kurdish women in Turkey is converging, then, in terms of external facts, a plausible conclusion is that what Işık and Pınarcıoğlu (2006) called “modernization dynamics” (p. 417) have entered the demographically Kurdish regions of Turkey with greater force than in the past. The regression model employed by Işık and Pınarcıoğlu supported the inference that increased literacy and the increased participation of women in the workforce would result in reduced fertility for women throughout Turkey. As such, in terms of fertility, the distinction between ethnic Kurdish and Turkey women has been due to different levels of exposure to education and employment in Turkey’s modernizing economy. One of the open theoretical questions among scholars of fertility convergence is the relative contribution of direct cultural mimicry, or, in other words, following the fertility behavior of others (Garssen & Nicolaas, 2008) versus exposure to the modernization dynamics mentioned by Işık and Pınarcıoğlu (2006). Theoretically, at least, these forces appear to be distinct. The mimicry approach seems driven by collectively oriented considerations such as social cognition, acculturation, and peer pressure while the modernization dynamics approach seems driven by something akin to individual cost-benefit analysis (Klaus, 2010). There does not appear to be an empirical way to test the distinction between these two theoretical drivers of fertility convergence; the data are compatible with both approaches. For example, in Garssen and Nicolaas’s discussion of the fertility convergence between native Dutch women and immigrant Turkish and Moroccan women, the observed convergence can plausibly be explained either as cultural mimicry or as immigrant women’s self-interested exploitation of new educational and professional opportunities. For that matter, these two theoretical themes can be blended, as the decision to pursue work and education can also be understood as a form of cultural mimicry. In the Turkish context, cultural mimicry seems theoretically inadequate to explain both the divergence and the subsequent convergence between ethnic Turkish and ethnic Kurdish fertility rates. Turks and Kurds have cohabitated in various parts of the Near and Middle East for over a millennium (Peretz, 1994), and, as a numerically and politically weaker people, the Kurds of Anatolia have long been under Turkish regimes including those of the Seljuqs, Ottomans, and, most recently, the Turkish Republic (Natali, 2005). The historical resistance of Kurds to assimilation noted by Natali suggests that mimicry is not a worthwhile explanation of any form of demographic convergence between the two populations. Rather, it seems more plausible to argue that delayed Kurdish exposure to the fruits of Turkish modernization is responsible for the observed convergence in fertility, crude marriage rate, and total number of births per year.
Conclusion
That ethnic Kurds in Turkey are the inheritors of a rich and distinct cultural tradition is not in serious doubt, despite the risible claims of early Turkish nationalists that the Kurds were so-called mountain Turks (Kirişci & Winrow, 1997, p. 103). At the same time, it is perhaps the case that the demographic differences between Turks and Kurds in Turkey have been utilized to overstate the boundaries between these populations. For that matter, the cultural theory of demographic difference between Turks and Kurds is overdetermined. The more parsimonious explanation of the earlier period of fertility divergence and the nascent period of fertility convergence is that Turkish Kurds have only lately been exposed to the forces of modernization that are associated with lower fertility. To be sure, the particularities of the Kurdish response to modernization can benefit from cultural explanation, but, at the same time, the power of modernization to flatten culture has to be respected.

References

Abrams, P. (1982). Historical sociology. Ithaca: Cornell University Press.
Altıok, Ö. (2013). Reproducing the nation. Contexts, 12(2), 46-51.
Angin, Z., & Shorter, F. C. (1998). Negotiating reproduction and gender during the fertility decline in Turkey. Social Science & Medicine, 47(5), 555-564.
Garssen, J., & Nicolaas, H. (2008). Fertility of Turkish and Moroccan women in the Netherlands: Adjustment to native level within one generation. Demographic Research, 19(33), 1249-1280.
Ghosh, P. (2012). A Kurdish majority in Turkey within one generation? International Business Times, 1. http://www.ibtimes.com/kurdish-majority-turkey-within-one-generation-705466
Gunter, M. M. (2013). The Kurdish Spring. Third World Quarterly, 34(3), 441-457.
Hacettepe. (1978). Turkish Fertility Survey. Ankara.
Hacettepe. (1983). Turkish Fertility Survey. Ankara.
Hacettepe. (1988). Turkish Population and Health Survey. Ankara.
Hacettepe. (1993). Turkish Demographic and Health Survey. Ankara.
Hacettepe. (1998). Turkish Demographic and Health Survey. Ankara.
Hacettepe. (2003). Turkey Demographic and Health Survey. Ankara.
Hacettepe. (2008). Turkey Demographic and Health Survey. Ankara.
Işık, O., & Pınarcıoğlu, M. M. (2006). Geographies of a silent transition: A geographically weighted regression approach to regional fertility differences in Turkey. European Journal of Population/Revue europeenne de demographie, 22(4), 399-421.
Kırdar, M., Dayıoğlu, M., & Koç, İ. (2012). The effect of compulsory schooling laws on teenage marriage and births in Turkey. (38375).
Kirişci, K., & Winrow, G. (1997). The Kurdish question and Turkey: An example of trans-state ethnic conflict Portland: Frank Cass.
Klaus, D. (2010). Changing Value of Children and Fertility Transition in Turkey. Journal of Comparative Family Studies, 799-815.
Koç, I., Hancıoğlu, A., & Cavlin, A. (2008). Demographic differentials and demographic integration of Turkish and Kurdish populations in Turkey. Population Research and Policy Review, 27(4), 447-457.
Laciner, S., & Bal, I. (2004). The ideological and historical roots of the Kurdist movements in Turkey: Ethnicity, demography, and politics. Nationalism and Ethnic Politics, 10(3), 473-504.
Natali, D. (2005). The Kurds and the state: Evolving national identity in Iraq, Turkey, and Iran. Syracuse: Syracuse University Press.
Peretz, D. (1994). The Middle East today. Westport: Greenwood Publishing Group.
Taheri, A. (2014). The Turkish Roller-Coaster. American Foreign Policy Interests, 36(2), 138-147.
Taşpınar, Ö. (2005). Kurdish nationalism and political Islam in Turkey: Kemalist identity in transition. New York: Psychology Press.
TurkStat. (2014). Main Statistics. from http://www.turkstat.gov.tr/UstMenu.do?metod=temelist
Westerhaus, M., Finnegan, A., Zabulon, Y., & Mukherjee, J. (2008). Northern Uganda and paradigms of HIV prevention: The need for social analysis. Global public health, 3(1), 39-46.
Yüceşahin, M. M., & Özgür, E. M. (2008). Regional fertility differences in Turkey: Persistent high fertility in the southeast. Population, Space and Place, 14(2), 135-158.

Table 1
Observed and Predicted Fertility, Five Regions of Turkey
| |Western |Southern |Central |Northern |Eastern |
|1978 |2.89 |3.77 |4.26 |4.99 |6.31 |
|1983 |3.67 |4.87 |4.83 |4.93 |5.49 |
|1988 |3.53 |4.48 |4.14 |3.56 |5 |
|1993 |2 |2.4 |2.4 |3.2 |4.4 |
|1998 |2.03 |2.55 |2.56 |2.68 |4.19 |
|2003 |1.88 |2.3 |1.86 |1.94 |3.65 |
|2008 |1.73 |2.09 |2.2 |2.08 |3.27 |
|2013* | | | | |2.67* |
|2018* | | | | |2.19* |

Note: * Predicted
Table 2
Correlation Matrix
| |West |East |North |South |Central |
|West |1 |.7717** |.8090** |.9981*** |.9565*** |
|East |.7717** |1 |.9625*** |.7866** |.8796* |
|North |.8090** |.9625*** |1 |.8279** |.9239*** |
|South |.9981*** |.7866** |.8279** |1 |.9667*** |
|Central |.9565*** |.8796* |.9239*** |.9667*** |1 |

Note: *** p < .001, ** p < .05, * p < .10

Table 3
Comparison of Crude Marriage Rates, Istanbul versus Southeast Anatolia
|Year |Istanbul |Southeast Anatolia |
|2007 |8.02 |10.55 |
|2008 |8.54 |10.09 |
|2009 |7.81 |9.48 |
|2010 |7.97 |8.73 |
|2011 |7.88 |8.97 |
|2012 |8.04 |9.1 |
|2013 |7.94 |8.63 |

Table 4
Comparison of Births per Year, Istanbul versus Southeast Anatolia
|Year |Istanbul |Southeast Anatolia |
|2001 |188544 |216372 |
|2002 |181993 |205800 |
|2003 |187815 |198399 |
|2004 |197056 |201000 |
|2005 |205212 |206842 |
|2006 |207823 |209262 |
|2007 |221885 |208294 |
|2008 |225910 |209211 |
|2009 |209945 |208640 |
|2010 |213110 |210588 |
|2011 |211874 |210327 |
|2012 |224469 |217395 |

Table 5
Comparison of Age at First Marriage for Brides, Istanbul versus Southeast Anatolia
|Year |Istanbul |Southeast Anatolia |
|2004 |23.3 |23.8 |
|2005 |23.5 |23.5 |
|2006 |23.6 |23.2 |
|2007 |23.7 |23 |
|2008 |23.9 |22.8 |
|2009 |24.1 |22.7 |
|2010 |24.3 |22.7 |
|2011 |24.5 |22.6 |
|2012 |24.6 |22.7 |
|2013 |24.8 |22.7 |

[pic]
Figure 1. Line Graph, Fertility Rates for Five Regions of Turkey, 1978-2008. Note: Based on data from Turkish Demographic and Health Surveys.

[pic]
Figure 2. Line Graph, Crude Marriage Rates in Istanbul and Southeast Anatolia, 2007-2013.
Note: Based on data from Turkish Statistical Institute.

[pic]
Figure 3. Line Graph, Total Number of Births in Istanbul and Southeast Anatolia, 2001-2012.
Note: Based on data from Turkish Statistical Institute.

[pic]
Figure 4. Line Graph, Bride’s First Marriage age in Istanbul and Southeast Anatolia, 2004-2013.
Note: Based on data from Turkish Statistical Institute.