22

Click here to load reader

Post-Soviet Demographic Paradoxes: Ethnic Differences in Marriage and Fertility in Kazakhstan

Embed Size (px)

Citation preview

Page 1: Post-Soviet Demographic Paradoxes: Ethnic Differences in Marriage and Fertility in Kazakhstan

Sociological Forum, Vol. 14, No. 3, 1999

Post-Soviet Demographic Paradoxes: EthnicDifferences in Marriage and Fertility in Kazakhstan

Victor Agadjanian1

The dramatic and generalized socioeconomic and political crisis that followedthe collapse of the socialist system has produced unique demographic re-sponses in the former Soviet Union. This study addresses the differences innuptiality and early fertility before and after the onset of the crisis and betweenthe indigenous and European-origin population in Kazakhstan. Drawing ondata from the 1995 Kazakhstan Demographic and Health Survey, this studyfinds that, contrary to the logical expectation, European-origin women aresignificantly more likely to marry earlier than indigenous women, and thisdifference has become even more pronounced during the crisis. However,the crisis is also associated with a longer interval between first marriageand first birth among European-origin women. The analysis shows thatEuropean-origin women prolong this interval through increasing use ofcontraception and abortion. The study attempts to link these findings to thesociopolitical and ethnic climate in Kazakhstan and to changing meaningsof and attitudes toward marriage and childbearing.

INTRODUCTION: CONCEPTS AND BACKGROUND

Since Malthus, the postponement of marriage and childbearing hasbeen seen as a natural demographic response to economic hardships andsocial upheavals. Several historical studies of western societies have demon-strated a strong association between socioeconomic oscillations and mar-

1Department of Sociology, Arizona State University, Tempe, Arizona 85287-2101.

KEY WORDS: ethnicity; marriage; fertility; Kazakhstan; Central Asia; former Soviet Union.

425

0884-8971/99/0900-0425$16.00/0 © 1999 Plenum Publishing Corporation

Page 2: Post-Soviet Demographic Paradoxes: Ethnic Differences in Marriage and Fertility in Kazakhstan

riage and fertility rates at the aggregate level (e.g., Galloway, 1988;Richards, 1983; Weir, 1984; Wrigley and Schofield, 1981). In more recentepochs such an association—albeit with variable strength and consistency—has manifested itself in several parts of the world (e.g., Easterlin, 1968;Lesthaeghe, 1989; Palloni and Tienda, 1991).

The collapse of the Communist bloc and the rapid and painful changesin the countries that once constituted it offer potentially rich materialfor an examination of the effects of dramatic sociopolitical and economictransformations on marital and fertility behavior. At first glance, it appearsthat the Malthusian thesis holds true there as well. For instance, in EasternEurope—and most conspicuously, in East Germany—a significant drop inmarriage and fertility rates has been observed, although as Conrad et al.(1996) argued, this precipitous decline might in fact represent a first stepin these societies' adaptation to western patterns of nuptiality and fertility.In Russia, the largest and the best studied of the states that emerged fromthe rubble of the Soviet Union, a recent decline in fertility has also beenobvious, although, as in the case of East Germany, it was said to exemplifythe "westernization" of Russia's reproductive dynamics (Vishnevsky,1996:44).

However, relatively little is known about recent trends in reproductiveand marital behavior in other Soviet successor states, especially in CentralAsia. My study attempts to fill this gap by focusing on age at entry intomarriage and early stages of fertility in Kazakhstan, the second largest interritory and third largest in population among the former Soviet republics.Like most other regions of the former USSR, post-Soviet Kazakhstan hasexperienced a dramatic and generalized economic decline, rising unemploy-ment, and a virtual collapse of the social welfare system (United NationsDevelopment Program, 1995). Also, like many of the ex-Soviet republics,Kazakhstan is a multiethnic nation, which offers the opportunity to examinedifferences in the demographic response to the sociopolitical and economictransition among its constituent ethnic groups.

My general hypothesis is twofold. First, I propose that when confrontedwith a prolonged social crisis or at least with a clear prospect of such acrisis, people do not change all components of their demographic behaviorequally and uniformly. Instead, they adjust specific aspects of their behaviorto the degree that they deem necessary and in a manner determined by aparticular set of contextual constraints. Thus, depending on the context andthe perceptions of the difficulties, changes in nuptiality may not necessarilyresonate with those in reproductive behavior. And second, I argue thatbecause different ethnic groups perceive these cataclysms and their implica-tions differently, their demographic reactions should also differ. This secondhypothesis is theoretically grounded in the vast literature on the relationship

Agadjanian426

Page 3: Post-Soviet Demographic Paradoxes: Ethnic Differences in Marriage and Fertility in Kazakhstan

between minority-group status and demographic behavior, particularly fer-tility. The minority-group status hypothesis posits that socially disadvan-taged ethnic, religious, or other minorities adjust their reproductive behav-ior to maximize their security or social mobility. Although the cross-culturalevidence is inconclusive and often difficult to explain and generalize, thisliterature provides numerous examples of ethnic differences in fertilitybehavior and outcomes regardless of individual-level sociodemographiccharacteristics (e.g., Aneshensel et al., 1989; Bean and Marcum, 1978;Goldscheider and Uhlenberg, 1969; Goodkind, 1995; Halli, 1987).

More specifically, I intend to examine differences in the onset of mar-riage and childbearing between the indigenous, traditionally Muslim andpredominantly ethnic Kazakh population of Kazakhstan and its populationof European origin—mainly Russians, but also Ukrainians, Belorussians,Germans, and others. Because both groups—to whom I will also referas Kazakhs and Europeans for shortness—have faced similar economichardships since the Soviet system began to crumble in the late 1980s, atypical Malthusian view would suggest that both should try to postponeentry into marriage and childbearing. However, because for the European-origin groups the economic crisis has been compounded by increasingethnocultural insecurity, I anticipate that their demographic reaction willbe more pronounced than among the indigenous population. At the sametime, the adjustment of both segments of Kazakhstani society may differfrom the classical Malthusian model because of different meanings thateach group attaches to marriage, family, and childbearing in that sociocultu-ral context.

Much has been written on the growing discomfort of Russians andrepresentatives of other ethnic groups of European stock in Kazakhstan.Ethnic tensions, simmering for decades, intensified considerably after theoutset of perestroika in the mid-1980s. In December 1986, Kazakhstan wasshaken by the first major explosion of ethnically driven violence when Kazaknationalists clashed with the police in protest against the appointment ofan ethnic Russian to the top post in the republic's power hierarchy—thatof the first secretary of the Kazakhstani Communist Party (Olcott,1995:251-253). Although the bleak prognoses of an inevitable and general-ized ethnic bloodbath to follow the 1986 troubles never materialized, rapidlyrising Kazakh nationalism and the increasing Kazakhization of all spheresof public life, especially after Kazakhstan's independence in 1991, havegreatly contributed to the uneasiness of Europeans living there (Masanovand Savin, 1997). And the generalized economic crisis that has characterizedindependent Kazakhstan has been particularly devastating in large indus-tries, where Russian and other European-origin labor has traditionallyconcentrated (Subbotina, 1997; Tishkov, 1994). As a result, Russians and

427Post-Soviet Demographic Paradoxes

Page 4: Post-Soviet Demographic Paradoxes: Ethnic Differences in Marriage and Fertility in Kazakhstan

other Europeans have expressed greater dissatisfaction with both the stateof ethnic relations and their material conditions than have Kazakhs (Arenovand Kalmykov, 1998; Kazakhstan, 1998).

Migration trends have always been the most "natural" and visibledemographic barometer of ethnopolitical and socioeconomic climate in thecountry. Until the end of the 1970s, the net balance of Europeans' migrationto Kazakhstan remained positive. The first large in-migration wave arrivedin the territory of modern Kazakhstan in the late 19th century, activelyencouraged by Russia's tsarist regime seeking to strengthen its grip onCentral Asia. The second major influx of Europeans occurred in the 1950sand 1960s, this time vigorously promoted by the Soviet government's eager-ness to tap Kazakhstan's agricultural and industrial potential. As a resultof this century-long in-migration Europeans came to constitute a sizableshare of the Kazakhstani population, with a particularly large concentrationin the northern areas of Kazakhstan and in then the capital city Alma-Ata(presently Almaty). According to the last all-Union census conducted in1989, people of European extraction made up about one half of the repub-lic's population, whereas ethnic Kazakhs constituted nearly 40% (Goskom-stal, 1991a:102). However, by then the net migration rate of Russians andother European-origin groups had already been negative for at least adecade.

Out-migration of European ethnic groups reached avalanche propor-tions in the years following independence, as the economic decline deep-ened, unemployment rose, and the nationalist stance of the political leader-ship hardened. However, the majority of Europeans still remain in thecountry—because of loyalty to their new nation, continuing hopes for abetter future, or the valuable personal assets that are difficult to transferabroad. Many are also well aware of the difficulties involved in adaptingto life in Russia or other European former Soviet republics following reloca-tion (e.g., Vesti Kazakhstana, 1996).

Numerous studies have discussed the causes, dimensions, and trendsof Europeans' migration from Kazakhstan to Russia and other countries(Demakov, 1997; Esenova, 1996; Skopin, 1994; Subbotina, 1997). Althoughmy study focuses on other aspects of demographic adjustment, an indirectimpact of migration on nuptiality and fertility is implied. Europeans whohave left Kazakhstan in recent years may have been demographically andculturally different from those who have stayed. For a large part of Euro-pean-origin people who still live in Kazakhstan, relocation abroad—eitheras a planned move or a precipitous flight—-remains a potential option(Arenov and Kalmykov, 1998). This never-vanishing prospect may haveaffected many Europeans' marital and reproductive plans and expectations.

For most of this century, Europeans' fertility levels were consistently

Agadjanian428

Page 5: Post-Soviet Demographic Paradoxes: Ethnic Differences in Marriage and Fertility in Kazakhstan

below those of Kazakhs. However, it has been noted that Russians and otherEuropeans in Central Asia have traditionally had lower age at marriage andhigher fertility than their coethnics in the European part of the formerSoviet Union. These differences are commonly explained by the influenceof the more pronatalist indigenous sociocultural milieu in the region (Bond-arskaya and Ilyina, 1979:17; Goskomstat, 1991b; Tishkov, 1994:144). At thesame time, it has been acknowledged that the large European populationin a republic like Kazakhstan could also influence marital and fertilitybehavior of the indigenous group (Bondarskaya and Ilyina, 1979:17).

Recent years have seen a decline in fertility levels in Kazakhstan. TheKazakhstan Demographic and Health Survey conducted in 1995 estimatedthe total fertility rate (TFR)2 for Kazakhstan as a whole at 2.5 comparedto 2.9 children per woman calculated from the 1989 All-Union census.Although fertility declined in all ethnic groups, the gap between Europeansand Kazakhs has remained. Thus ethnic Kazakhs' TFR in 1989 and 1995was 3.6 and 3.1, respectively, whereas the corresponding figures for ethnicRussians, by far the most numerous group among Europeans, were 2.2 and1.7 (National Institute of Nutrition and Macro International, 1996:37).

Traditionally, the indigenous Central Asian peoples were characterizedby earlier female marriage and a lower incidence of spinsterhood thanRussians and other Europeans (Akiner, 1997; Fisher, 1980:122-125). How-ever, in the last several decades age at first marriage among indigenousCentral Asian women has gradually risen while the Russian and otherEuropean women's age at first marriage, on the contrary, has declined. Infact, by 1995 the ethnic Kazakhs' median age of entry into first marriagewas more than a year higher than among Russians—21.7 and 20.5 years,respectively—and the differences were even more pronounced amongyounger women (National Institute of Nutrition and Macro Interna-tional, 1996:81).

DATA AND METHODS

This study uses data from the 1995 Kazakhstan Demographic andHealth Survey (KDHS). The KDHS interviewed 3771 women aged 15-49and was the first nationally representative survey of the kind conducted inthe country. The KDHS individual file, used in my analysis, contains rich,high-quality information on women's childbearing, including type and timeof pregnancy outcomes, and some key facts about women's marital history

2The total fertility rate indicates the average number of children born to a woman if she wereto experience the current age-specific fertility rates during her reproductive span.

429Post-Soviet Demographic Paradoxes

Page 6: Post-Soviet Demographic Paradoxes: Ethnic Differences in Marriage and Fertility in Kazakhstan

and their sociodemographic background. Unfortunately, however, theKDHS offers little information on the sociocultural peculiarities of thecontext and on the changes in the status and type of women's employmentand in their material conditions. Although these deficiencies impose somelimits on my analysis the KDHS in general supplies sufficient informationto test the proposed hypotheses.

To make a valid comparison of age at marriage between the twoperiods of Kazakhstan's recent history, I restrict my analysis to womenwho had ever been in a marital union. Considering the high female marriagerate and relatively early marriage that has characterized most of the formerSoviet Union, this restriction will leave out mainly the youngest respon-dents: close to half of the never-married women in the DHS sample wereyounger than 18, the legal age of marriage in Kazakhstan, and over 90%of them could not have legally married before 1987. Of course, the maritalstrategy of this group may somewhat differ from the rest but, given thenature of the data, these differences are impossible to assess reliably. Also,because I am interested in the onset of childbearing within marriage, Iexclude the small number of women who had births prior to their firstmarital union. The KDHS collected information on age at first maritalunion but, concerned with marriage primarily as a proximate determinantof fertility, it did not differentiate between the beginning of informal cohabi-tation and the contraction of a formal marriage.3 Thus for a number ofcases it is impossible to ascertain when the formal marriage took place.However, because premarital cohabitation was and still is relatively rarein Kazakhstan, I assume that this problem does not affect my results.

For the purpose of my analysis I define two ethnic conglomerates—Europeans and Kazakhs. The first group includes ethnic Russians, Ukraini-ans, Germans, Belorussians, and other ethnic groups of European extrac-tion. The second group is composed mainly of ethnic Kazakhs as well assmall groups of other, closely related, indigenous Central Asian groupssuch as Uzbeks and Uigurs. All other, smaller ethnic groups, whose positionin Kazakhstani society and reaction to the social transition cannot be un-equivocally predicted within the proposed conceptual framework, are ex-cluded from the analysis.

Table I illustrates the sociocultural and demographic differences be-tween the two ethnic groups in the resulting sample of 2607 women. ThusKazakh women were noticeably younger than Europeans, which is largelydue to the difference in the timing of fertility transition between the twogroups, but may be also attributed, at least to some extent, to the age

3The distinction between "being married" and "living with a man" in the KDHS individualfile is made only for the current marital status.

Agadjanian430

Page 7: Post-Soviet Demographic Paradoxes: Ethnic Differences in Marriage and Fertility in Kazakhstan

Post-Soviet Demographic Paradoxes

Table I. Selected Characteristics of Ever-Married or Ever-Cohabited Women withno Live Births Prior to First Marriage/Cohabitation by Ethnicity, Kazakhstan

Demographic and Health Survey, 1995

Median ageEducation, %

Secondary onlySecondary specialHigher

Area of de facto residence, %RuralTown or small cityLarge city

Employment, %Blue-collar occupationWhite-collar occupationDoes not work outside the home

Car ownership, %Household owns a carHousehold does not own a car

Median age at first marriageMean number of pregnancies (SD)Mean number of children ever born (SD)Use of modern contraceptives, %

Used at least onceNever used

Abortion history, %Had at least one abortionHad no abortions

Mean number of abortions (SD)

Kazakhs(n = 1419)

33.6

34.343.822.0

61.429.5

9.2

35.941.023.1

22.877.221.1

3.9 (2.7)2.8 (2.0)

68.531.5

34.265.80.8 (1.5)

Europeans(n = 1188)

36.1

26.453.320.3

25.249.025.8

43.139.917.0

30.469.620.1

4.7 (3.9)1.8 (1.2)

80.919.1

71.029.0

2.4 (3.1)

structure of the European out-migration. Europeans had a higher propor-tion of those with complete secondary education but Kazakhs held a slightadvantage in the share of women with college education. Europeans wereconsiderably more urbanized than Kazakhs, whose absolute majority re-sided in rural areas. Both groups demonstrated a high level of participationin the labor force—a typical feature of Soviet and post-Soviet societies—butthe share of women not working outside the home was still higher amongKazakhs. Europeans were more likely to hold a blue-collar occupation,reflecting this group's traditionally strong presence in the industrial sector.The similar levels of white-collar occupations, on the other hand, hide thepredominance of Kazakhs in clerical and bureaucratic occupations and theoverrepresentation of Europeans among the nation's technical intelli-gentsia. Finally, if we assume that car ownership is an adequate indicatorof a household's material well-being, Europeans lived generally in bettermaterial conditions than Kazakhs.

431

Page 8: Post-Soviet Demographic Paradoxes: Ethnic Differences in Marriage and Fertility in Kazakhstan

The remaining characteristics in Table I provide a marital and repro-ductive profile of women in the study. Because the ethnic classificationemployed here does not depart dramatically from the conventional Russiansvs. Kazakhs taxonomy used in the KDHS final report, the ethnic differencesthat I find in my sample resemble those between ethnic Russians andKazakhs mentioned earlier. Thus, half of all Kazakhs were married by age21.1, whereas Europeans' median age was a year earlier, at 20.1. Althoughon average Europeans had more pregnancies than Kazakhs, their meannumber of children ever-born was considerably lower. This was clearly dueto a much more frequent recourse to induced abortion among Europeans.Finally, Europeans were significantly more likely to have ever used moderncontraception.

To analyze the trends in the timing of first marriage (or marital cohabi-tation) and the onset of childbearing, I subdivide the sample into twocategories—those who entered their first marriages before 1987 (1727women) and those married for the first time in 1987 or after (880 women).The cutoff year is chosen for two reasons. First, the December 1986 confron-tations in Alma-Ata are seen here as the starting point for the rapid escala-tion of ethnic tensions in the republic. And second, 1987 was a watershedyear in fertility trends in most of the former Soviet Union. After 1987,following half a decade of a slight fertility increase, spurred in part by thepronatalist legislation of the early 1980s, fertility resumed its post WorldWar II decline, dipping to unprecedentedly low levels by the mid-1990s(Borisov and Sinelnikov, 1996:68-72).

Admittedly, this classification of marriage periods is imperfect—as isany attempt to express a continuous development in discreet terms. It ispossible, for instance, that the perception of the social crisis and ethnictensions and the demographic reaction to them have changed somewhatafter Kazakhstan's independence compared to the late 1980s. Unfortu-nately, any further breakdown of this post-1986 period is made unfeasibleby the small number of available observations. Nonetheless, I believe thissimple dichotomy is sufficient to capture the changes in ethnically specifictrends in nuptiality and fertility as needed to test my hypotheses.

I start my analysis with an overview of ethnic- and period-specifictrends in age at first marriage and in the interval between first marriageand first birth. I then move on to testing ethnic and period differences ina multivariate framework. I employ ordinary least squares regression totest the influence of ethnicity, time of marriage, and of their interaction onage at first union. To examine the influence of these factors on the probabil-ity of having a first birth, I use a proportional hazards model. In this model,I include both women who had already had children and those who werestill childless at the moment of the survey.

Agadjanian432

Page 9: Post-Soviet Demographic Paradoxes: Ethnic Differences in Marriage and Fertility in Kazakhstan

Both multivariate models include a set of control variables—educational level, type of employment, place of residence, region of resi-dence, and the household's ownership of an automobile. Lamentably, theKDHS provides only current status for the last four variables. However,in a context of relatively limited socioeconomic and spatial mobility, such asthat of Kazakhstan, these current characteristics can still serve as reasonableapproximations of a woman's sociodemographic background.

The last part of my analysis focuses on how the ethnic-period differen-tials in the first birth interval are affected by contraception and abortion.

AGE AT FIRST MARITAL UNION

As we learned from Table I, on the whole, Europeans tended to entertheir first marriage at an earlier age than Kazakhs. If we now compare thetrends in each group over the two marriage periods, we discover that thetwo groups moved in the opposite directions: Kazakhs' median age at firstmarriage rose from 21.0 to 21.3, whereas Europeans' median age at firstmarriage declined from 20.2 to 19.7.

To examine the effect of ethnicity and period on age at marriage netof other intervening factors, I employ ordinary least square regression.Table II presents the parameter estimates for three models: one with eth-nicity and period as the only predictors (Model A), one in which sociodemo-graphic variables are added as controls (Model B), and one that includesthe interaction between period of marriage and ethnicity in addition to thesepredictors' main effects and the effects of the control variables (Model C).

In the baseline model (Model A), both predictors exert statisticallysignificant effects on age at first union. As could be expected in a situationof a dramatic social upheaval, marrying in 1987 or later increased the ageat first marriage. On the contrary, being a European had a significantnegative impact on the age at first marriage. These effects remain significantafter the addition of education, area, and region of residence, occupation,and car ownership as control variables (Model B). In fact, in this modelthe effect of ethnicity increases compared to the baseline model. In ModelC, an interaction term for ethnicity and the period of marriage is added.The corresponding coefficient is statistically significant, indicating that thenegative association between being a European and age at first marriagebecomes stronger after the onset of the crisis. At the same time, even afterthe addition of the interaction term, the independent effects of ethnicityand of the period of marriage remain significant, and the latter, which nowrefers to Kazakhs only, considerably increases in magnitude.

The effects of some of the other variables included in the model are

433Post-Soviet Demographic Paradoxes

Page 10: Post-Soviet Demographic Paradoxes: Ethnic Differences in Marriage and Fertility in Kazakhstan

also worth mentioning. Not surprisingly, college education, white-collaroccupation, and residence in a big city all tend to increase age at firstmarriage. Regional differences in age at first marriage do not appear strong,with only the southern region somewhat standing out. Finally, household carownership, admittedly an imperfect measure of the respondent's materialenvironment, shows no discernible effect. In all, however, even Model C,the most comprehensive of the three models, explains but a small fractionof the variation in age at first marriage (adjusted R2 = 0.12), indicating

Agadjanian434

Table II. Ordinary Least Squares Regression Analysis of Age at First Union Among Ever-Married or Ever-Cohabited Women with No Live Births Prior to First Marriage/Cohabitation,

Kazakhstan Demographic and Health Survey, 1995

Variable

Ethnic groupKazakhs (reference)Europeans

Year of first unionBefore 1987 (reference)In or after 1987

Interaction: Europeans • Marriedafter 1987

EducationHigher (reference)Secondary specialSecondary only

EmploymentDoes not work (reference)Blue-collar occupationWhite-collar occupation

Household's car ownershipHousehold does not own a

car (reference)Household owns a car

Area of residenceRural (reference)Town or small cityLarge city

RegionCentral (reference)South (including Almaty-City)WestNortheast

ConstantAdjusted R2

Number of cases

Model A

b

—-0.99b

—0.27"—

———

———

———

————

SE

—0.12

—0.13

———

——

———

————

21.000.03

2607

Model B

b

—-1.22b

—0.24"—

—-1.10b

-1.83b

—0.090.79b

-0.21

—0.080.95b

—-0.45"

0.270.08

SE

—0.14

—0.13—

—0.160.19

—0.160.17

0.14

—0.140.22

—0.190.190.20

21.820.11

2607

Model C

b

—-0.91b

—0.66b

-1.01*

—-1.13b

-1.82b

—0.110.80*

-0.22

—0.080.97*

—-0.47a

0.270.08

SE

—0.16

—0.170.26

—0.160.19

—0.160.17

0.14

—0.140.22

—0.190.190.20

21.660.12

2607aSignificant at p < 0.05.bSignificant at p < 0.01.

Page 11: Post-Soviet Demographic Paradoxes: Ethnic Differences in Marriage and Fertility in Kazakhstan

that other relevant factors need to be accounted for to improve the model'sfit. This task, however, lies beyond the scope of this study.

FIRST BIRTH INTERVAL

Among women who had at least one child by the time of the survey,before-1987 Europeans and Kazakhs had the same median interval betweenfirst marriage and first birth—12 months—whereas after 1987 the Europe-an's median interval is longer than that of Kazakhs—12 vs. 10.5 months.These comparisons, however, may be misleading because they ignore thelarger number of still childless women among the Europeans, especially inthe post-1987 subsample. In addition, the nulliparous women may differfrom those who already had their first births in reproductive attitudes andbehavior. Thus, among respondents who married in 1987 or later and werestill childless at the time of the survey, the median duration of marriagewas 21.5 months for European but only nine months for Kazakhs. To refinethe analysis—by including women with an open first birth interval and bycontrolling for other relevant factors—a multivariate test is needed.

Table III displays the results of three proportional hazards modelsthat predict the risk of having a first child. The three models are constructedin the same way as are the models in the multivariate test of age at firstmarriage. In the baseline model (Model A) with ethnicity and marriageperiod as the only predictors, both variables show a negative influence onthe risk of having a first child. The impact is nontrivial for ethnicity: Europe-ans are significantly less likely to have a first child at any moment in time.In contrast, the effect of period of marriage in the baseline model is trivialand statistically nonsignificant. The addition of control variables, this timeincluding age at first marriage (Model B), does not raise the significancelevel of the coefficient for marrying in the later period. In contrast, thestatistically significant negative effect of ethnicity persists virtually unal-tered. Finally in the most comprehensive model that includes the ethnicity-period interaction (Model C), the independent effect of marriage periodchanges the sign but still remains nonsignificant, whereas the negativeimpact of being a European, even though considerably weakened, remainsstatistically significant. Most notably, however, the coefficient for the inter-action term is both large and highly significant. The addition of the interac-tion term also significantly improves the overall model fit compared to themodel with no interaction, as can be seen from the change in the log-likelihood statistic.

These results support the hypothesis about ethnic differentials in repro-ductive response to the social crisis, as they indicate that Europeans who

435Post-Soviet Demographic Paradoxes

Page 12: Post-Soviet Demographic Paradoxes: Ethnic Differences in Marriage and Fertility in Kazakhstan

married in or after 1987 had an especially large increase in the intervalbefore first birth. At the same time, the significance of the main effect ofethnicity points to a certain continuity in observed trends. Interestingly,among other characteristics, only area of residence displays a statisticallysignificant effect: women living in towns, and especially in large cities, havea much lower risk of having a first child than those who reside in ruralareas. In contrast, education—ostensibly the most likely predictor—doesnot appear to exert any net impact on the hazard of having a first child.

Table III. Proportional Hazards Analysis of the Risk of Having a First Child Among Ever-Married or Ever-Cohabited Women with no Live Births Prior to First Marriage/Cohabitation,

Kazakhstan Demographic and Health Survey, 1995

Variable

Ethnic groupKazakhs (reference)Europeans

Year of first unionBefore 1987 (reference)In or after 1987

Interaction: Europeans • Marriedafter 1987

Age at first unionEducation

Higher (reference)Secondary specialSecondary only

EmploymentDoes not work (reference)Blue-collar occupationWhite-collar occupation

Household's car ownershipHousehold does not own a

car (reference)Household owns a car

Area of residenceRural (reference)Town or small cityLarge city

RegionCentral (reference)South (including Almaty-City)WestNortheast

-2 Log likelihoodNumber of cases

Model A

b

—-0.20b

—-0.06

———

———

———

————

SE

—0.04

—0.04—

———

———

———

————

336422607

Model B

b

—-0.19b

—-0.05

-0.01

—-0.05-0.10

—0.010.04

0.01

—-0.18b

-0.20b

—-0.10

0.080.13

SE

—0.05

—0.05—

0.01

—0.060.07

—0.060.06

0.05

—0.050.08

—0.070.060.07

335972607

Model C

b

—-0.12"

—0.07

-0.28b

-0.01

—-0.06-0.10

—0.020.04

0.01

—-0.17b

-0.19a

—-0.11

0.080.13

SE

—0.06

—0.060.09

0.01

—0.060.07

—0.060.06

0.05

—0.050.08

—0.070.060.07

335872607

aSignificant at p < 0.05.bSignificant at p < 0.01.

436 Agadjanian

Page 13: Post-Soviet Demographic Paradoxes: Ethnic Differences in Marriage and Fertility in Kazakhstan

CONTRACEPTION AND ABORTION

How do Europeans achieve this increasing interval between first mar-riage and first birth? The answer appears to lie in contraception and abor-tion. Table IV presents the changes in the proportions of women who usedcontraception for the first time in their lives before their first birth. Thecorresponding proportions of those who started using contraception afterhaving one child are provided for comparative purposes.4 (Recall that onlyever-married women who had not had a live birth prior to first marriageare included in these comparisons.)

We can see from Table IV that the share of women who began usingcontraception before the birth of their first child increased considerablybetween women who married before and after 1987—from 6.7 to 22.7%.Overall, Europeans were almost three times as likely to use contraceptionprior to their first birth as were Kazakhs. In both ethnic groups, the changesacross the marriage periods in contraceptive incidence before the onset ofchildbearing are impressive. However, the difference between Europeansand Kazakhs changed little: while among women married before 1987Europeans were 3.8 times more likely to have contracepted before thebirth of their first child, among women who married in the later periodthis ratio was 3.4. In comparison, the ethnic gap narrowed considerably forwomen who started contracepting after giving birth to their first children—

Table IV. Percent of Women who First Used Contraception Before and After FirstBirth Among Ever-Married or Ever-Cohabited Women with no Live Births Prior to

First Marriage/Cohabitation, Kazakhstan Demographic and Health Survey, 1995

All marriagesAll women

KazakhsEuropeans

Married before 1987All women

KazakhsEuropeans

Married in 1987 or afterAll women

KazakhsEuropeans

Contracepted first timebefore first birth

(n = 2607)

12.16.5

18.8

6.72.8

10.5

22.712.141.4

Contracepted first timeafter first birtha

(n = 2401)

38.225.753.5

37.320.853.9

40.634.552.2

aNulliparous women excluded.

437Post-Soviet Demographic Paradoxes

4The KDHS data do not allow one to determine whether the first use of family planninginvolved a modern or traditional contraceptive method.

Page 14: Post-Soviet Demographic Paradoxes: Ethnic Differences in Marriage and Fertility in Kazakhstan

the corresponding share of Kazakhs rose significantly from 20.8 to 34.5%,whereas that of Europeans declined slightly from 53.9 to 52.2% (nulliparouswomen are excluded from this test). This suggests that a large proportionof recently married Europeans attempt to postpone the initiation ofchildbearing within marriage through contraception, whereas the over-whelming majority of Kazakhs begin using contraception—and delayingbirths—only after their first or higher-order children are born.5

The role of perceived economic hardships and social discomfort inchanging contraceptive prevalence is difficult to ascertain because thesweeping political and socioeconomic transformation of this society hascoincided with rapidly increasing access to contraceptives. In contrast,induced abortion was traditionally a major method of birth control inthe Soviet Union (Henshaw, 1990; Remennick, 1991), and even today,despite rising contraceptive use, it remains a major form of fertilityregulation in the Soviet successor states (Interstate Statistical Committee,1995; Popov, 1996; Westoff et al., 1998). The KDHS found strongdifferences in recourse to abortion between ethnic Russians and Kazakhs.Thus 61% of surveyed Russians had at least one abortion, whereas only25% of Kazakhs did. The total abortion rates6 were 2.7 for Russiansbut only 1.1 for Kazakhs (National Institute of Nutrition and MacroInternational, 1996:69-70).

Ethnic differentials in abortion in Kazakhstan have been analyzed indetail elsewhere (Agadjanian and Qian, 1997), and here I only focus onthose aspects relevant to this discussion. Specifically, I look at the abortionratio—that is, the number of abortions per total number of pregnanciesthat end in either a live birth or an abortion. Despite a decline in the absolutenumber of abortions in Kazakhstan in the recent years, the abortion ratiodid not change much throughout the early 1990s (Darsky and Dworak,1993; Ministry of Health of Kazakhstan, 1995). Figure 1 depicts abortionratios for first pregnancy by period and ethnicity. For purposes of compari-son, Fig. 1 also displays the same information on second pregnancy forwomen whose first pregnancy ended in a live birth (note that differentscales are used in the two graphs). Because my focus is on early stages offertility and fertility regulation within marriage, I exclude the few womenwho had pregnancies that ended before marriage.

5It is impossible to determine from the KDHS data whether and to what extent the increasein contraceptive use is related to increasing premarital sexual activity and premarital contra-ception in recent years. Premarital sex was relatively unusual in the former Soviet Union,but Russians and other Europeans have traditionally showed greater tolerance to it than theindigenous groups of Central Asia.

6Similarly to the TFR, the total abortion rate indicates the number of abortions a womanwould have during her reproductive span if she were to experience the current age-specificabortion rates.

Agadjanian438

Page 15: Post-Soviet Demographic Paradoxes: Ethnic Differences in Marriage and Fertility in Kazakhstan

Post-Soviet Demographic Paradoxes

Fig. 1. Abortion ratios for first and second pregnancies among ever-married orever-cohabited women with no pregnancies prior to first marriage/cohabitation,by year of marriage and ethnicity. (Kazakhstan Demographic and Health Sur-vey, 1995).

The differences between Europeans and Kazakhs are large for bothperiods and both pregnancies. However, what is perhaps even more strikingis the dramatic increase in the Europeans' abortion ratio for first pregnancywhen comparing women who married before and after 1987. Whereas 6%of European women who were married by 1987 aborted their first-in-lifepregnancies within their first marriages, this proportion almost doubledamong women marrying in 1987 and after. In contrast, the abortion ratio

439

Page 16: Post-Soviet Demographic Paradoxes: Ethnic Differences in Marriage and Fertility in Kazakhstan

for first pregnancy among before-1987 Kazakhs was a relatively low 2%and declined slightly among those who married in the later period. Althoughthe ethnic differences are no less impressive in abortion ratios for secondpregnancy, these have remained remarkably stable over the two periods,hovering just above 60% among Europeans and around 18% amongKazakhs. Therefore, as the sociopolitical and economic crisis in Kazakhstanexacerbated, the biggest change occurred in the outcome of first pregnancyamong Europeans. This finding confirms my hypothesis that Europeanswere demographically more susceptible to this multifaceted crisis. It alsoindicates that Europeans reacted by changing their behavior in the earlystage of childbearing.

To summarize the analysis of ethnic differences in contraceptive useand in recourse to abortion, women of European origin had overall higherusage of both contraception and abortion as means of delaying first birthsthan did Kazakh women. However, in recent years the ethnic gap hastended to narrow with respect to contraception but to widen with respectto abortion.

DISCUSSION

The foregoing analysis lends considerable support to my hypothesesabout ethnic differences in the demographic response to the dramaticsocioeconomic and political cataclysms in Kazakhstan. Overall, the crisistended to raise the age at first marriage, as similar crises did in othercomparable settings. However, European women were markedly differentfrom Kazakh women in marrying significantly earlier in the time of thecrisis. Notably, this difference in Kazakhs' and Europeans' age at firstmarriage was built upon the crisis-independent ethnic differences in ageat first marriage.

At the same time, I found that the crisis had little, if any, discernibleinfluence on the general timing of the first birth within marriage, but didsignificantly increase the already larger interval between the entry intomarital union and the birth of first child among Europeans compared toKazakhs. To compensate for the extended exposure to the risk conceptionand to maintain the desired delay in the arrival of the first child, Europeanwomen increasingly resorted to contraception and, when necessary, toaborting their first pregnancies. Thus the seemingly irrational change inEuropeans' nuptiality is offset by the rational strategies employed to post-pone the onset of childbearing.

But how to reconcile these two seemingly incompatible strategies—arelatively early age of entry into marriage and delayed childbearing?

Agadjanian440

Page 17: Post-Soviet Demographic Paradoxes: Ethnic Differences in Marriage and Fertility in Kazakhstan

Some tentative sociological and demographic explanations can be pro-posed. First, the protracted socioeconomic crisis and simmering ethnictensions may have affected the meaning of marriage for Europeans andKazakhs differently. Marriage was traditionally a very popular institutionin the Soviet Union (Fisher, 1980:117) and probably remains so despitethe radical transformation of the post-Soviet societies. Although littlereliable sociological data from the former Soviet Union are available tounderstand attitudes toward marriage, there are indications that, incontrast to indigenous Central Asians, childbearing was not seen as theprimary goal of young spouses among Russians and other ethnic groupsof European extraction. Thus, according to a survey of newly marriedcouples conducted in Moscow in 1983, the desire to have children wasonly third in importance among reasons for marriage chosen by womenand men alike—after the desire to receive and offer love and care andthe need for mutual understanding, psychological support, and protection(Sisenko, 1986:134-137). Social pressure to marry, especially for women,was traditionally high and marriage was an important factor in cementingyoung people's social identity. Marriage was not perceived as an immedi-ate economic burden, as young couples often shared both residenceand resources with their parents. At the same time, marriage was animportant—if not the only—channel for obtaining housing through thestate-run distribution system. Finally, under conditions of restrainedpremarital sexuality, marriage also offered many young people a culturallyacceptable sexual outlet.

Although the situation in Kazakhstan has changed considerably withrespect to such aspects as the government's housing policy and premaritalsexuality, marriage continues to be an important element of survivalstrategy for all segments of the Kazakhstani population. But given theparticularly vulnerable socioeconomic position in which Europeans seethemselves in Kazakhstan, they may be especially reliant on the institutionof marriage in confronting the current hardships and uncertainties—bothsocioeconomic and ethnocultural. The perception of and attitudes towardmarriage that had been shaped in the socialist years may therefore havebeen further reinforced by the conditions of the postsocialist crisis.

It is also possible that the observed difference in age at first marriagebetween Europeans and Kazakhs is indirectly related to out-migrationof the former group from Kazakhstan in recent years. Migration dataof necessary detail are lacking, but some speculation can be proposed.Migration is selective in terms of family status—childless people arepotentially more mobile whereas, other things being equal, those whomarry and have children are more likely to stay behind. Therefore, theEuropeans' relatively early age at marriage may reflect, at least in part,

441Post-Soviet Demographic Paradoxes

Page 18: Post-Soviet Demographic Paradoxes: Ethnic Differences in Marriage and Fertility in Kazakhstan

the marital preferences of nonmigrants. However, the emigration ofothers may also affect the views on marriage and childbearing of thoseEuropeans who stay. In the uncertain sociopolitical and economic climateof today's Kazakhstan, leaving the country must be considered onepossible way out of trouble. Thus by marrying relatively early, Europeansform a potential migration unit. At the same time, by delaying a first birth,they try to leave their options open. Ostensibly irrational, Europeans' ageat first marriage begins to make sense if it is examined separately fromreproduction. In fact, given the specific conditions in which Kazakhstan'sEuropeans live today, marriage becomes an even more important mecha-nism of social support, protection, and mobilization, especially for women.In a sense, it can be argued that the same uneasiness that forcesEuropean women to postpone childbearing may encourage them to entermarriage earlier.

Putting off the first birth at such a large scale is a new phenomenonbrought about by the constraints of a rapidly and painfully changing society.But although postponing first births within marriage was not common inprecrisis Soviet society (Borisov and Sinelnikov, 1996:71-73; Olson andMatkovsky, 1994:24), it was not altogether unknown: the practice of de-laying first births, including through use of abortion, was reported in surveyscarried out in various parts of the Soviet Union, including one survey ofRussians and generally more russified Kazakhs of Alma-Ata conducted inthe early 1970s (Katkova, 1977:87-89).

It should be noted here that despite the widening practice of postponinga first birth, especially among European-origin women, a sizable share ofboth Europeans and Kazakhs give birth very soon after marrying. In fact,this share may be rising because of increasing premarital sexual activitynot fully compensated by the spread of contraception. Also, the focus onethnic differences in delaying the first birth should not obscure the factthat even after the birth of the first child Europeans are generally morelikely to avoid or terminate further pregnancies than are Kazakhs, and thedifferences in total fertility rates mentioned earlier are shaped during theentire reproductive span. At the same time, Kazakhs have not been immuneto the hardships of the profound socioeconomic transition as the declinein their fertility rates and, to some extent, an increase in contraceptiveprevalence indicate.

Due to the limited number of available observations, this analysis didnot consider the sociocultural diversity within each of the two ethnic groups.Agadjanian and Qian (1997) showed, for example, that Kazakh women'srecourse to induced abortion depends significantly on the degree of culturalrussification. Similar nuances may exist in both Kazakhs' and Europeans'marital and fertility behavior. The same type of data constraints has also

Agadjanian442

Page 19: Post-Soviet Demographic Paradoxes: Ethnic Differences in Marriage and Fertility in Kazakhstan

prevented me from using a more detailed time scale for my analysis. Itwould have been particularly interesting to dissect the after-1987 periodand examine trends in its shorter segments.

These reservations notwithstanding, this study provides useful insightsinto the complexity of the demographic response to the crisis of socialtransition in Kazakhstan. This analysis also critically revisits the minority-group status hypothesis. Along the lines traced by Goodkind (1995), myanalysis demonstrates that the reproductive adjustment of a disadvantagedminority group should not be sought only in completed fertility but alsoin other fertility-related events that occur during the life course, such asentry into marriage and timing of births. These forms of adjustment aredictated and constrained by the specificity of the sociocultural context. Thefundamental transition underway in Kazakhstan and in other parts of theformer Soviet Union is historically unique—and so are the demographicdynamics that it generates. More sociological and ethnographic studies aretherefore needed to understand fully the mechanisms and directions ofthese processes.

Of course, these dynamics in Kazakhstan will change as the socioeco-nomic and ethnopolitical situation evolves. At this time nothing seemsto foretell a dramatic worsening of the ethnic climate in the country(Charney, 1997), although Russians and other Europeans are more criticalabout the state of affairs in Kazakhstan and more pessimistic about thefuture than are Kazakhs (Arenov and Kalmykov, 1998; Kazakhstan,1998). Kazakhstan possesses vast mineral resources, including oil andnatural gas, that can fuel robust and rapid economic growth. If thesocioeconomic situation improves and ethnic peace is safeguarded, thenchanges in both Europeans' and Kazakhs' demographic behavior maybe expected. Under such conditions, these changes are more likely tobe in the same direction—rising age at marriage, rising age at the onsetof childbearing, and a stabilization of fertility levels. However, if ethnictensions intensify and the Europeans' perception of social and culturalinsecurity exacerbates, a continuing and even deepening divergence inethnic strategies of demographic adjustment should be anticipated.

ACKNOWLEDGMENTS

I would like to thank Ekaterina Makarova, Cecilia Menjivar, LarissaRemennick, Zhenchao Qian, Rose Weitz, and three anonymous reviewersfor their helpful suggestions and comments.

443Post-Soviet Demographic Paradoxes

Page 20: Post-Soviet Demographic Paradoxes: Ethnic Differences in Marriage and Fertility in Kazakhstan

444 Agadjanian

REFERENCES

Agadjanian, Victor and Zhenchao Qian1997 "Ethnocultural identity and induced

abortion in Kazakstan." Studies inFamily Planning 28:317-329.

Akiner, Shirin1997 "Between tradition and modernity:

The dilemma facing contemporaryCentral Asian women." In MaryBuckley (ed.), Post-Soviet Women:From the Baltic to Central Asia: 261-304. Cambridge, UK: Cambridge Uni-versity Press.

Aneshensel, Carol S., Eve P. Fielder, andRosina M. Becerra1989 "Fertility and fertility-related behav-

ior among Mexican-Americans andnon-Hispanic white female adoles-cents." Journal of Health and SocialBehavior 30:56-76.

Arenov, M. M. and S. K. Kalmykov1998 "Ethnosocial reality of Kazakhstan."

(in Russian) Sotsiologicheskie Issledo-vania 3:45-58.

Bean, Frank D. and John P. Marcum1978 "Differential fertility and the minority

group status hypothesis: An assess-ment and review." In Frank D. Beanand W. Parker Frisbie (eds.), The De-mography of Racial and EthnicGroups: 189-211. New York: Aca-demic Press.

Bondarskaya, G. A. and L. P. Ilyina1979 "Ethnic differentiation in female nup-

tiality in the U.S.S.R.: The demo-graphic aspect" (in Russian). In A. O.Volkov (ed.), Demographic Develop-ment of the Family: 7-38. Moscow:Statistika.

Borisov, V. A. and A. B. Sinelnikov1996 Nuptiality and Fertility in Russia: A

Demographic Analysis (in Russian).Moscow: Institute for the Research ofthe Family.

Charney, Craig1997 "Survey addresses ethnic relations in

Kazakhstan." Surviving Together 15:52-53.

Conrad, Christoph, Michael Lechner, andWelf Werner1996 "East German fertility after unifica-

tion: Crisis or adaptation." Populationand Development Review 22:331-358.

Darsky, Leonid and N. B. Dworak1993 Kazakhstan: Fertility Indicators and

Characteristics of the Potential Mar-ket for Contraception. Washington,DC: The Futures Group.

Demakov, Georgui1997 "Population migration and its social

consequences in Kazakhstan" (in Rus-sian). Tsentralnaya Azia 4:28-37.

Galloway, Patrick1988 "Basic patterns of annual variations in

fertility, nuptiality, mortality andprices in pre-industrial Europe." Pop-ulation Studies 42:275-302.

Goldscheider, Calvin and Peter H. Uhlenberg1969 "Minority-group status and fertility."

American Journal of Sociology 74:361-372.

Goodkind, Daniel M.1995 "The significance of demographic triv-

iality: Minority status and zodiacal fer-tility timing among Chinese Malay-sians." Population Studies 49:45-55.

Goskomstat [State Statistical Committee]1991a National Composition of the Popula-

tion of the U.S.S.R. According to theAil-Union Population Census of 1989[in Russian]. Moscow: Finansi i stat-istika.

1991b "Territorial and ethnic differentiationof fertility and mortality in theU.S.S.R." [in Russian]. Vestnik Stat-istiki 8:67-69.

Easterlin, Richard A.1968 Population, Labor Force, and Long

Swing in Economic Growth. NewYork: National Bureau of EconomicResearch.

Esenova, Saulesh1996 "The outflow of minorities from the

post-Soviet state: The case of Kazakh-stan." Nationalities Papers 24:691-708.

Fisher, Wesley A.1980 The Soviet Marriage Market: Mate-

Selection in the USSR. New York:Praeger.

Halli, Shivalingappa S.1987 How minority status affects fertility:

Asian groups in Canada. Contribu-tions in Ethnic Studies, No. 18. NewYork, NY, and London, England:Greenwood Press.

Page 21: Post-Soviet Demographic Paradoxes: Ethnic Differences in Marriage and Fertility in Kazakhstan

Post-Soviet Demographic Paradoxes 445

Henshaw, Stanley K.1990 "Induced abortion: A world review,

1990." Family Planning Perspectives22:76-89.

Interstate Statistical Committee1995 Demographic Yearbook, 1995. Mos-

cow: Interstate Statistical Committeeof the Commonwealth of Indepen-dent States.

Katkova, I.1977 "Peculiarities of demographic behav-

ior of the family in early years of mar-riage" [in Russian]. In D. I. Valentei(ed.), Young Family, Issue 18:84-95.Moscow: Statistika.

Kazakhstan1998 "Kazakhstan." The Current Digest of

the Post-Soviet Press 50:23.Lestahaeghe, Ron1989 "Social organization, economic crisis,

and the future of fertility control inAfrica." In Ron Lesthaeghe (ed.), Re-production and Social Organization inSub-Saharan Africa: 475-505. Berke-ley: University of California Press.

Masanov, Nurbulat and Igor Savin1997 Kazakhstan: A model for ethnological

monitoring (in Russian). Moscow: In-stitute of Ethnology and Anthropol-ogy of the Russian Academy of Sci-ences.

Ministry of Health of Kazakhstan1995 Health of the Population of the Re-

public of Kazakhstan and Activities ofHealth Care Institutions in 1994 (inRussian). Almaty, Kazakhstan: Minis-try of Health of the Republic ofKazakhstan.

National Institute of Nutrition and Macro In-ternational1996 Kazakstan Demographic and Health

Survey. Calverton, MD: National In-stitute of Nutrition [Kazakhstan] andMacro International.

Olcott, Martha Brill1995 The Kazakhs (2nd ed.). Stanford, CA:

Houver Institution Press.Olson, David H. and Mikhail S. Matskovsky1994 "Soviet and American families: A

comparative overview." In James W.Maddock, M. Janice Hogan, AnatolyiI. Antonov, and Mikhail S. Matskov-sky (eds.), Families Before and AfterPerestroika: Russian and U.S. Perspec-tives: 9-35. New York and London:The Guilford Press.

Palloni, Alberto and Marta Tienda1991 "Demographic responses to eco-

nomic recessions in Latin Americasince 1900." Sociological Inquiry 62:246-270.

Popov, Andrej A.1996 "Family planning and induced abor-

tion in post-Soviet Russia of the early1990s: Unmet needs in informationsupply." In Julie DaVanzo and Gwen-dolyn Farnsworth (eds.), Russia's De-mographic "Crisis": 84-112. SantaMonica, CA: RAND.

Remennick, Larissa I.1991 "Epidemiology and determinants of

induced abortion in the U.S.S.R." So-cial Science and Medicine 33:841-848.

Richards, Toni1983 "Weather, nutrition and the economy:

Short-run fluctuations in births, death,and marriages, France 1740-1909."Demography 20:197-212.

Sisenko, Viktor A.1986 Youth Enters Marriage (in Russian).

Moscow: Mysl'.Skopin, A.1994 "Migration processes in Kazakhstan:

Past, present, and future" (in Rus-sian). In Zhanna Zayonchkovskaya(ed.), Migration Processes After theBreakup of the USSR: 167-177. Mos-cow: Institute for Economic Forecast-ing of the Russian Academy of Sci-ences and RAND.

Subbotina, I. A.1997 "Russians in Kazakhstan: The Migra-

tion situation on the eve and after thebreakup of the USSR" (in Russian).In S. S. Savoskul (ed.), Russians in theNew Abroad: The Migration Situa-tion, Relocation, and Adaptation inRussia: 154-175. Moscow: Institute ofEthnology of the Russian Academyof Sciences.

Tishkov, Valeri A.1994 "Russians in Central Asia" (in Rus-

sian). In V. I. Kozlov and E. A. Sher-voud (eds.), Russians in the NearAbroad: 140-149. Moscow: Instituteof Ethnology and Anthropology of theRussian Academy of Sciences.

United Nations Development Program1995 Kazakstan: The Challenge of Transi-

tion. Human Development Report1995. Almaty: UNDP.

Page 22: Post-Soviet Demographic Paradoxes: Ethnic Differences in Marriage and Fertility in Kazakhstan

446 Agadjanian

Vesti Kazakhstana1996 Vesti Kazakhstana (In Russian),

Jan. 16.Vishnevsky, Anatoli G., ed.1996 The Population of Russia, 1995 (in Rus-

sian). Moscow: Center for Demogra-phy and Human Ecology of the Insti-tute for Economic Forecasting of theRussian Academy of Sciences.

Weir, David R.1984 "Life under pressure: France and En-

gland, 1670-1870." Journal of Eco-nomic History 44:27-47.

Westoff, Charles F., Almaz T. Sharmanov,Jeremiah M. Sullivan, and Trevor Croft1998 Replacement of Abortion by Contra-

ception in Three Central Asian Re-publics. Calverton, MD: The PolicyProject and Macro International,Inc.

Wrigley, Anthony E. and Roger S. Schofield1981 The Population History of England,

1541-1871. Cambridge, MA: HarvardUniversity Press.