1

Non-Marital Births in East Germany after Unification

Dirk Konietzka and Michaela Kreyenfeld★

July 2001

Abstract

In comparison to other European countries, West Germany displays relatively low ratesof non-marital childbearing. Since the 1960, there has been a postponement of firstbirth, an increase in the age at first marriage and an increase in childlessness.Nevertheless, childbearing and marriage remained strongly coupled. In the former EastGermany, on the other hand, non-marital childbearing was relatively high compared toother European countries and particularly compared to West Germany. In 1989, theratio of non-marital births had reached 33 percent. Overwhelmingly, researchers blamedGDR policies for high non-marital birth rates. However, after the breakdown of theGDR regime, the high East German non-marital birth rates did not rebound to WestGerman levels but they sky-rocked, reaching 50 percent in 1999. Using data from theGerman micro-census of the year 1997, we investigate the hypothesis that high non-marital births reflect a high labor market orientation among East German women withchildren. Our empirical results reveal two different patterns in East and West Germany.While in the West a high labor market orientation is indeed related to lower marriagerisks, we find the reversed pattern in East Germany. East German women with a collegedegree and/ or women who have a relatively higher educational attainment than theirmale partners are more likely to get married when they have children.

Keywords: Cohabitation, Female Employment, Germany, Non-Marital Childbearing

★ Authors’ correspondence address: Michaela Kreyenfeld, Max Planck Institute for DemographicResearch, Doberaner Str. 114, 18057 Rostock, Germany. Telephone: +49-381-2081-136. Fax:+49-381-2081-436. E-mail: kreyenfeld@demogr.mpg.de; Dirk Konietzka, Universität Rostock,Institut für Soziologie, 18051 Rostock, E-mail: dirk.konietzka@wisofak.uni-rostock.de

2

Introduction

In comparison to other European countries, West Germany displays relatively low rates

of non-marital childbearing. Since the 1960, there has been an increase in the age at

first marriage, a postponement of first birth, and an increase in childlessness (COUNCIL

OF EUROPE 1999; DORBRITZ 2000: 257). Nevertheless, childbirth and marriage

remained strongly coupled, which induced researchers to speak of a child oriented

marriage in West Germany (NAVE-HERZ 1994: 9). In the former East Germany, on the

other hand, non-marital childbearing was relatively high compared to other European

countries and particularly compared to West Germany. Since the 1970s, the ratio of

non-marital births has steadily increased, reaching 33 percent in 1989.

Overwhelmingly, researchers attributed high non-marital birth rates to GDR policies

(CROMM 1998; TRAPPE 1995: 210). While the intended aim of the government was to

support single mothers, at the same time, these regulations encouraged women not to

get married. With the breakdown of communism and the take over of the West German

institutional framework, it seemed straightforward to believe that East German women

would adapt to the western demographic patterns, i.e. non-marital birth rates would

soon fall to West German levels (HÖHN/ DORBRITZ 1995: 171; WITTE/ WAGNER 1995:

395). Surprisingly, after unification the opposite happened and non-marital birth rates

sky-rocked, reaching 50 percent in 1999 (see Figure 1).

The steady increase in non-martial birth rates in East Germany after Unification poses

several unresolved questions. In this paper, our focus is on two aspects – family policies

and female labor force participation – which may possibly explain the unexpected

development of out-of-wedlock childbearing. With respect to family policies, the crucial

question is why non-marital births increased, although the incentive structure in

3

contemporary Germany is supposed to strongly support marital childbearing (HUININK

1998: 301)? Is this an irrational response of East German women or couples to the new

family polices? Or do East Germans use the new incentive structure strategically, while

West Germans have, for decades, failed to do so? Are East Germans simply turning

their backs on “traditional family forms”? What role do high female employment rates

play for understanding the marriage pattern in the East?

In order to answer these questions, it is indispensable to distinguish different “types” of

non-marital births. Taking into account the modernization of family forms and the

increase in new living arrangements like cohabiting unions (e.g. KAUFMANN ET AL.

1997; RALEY 2001; SELTZER 2000; SMOCK 2000; TYRELL 1985), it makes sense to

distinguish births to single mothers, births in marital and births in cohabiting unions.

Furthermore, from a longitudinal perspective, the cohabiting couple may get married

shortly after the birth of their first (or second) child or may permanently remain

unmarried. While researchers initially classified cohabiting unions as “trial marriages”

(e.g. BENNETT/ BLANC/ BLOOM 1988), it has by now been acknowledged that they –

partially- represent independent and permanent family forms. In this study, our primary

focus is on women who permanently live in cohabiting unions with children. The

remainder of this paper is structured as follows. In part 1, we give an overview of the

family policies that define(d) relevant conditions of non-marital childbearing in East

Germany before and after Unification. In this context, we sketch our main hypothesis

that non-marital parenthood can be related to a high labor market orientation among

East German women. In part 2, we empirically investigate how the woman’s and her

male partner’s employment status is related to marriage decisions in East and West. Part

3 contains the concluding discussion.

4

Figure 1: Non-marital Births in East and West Germany (in percent of all births)

0

0.1

0.2

0.3

0.4

0.5

0.6

1945 1950 1955 1960 1965 1970 1975 1980 1985 1990 1995 2000

East

West

Source: STATISTISCHES BUNDESAMT (2000; 2001a)

1 Family Policies and Non-Marital Births in East and West

Germany

1.1 Family Policies before and after Unification

In the GDR, family policies were out-spoken pro natalistic and contained various

regulations that supported an early marriage. Upon marriage, couples received a “home

furnishing loan” of 7,000 Marks (5,000 Marks until 1986) and received priority access

to an own flat. However, there were also several important regulations, which fostered

single parenthood. Children of single mothers received priority access to public

daycare. When a child was sick or when no slot in daycare could be provided with, she

was guaranteed paid leave (GYSI/ SPEIGNER 1983: 75). The most important policy

5

measure was, however, the “Babyjahr”, which was introduced in 1976. 1 After the birth

of a child, a single mother was allowed to take a year of paid leave. Married mothers,

on the other hand, were only allowed to take advantage of this regulation after the birth

of a second or higher order birth. Since married and non-marital couples were treated

alike once the second child was born, the birth of the second child was an important

occasion to finally get married (HUININK 1999: 127; HUININK/WAGNER 1995).

The “Babyjahr” was regarded as the main reason for the rapid increase in non-marital

births to roughly 30 percent in the 1980s (HÖHN 1992: 9). Changes in the housing

“market” were presumably another important factor supporting this development. While

in the 1970s, marriage was still a major tool to get access to an own flat in the strongly

regulated East German housing market, in the 1980s, the housing shortage was partially

relieved and access to own apartments was facilitated for couples without being

married. The increase in non-marital birth rates was largely considered as a very

unintended effect of the East German family policies (TRAPPE 1995: 210). In 1986, the

GDR government responded to it and since then also married women could take a year

of paid leave after the birth of the first child. The extension of the “Babyjahr” put a halt

to a further increase in non-marital birth rates, but, no sizeable decline can be observed

until the breakdown of the GDR.

In October 1990, the former two German states were united and the East German legal

and political system was basically replaced by the West German one.2 Similar to former

1 Since 1961, women were allowed to take one year of unpaid leave after childbirth. Since 1976,mothers with two and more children were entitled to a year of paid leave (which amounted to theusual sick pay that was granted after the 7th week of sickness). Since 1984, mothers with morethan two children were entitled to 18 months of paid leave. In 1986, paid leave was extended to allmothers. If no slot in daycare could be found, all mothers were entitled to extend their period ofunpaid leave up to the third birthday of the child (CROMM 1998; FRERICH/ FREY 1993).

2 At German Unification, the Unification Treaty (“Einingungsvertrag”) went into force, whichprescribed the take-over of West German rules and regulations in East Germany. However, itshould be noted that some East German regulations were only gradually abolished. Paid leave

6

East Germany, single mothers are subject to special treatments. They are still entitled to

take paid leave when a child is sick and single parenthood is one key characteristic to

receive a slot in a public day care institution (DORBRITZ 1997: 243).3 Furthermore, there

are several transfer payments (such as maternity leave, social welfare, and housing

benefits) which are means tested. Single parents who are not working have priority

access to these treatments, since they do not have a partner whose income is assessed.

Although one should bear these regulations in mind for the subsequent analysis, it is

however unlikely that they are able to explain the East-West differences in marital

patterns. The most important reason is that in contrast to former East Germany, most

regulations in present-day Germany make an explicit distinction by cohabiting unions

(“Nicht-eheliche Lebensgemeinschaft”), marital unions and singles (PEUCKERT 1999;

SCHNEIDER/MATTHIAS-BLECK 1999). Child-rearing benefits, social welfare and

housing benefits are means tested and the non-married partner’s income (as long as he is

cohabiting) is, in general, assessed as well. A similar matter applies to priority access to

children’s day care. While single motherhood entitles to priority access to a daycare

slot, children of couples in cohabiting and marital unions will in most cases be treated

alike. Since the large majority of non-married East German women with children are

living in cohabiting unions (see below), they will not be able to take advantage of these

regulations. Finally, social welfare and housing benefits are only of interest for couples

who expect very low labor market incomes and/ or expect that they will be unemployed

for an indefinite period of time. Despite relatively unfavorable labor market conditions,

only a minority of (male) East Germans are permanently out-of job (BRINKMANN/

when a child was sick was still valid until July 1991. Parental leave regulations and child benefitswere changed in January 1991 (Berghahn 1992: 78ff.; Frerich/ Frey 1996).

3 Since 1992 married women (and in principle also married men) are entitled to take 10 days of paidleave to care for the sick child (20 for parents with more than one child). Single parents can take25 day paid leave to care for the sick child (50 days for single parents with more than one child)(BMA 2000: 152f.).

7

WIEDEMANN 1995: 330; MAYER/ DIEWALD/ SOLGA 1999). In sum, the German tax and

transfer system sets some incentives for non-working single mothers to avoid moving

together with the partner. But for the large majority of couples, these incentives should

not discourage them from getting married.

The very opposite is true. There are several important transfer payments which strongly

encourage non-working women with employed partners to get married. This pertains to

the system of income splitting, which allows married couples to file their taxes jointly,

i.e. the man’s and the woman’s income are added together, divided by two and taxed as

individual incomes. Due to progressive taxation, this regulation result into strong tax

relieves for married couples given that the man and the woman earn very unequal wages

or one of the partners is not employed. Moreover, married housewives are

automatically insured in the health insurance of their partners and entitled to a

������������������������� 2000; SAINSBURY 1997).

In sum, the German institutional framework encourages couples to get married

particularly when one of the partners withdraws from full-time employment after

childbirth. Although these regulations basically apply in the same manner to the

Eastern and Western states of Germany, there is, however, a major aspect, which

renders these regulations less decisive for the East. East German women are

substantially more likely to be full-time employed or to seek for full-time employment.

In the following, we discuss this issue in greater detail.

1.2 Female Employment in East and West Germany

Particularly in the early 1990s, East German women were subject to very discouraging

labor market constraints, i.e. high female unemployment rates, low re-employment rates

and high risks of status downward mobility once unemployed (BECKMANN/

8

ENGELBRECH 1999: 206; 6; HOLST/ SCHUPP 1995). Against this background, it was

widely expected that East German women would soon, similar to their West German

counterparts, withdraw from the labor market after childbirth and follow the traditional

“male-breadwinner model” (DORBRITZ 1997: 243; HUININK 1999: 129).

Although macroeconomic conditions remained less advantageous throughout the 1990s,

there are important factors which kept women’s employment rates on a high level. First

of all, in the East, the labor market situation of males also remained relatively

unfavorable, which basically pulled women into the labor market due to financial

pressures. It has also been argued that East German women have, as a holdover from

former socialist times, a high labor market orientation and consider economic

independence and a life long full-time employment as a “matter of course” (ADLER

2001; BRAUN/ SCOTT/ ALWIN 1994). Perhaps the most important aspect in this context

is the relatively high provision with public day care, which allows East Germans more

easily to reconcile childrearing and employment. In 1998, the provision rate with public

day care for infants (age 0-3) is only 3 percent in the West, but 35 percent in the East.

There is complete coverage with full-time care for pre-school children (age 3-6½) in the

eastern states but only a coverage of 16 percent in the Western states (STATISTISCHES

BUNDESAMT 2001b).

Whatever the more apt reason for the East-West differences in women’s employment

rates may be, a high labor market orientation among East German women should have

important implications for the decision to get married. As discussed above, the German

tax and transfer system is particularly beneficial for married couples who follow a

gender-specific division of labor, i.e. one of the partners devotes most of her (or his)

time to childrearing tasks while the other one is continuously full-time employed. If,

however, both partners are continuously full-time employed, there are only few

economic benefits that accrue to marriage. Therefore, it seems straightforward to

9

suspect that the East-West differences in non-marital childbearing relate to East-West

differences in the employment behavior of women. Since this line of argumentation

strongly relies on assumptions regarding the employment pattern of East and West

German women with children, we embark on this issue in more detail below.

Table 1 displays the employment rates of women of the birth cohorts 1961 to 1980 (the

cohorts we will use for the subsequent analysis as well). The Table displays a, by now,

well known pattern: East German women are more likely to be full-time employed than

their West German counterparts. In addition to this, those who are currently not

employed are looking more actively for employment and they are more certain that they

want to return to the labor market.4 The large East-West differences in women’s labor

market participation can almost exclusively be explained by the employment behavior

of women with children. While only small East-West differences in labor force

participation exist among women of the cohort 1971-1980 of whom only few already

have children, huge differences can be observed for the older cohort. East German

women with children are almost four times as likely to be full-time employed, and those

who are not employed are more than three times as likely to seek for employment.

In Figure 2, we display the employment rates of mothers by the age of the youngest

child. The Figure even more clearly exposes the differences in employment patterns of

East and West German mothers. While in the West only 10 percent of the women are

full-time employed when the youngest child reaches primary school age,5 in the East,

almost half of all mothers are full-time employed by then.6

4 Other studies have also shown that East German women who are part-time employed would rather

like to work full-time and East German women in maternity leave are faster in returning to thelabor market after childbirth (ENGELBRECH 1997; HOLST/ SCHUPP 1999).

5 Full-time employement is defined as 35 working hours and more per week.6 The share of part-time employment is higher among West German women with children. Even

though the share of part-time employed East German women has increased since Unification, in1997, it is still substantially below West German levels. It is worth noting that the increase in part-time employment is often considered as a positive development fostering the integration of women

10

Table 1: Women’s Employment Rate in 1997 (in percent)

West EastCohort 1961-1970 Not employed 38 30 Part-time employed 21 18 Full-time employed 35 48 In education 5 4

Cohort 1971-1980 Not employed 16 16 Part-time employed 5 5 Full-time employed 28 21 In education 51 58

Women with Children*) In education 2 4 Full-time employed 10 37 Part-time employed 25 18 Not employed Seeking for employment 6 20 Not Seeking for employment/ „housewife“ 42 13 Missing information if seeking for employment 15 9

Note: (1) *) Cohort 1961-1980, youngest child age 0-10 (2) Full-time employed, i.e. 35 and morehours per week; Part-time, i.e. 1-35 hours working hours per week (3) Source: Mikrozensus1997 (own estimations)

���� ��� ����� ������ �� ������ ������� �������� ����� � !� "�� #����� ��$%��&�������however actively worked against a spread of part-time employment – arguing that it kept womenout of the labor market (Obertreis 1986: 305ff).

11

Figure 2: Female Employment Rate of the Cohorts 1961-1980; by Age of the

Youngest Child

Figure 2a: East Germany

77 7663

5246 43 38 34 35 29 30

9

12

1315 23

2021 18

19 21

812

2029 35

3136 41 44 49 46

0%

25%

50%

75%

100%

0 1 2 3 4 5 6 7 8 9 10

Figure 2b: West Germany

83 78 7367 63 61 59 57 55 52 50

11 1721 25 28 29 32 35

34 35

8 7 8 10 10 10 10 10 8 12 13

0%

25%

50%

75%

100%

0 1 2 3 4 5 6 7 8 9 10Not employed Part-time Full-time In education

Note: (1)Full-time employed, i.e. 35 and more hours per week; Part-time, i.e. 1-35 hours workinghours per week (2) Population: women of the birth cohorts 1961-1980 who have at least onechild (3) Source: Mikrozensus 1997 (own estimations)

12

1.3 Female Employment and Marriage Decisions – Hypotheses

Based on the previous empirical findings, one must ascertain that West German women

almost exclusively withdraw from a full-time employment career after the birth of the

first child. Since the German institutional framework sets high incentives to get married

when one partner is permanently not employed, West German women should be very

likely to get married when having the first child. Similar to their West German

counterparts, East German women take advantage of maternity leave regulations and

withdraw from the labor market after childbirth. But, in contrast to West Germans, 30

percent return to full-time employment when the youngest child reaches three years of

age. If they are not employed, they express a strong desire to return to gainful

employment then. Given that East German women are more likely to be full-time

employed or to seek for full-time employment, the long-term gains of marriage should

be relatively small for them.

In short, we expect that the differences in the labor market orientation among East and

West German mothers are able to explain the East-West differences in non-marital

childbearing.

2 Empirical Analysis

2.1 Procedure of the Analysis

Our empirical analysis consists of three broad parts:

(1) In the first step, we investigate how marriage and first birth are coupled. This

aspect is crucial to encircle our research topic.

(2) In the second part, we investigate the hypothesis that high non-marital birth rates in

the East primarily reflect a high labor market orientation. The key indicator to

13

measure the labor market orientation of women is the educational attainment and

the labor market status. We estimate a multi-nominal logit model on the probability

to be single, in a marital-union or in a cohabiting union at the date of interview.

(3) In a third step, we encircle our research topic more narrowly. In this part of the

analysis, we exclude single mothers from the analysis and focus on women in

marital and cohabiting unions. We estimate several logistic regression models on

the probability to be married (versus not married) at the date of interview. Again,

our key independent variables are the employment status and educational attainment

of the woman. We make a distinction by relative and absolute educational

attainment. By “relative”, we understand the woman’s educational attainment in

relation to the one of her partner.

2.2 Data Source

Our empirical analysis is based on data from the German micro-census of the year 1997

(in the following “Mikrozensus”). The Mikrozensus is a one-percent sample of the

population residing in Germany. For West(ern) Germany, it has been conducted

annually since the year 1957 (except for the years 1975, 1983 and 1984). For Eastern

Germany, the first survey was conducted in 1991 (for details, see e.g. EMMERLING/

RIEDE 1997; SCHIMPL-NEIMANNS 1998). It covers standard demographic characteristics

(such as age, nationality, region of residence), employment status, educational

attainment etc. The major advantage of the Mikrozensus is its relatively large sample

size. One major disadvantage is that the Mikrozensus is a cross-sectional data set which

hardly provides any retrospective information. This particularly applies to the “fertility

history” of the respondents. However, it is possible to reconstruct a woman’s “fertility

history” from the number of children who live in the household at the time of survey.

14

Because this strategy involves a variety of problems – i.e. the older a women the more

likely her children have already moved out of the parental home – we restrict our

analysis to females born between 1961-1980, i.e. respondents who were between age 16

and 37 at the time of interview. Assuming that hardly any births occur before age 19,

children of these cohorts were at a maximum age 19 at the time of interview. We only

use persons who live in private households and we omit all cases, where a birth

occurred before age 17. For the multivariate analysis, we restrict the sample to mothers

with children age 3-6. We omit all mothers with children age 7 and older, since those

children already were born (or at least conceived) during GDR times. Mothers with

children age 2 or younger are omitted, since marriage rates are still rather high in the

months after a first child is born, while hardly any marriages occur after the first child

has reached three years of age (see below).

Table 2 comprises the composition of the sample (see Table 9 in the Appendix for a

more detailed overview on the variables used.) As expected, there is a much lower ratio

of cohabiting unions with children in West Germany. Most unmarried West German

women with children (age 3-6) are single mothers, i.e. they do not live together with a

male partner. In the East, however, the pattern is vice versa. The very large majority of

unmarried women with small children are living with a partner. Still, 16 percent of all

women with small children are single parents. East Germans have their first child at a

slightly younger age, but they are less likely to have a second child (see also

KREYENFELD 2000, 2001; SACKMANN 2000). East German mothers have on average a

higher educational attainment than their West Germans counterparts and a much higher

full-time employment rate. While 33 percent of them are full-time employed, this

applies to only 9 percent in the West. In light of the rather unfavorable employment

15

situation, it is not too surprising that the partner’s non-employment rates is higher in the

East than in the West.

Table 2: Composition of the Sample (in percent)

All Women Women with PartnerWest East West East

Family Form Married/ Widowed/ Divorced 88 64 96 74 Marital Union 4 21 4 26 Single 8 16 -- --

Age 20-24 7 11 6 9 25-28 23 40 22 39 29-32 39 37 39 39 32-36 31 13 32 13

Education No degree 23 7 22 6 Vocational degree 66 77 68 78 College degree 8 12 8 13 In education 2 5 2 4

Employment StatusIn education 2 5 2 4 Full-time employed 10 34 9 33 Part-time employed 26 19 26 20 Not employed 7 25 6 24 “Housewife” 42 12 44 13 Missing information*) 13 6 14 7

2nd Child 51 28 55 32

Education (Partner) No Degree -- -- 14 4 Vocational degree -- -- 67 79 College degree -- -- 16 14 In education -- -- 3 3

Employment Status (Partner) Full-time employed -- -- 88 84 Not employed or part-time employed -- -- 9 13 In education 3 3

Sample Size 6,509 1,000 5,735 774Note: (1) Selection of the sample: women of the birth cohorts 1961-1980 where the first child was

born between 1991 and 1994 (2) *) Not employed, but missing information whether therespondent is looking for employment (3) Source: Mikrozensus 1997

16

2.3 “Not married”: A Permanent Family Status?

In a first step, we address the question to what extent non-marital parenthood is a

permanent living arrangement. The non-marital birth rate provided by the German

Statistical Office (see Figure 1) is a very crude indicator for marriage patterns, since it

classifies births as “non-marital” even in case the marriage occurs shortly after

childbirth. I.e. it may very well be possible that West Germans avoid an “illegitimate”

birth, while East Germans are less concerned about postponing marriage to the period

shortly after the birth of the first child.

Furthermore, non-marital birth rates mix births of different orders. For East Germany,

one could argue that non-marital birth rates for first births are high, but that East

Germans get married at the birth of the second child. A possible rationale behind is that

two or more children are a serious impediment for the labor market career of a woman.

In order to empirically investigate this issue we consider all women who had a first (or

second) birth between 1990 and 1997. Using life table techniques, we show how

“marriage” is scattered around the birth of the first and the second child.

As can be seen from Panel 1 in Figure 3, roughly 80 percent of the West German

women are married at the end of the year when they have their first child. In the East,

this applies to roughly 45 percent only. When the first child reaches one year of age, 49

of the East and 81 of the West Germans are married. After three years, the picture looks

almost the same: 53 percent of the East and 83 percent of the West Germans are

married. This basically means that high non-marital birth rates in East Germany do not

result from a postponement of marriage to the period shortly after first birth.

In Panel 2 of Figure 3 we display the survival curves by the age of the second child.

Roughly 75 of the East and 90 percent of the West German women are married once the

second child is born. Since second birth risks rapidly declined after Unification

17

(KREYENFELD 2001; SACKMANN 2000), we can conclude that the high ratio of non-

marital births is partially a composition effect.

Figure 3: Transition to First Marriage by Age of Child

Panel 1: Transition to First Marriage by Age of First Child

0

0.5

1

-5 -4 -3 -2 -1 0 1 2 3 4 5

East

West

Panel 2: Transition to First Marriage by Age of Second Child

0

0.5

1

-5 -4 -3 -2 -1 0 1 2 3 4 5

East

West

Note: (1) Figure 3a: Women of the Cohort 1991-1997, who had a first birth beforecensoring in 1997; Figure 3b: Women of the Cohort 1991-1997, who had a secondbirth before censoring in 1997; (2) Source: Mikrozensus 1997

18

2.4 Marital Union, Cohabiting Union, and Single Parenthood

In the subsequent analysis, we investigate the determinants of living in a non-marital

union with children (versus single or marital union with children). Ideally, one would

presumably analyze the transition to the first marriage in an event history framework.7

Unfortunately, the Mikrozensus 1997 is a cross-sectional data set, which hardly

provides any retrospective information. Instead of an event history model, we therefore

simply estimate a logistic regression model on the probability to be married at the time

of interview. Since the analysis of the survival curves (see Figure 3) revealed that

hardly any marriages occur after the first child reaches three years of age, it seems safe

to consider women who are still not married when the first child reaches three years of

age to be “immune” against marriage.

We use the employment status and educational (vocational) attainment of the woman

and her partner as indicators for the labor market orientation and/ or labor prospects of

the respondents. This is, particularly in the East German case, very sensible, since labor

market chances are strongly correlated with educational attainment (BRINKMANN/

WIEDEMANN 1995: 330; MAYER/ DIEWALD/ SOLGA 1999). We distinguish between “no

degree”, “vocational degree”, “college degree” and “in education” at the date of

interview. We furthermore control for the employment status at the date of interview.

This is not completely sensible, because the marriage decision was taken in the past and

it is impossible that the current employment status influenced a past decision to get

married (see also HOEM 1996). Nevertheless, we interpret the current labor market

status as a valid indicator for the general labor market orientation and labor market

7 There is an often stated believe that it is more advantageous to analyze marriage risks in an event

history framework, using the ‘employment status’ and the ‘pregnancy status ’ (or an interaction ofthese two parameters) as covariates (see Andersson 1998; Blossfeld/ Huinink 1991). Although thedecision to get married might be motivated by the intention to have a child later on and tosubsequently withdraw from the labor market, most women are in fact still childless and in full-time employment at the time they get married. Against this background, it is questionable if anevent history model would in fact be more appropriate to reveal the causal mechanisms that relateemployment, marriage and fertility decisions.

19

prospects. To take into account that some women have deliberately withdrawn from the

labor market, we distinguish “housewives” from “not employed women”. While the

latter are seeking for employment, “housewives” are not employed and do not seek for

employment. Furthermore, we add the woman’s current age and an indicator variable

for having a second child to the regression.

In a first step, we estimate a multi-nominal logit model on the probability of a woman to

be single, to live in a marital union or to live in a cohabiting union (when her first child

is between 3 and 6 years of age). We make a distinction between single mothers and

women who are living in cohabiting unions, because the two groups are eligible for

different transfer payments (see above). According to our theoretical assumptions,

women with very poor labor market prospects are more likely to remain single (instead

of forming a marital or cohabiting union). On the other hand, women with good labor

market prospects should be more likely to remain in a cohabiting union (instead of

getting married).

Table 3 displays the results from the multinominal logit model on the family status of

East and West German women. Do we find the expect pattern, i.e. is the woman’s

employment and high educational attainment negatively correlated with living in a

marital (versus cohabiting) union? For West German women, we find the expected

pattern that “housewives” are more likely to be married. There is, however, no

significant impact of educational attainment on the probability of being married. For

East Germans, we find no correlation between the woman’s employment and marital

status, but a strong positive impact of having a college degree on the probability to be

married. This result opposes our expectations. We will come back to this issue further

down.

As said above, in the FRG there are basically no transfer payments which favor

cohabiting unions. However, some transfers might set incentives towards single

20

parenthood, especially for women who face poor labor market prospects. Yet, our

analysis provides rather equivocal evidence. In East Germany, single mothers are more

likely to be employed, which partially contradicts the widespread believe that single

mothers are, in the first place, welfare recipients. On the other hand, single mothers are

more likely to have no formal degree and, hence, very poor labor market prospects,

which fits this hypothesis very well.

Unfortunately, we do not know anything about the “partnership status” of single

parents, i.e. we do not know if they are single parents because they split up or whether

they simply did not move together with a partner. Therefore, we can only cautiously

interpret the results on the probability to remain a single mother. Since we are not able

to incorporate the partnership status and the disruption risk into the analysis, we will,

for the subsequent analysis, omit single parents from the analysis and concentrate on

women who are living in marital or cohabiting unions (when the youngest child is age

3-6).

21

Table 3: Multinominal Logit Model on “Marital Union”, “Single” versus

“Cohabiting Union”

West Eastb exp(b) t b exp(b) t

Marital Union versus Cohabiting UnionIntercept 3.47 32.26 17.80 *** 1.41 4.08 4.48 ***

Age 20-24 -0.59 0.55 -2.32 ** -1.11 0.33 -3.82 *** 25-28 -0.30 0.74 -1.75 * -0.54 0.58 -2.80 *** 29-32 0 0 32-36 0.39 1.47 2.12 ** -0.12 0.89 -0.41

Education In education -1.59 0.20 -5.01 *** -0.04 0.97 -0.08 No degree -0.03 0.97 -0.18 0.06 1.07 0.18 Vocational degree 0 0 College degree -0.23 0.80 -0.88 0.59 1.81 1.90 *

Employment Status Full-time employed -1.04 0.35 -4.60 *** -0.09 0.91 -0.31 Part-time employed -0.70 0.50 -3.58 *** 0.27 1.31 0.79 Not employed -0.70 0.50 -2.57 *** -0.33 0.72 -1.12 “Housewife” 0 0 Missing information -0.35 0.71 -1.31 0.66 1.94 1.40

2nd Child 0.77 2.16 4.81 *** 0.22 1.24 1.06

Single versus Cohabiting UnionIntercept 0.90 2.47 3.95 *** -0.66 0.52 -1.42

Age 20-24 0.19 1.21 0.67 0.25 1.29 0.73 25-28 0.06 1.06 0.30 -0.04 0.96 -0.15 29-32 0 0 32-36 0.36 1.43 1.65 * 0.14 1.15 0.36

Education In education -0.45 0.64 -1.22 0.58 1.79 0.97 No degree 0.39 1.47 1.87 * 0.83 2.30 2.14 ** Vocational degree 0 0 College degree -0.30 0.74 -0.94 -0.81 0.44 -1.49

Employment Status Full-time employed -0.33 0.72 -1.24 0.66 1.93 1.47 Part-time employed -0.39 0.68 -1.70 * 0.49 1.63 0.99 Not employed 0.29 1.34 0.97 0.17 1.19 0.39 “Housewife” 0 0 Missing information -0.46 0.63 -1.40 1.19 3.29 1.84 *

2nd Child -0.38 0.69 -1.96 * -0.62 0.54 -2.07 **Note: (1) Selection of the sample: women of the birth cohorts 1961-1980 where the first child was born between 1991 and

1994 (2) Method: multi-nominal logit model; dependent variable: marital union, single versus cohabiting union (3)Source: Mikrozensus 1997

22

2.5 The Role of the Woman’s Labor Market Orientation for living in a

Marital versus Cohabiting Union

In the subsequent analysis, we investigate the determinants of living in a cohabiting

versus a marital union. Apart from the woman’s characteristics, we also take into

account the partner’s educational attainment and his employment status. In this

analysis, we proceed in three steps. In a first step, we simply add the partner’s

characteristics to the model. In a second step, we focus on the “relative differences”

between the woman’s and the partner’s educational attainment. In a third step, we test

whether the East-West differences are significantly different from one another.

1. The Partner’s Characteristics

Table 4 displays the results from a logistic regression model on the probability to be

married (versus cohabiting). One could expect that women with partners with poor

labor market prospects have lower marriage risks. The empirical analysis only partially

supports this hypothesis. There is a strong negative effect of having an unemployed

partner on being married in East and West Germany. The partner’s educational

attainment has basically no effect on the marital status. Regarding the woman’s

characteristics, the results remain basically unchanged. Again, having a college degree

has a strong and positive impact on the probability to be married in East Germany.

23

Table 4: Logistic Regression Model “Marital Union versus Cohabiting Union”;

Focus of Model: Partner’s Characteristics

West East

b exp(b) t b exp(b) t b exp(b) t b exp(b) t

Intercept 3.42 30.59 17.46 *** 3.39 29.68 16.72 *** 1.28 3.61 3.99 *** 1.38 3.97 4.15 ***

Age

20-24 -0.64 0.53 -2.48 ** -0.53 0.59 -2.04 ** -1.14 0.32 -3.73 *** -1.18 0.31 -3.79 ***

25-28 -0.31 0.74 -1.74 * -0.25 0.78 -1.39 -0.52 0.59 -2.63 *** -0.55 0.57 -2.76 **

29-32 0 0 0 0

32-36 0.39 1.48 2.12 ** 0.37 1.44 1.97 * -0.13 0.88 -0.43 -0.12 0.89 -0.39

Education

In Education -1.72 0.18 -5.24 *** -1.72 0.18 -5.17 *** -0.42 0.66 -0.87 -0.44 0.64 -0.91

No Degree 0.01 1.01 0.08 0.05 1.05 0.26 -0.04 0.96 -0.11 -0.20 0.82 -0.50

Vocational Degree 0 0 0 0

College Degree -0.16 0.85 -0.62 -0.15 0.86 -0.54 0.61 1.84 1.88 * 0.70 2.01 2.01 **

Employment

Full-time employed -1.11 0.33 -4.92 *** -1.04 0.35 -4.55 *** -0.12 0.88 -0.40 -0.14 0.87 -0.46

Part-time employed -0.71 0.49 -3.64 *** -0.71 0.49 -3.63 *** 0.43 1.53 1.24 0.42 1.52 1.20

Not employed -0.78 0.46 -2.82 *** -0.66 0.52 -2.36 ** -0.37 0.69 -1.19 -0.39 0.68 -1.24

“Housewife” 0 0 0 0

Missing -0.31 0.73 -1.14 -0.36 0.70 -1.33 0.69 2.00 1.45 0.59 1.81 1.23

2nd child 0.81 2.26 5.06 *** 0.83 2.30 5.16 *** 0.33 1.39 1.53 0.40 1.50 1.85

Education (Partner)

In education -0.17 0.84 -0.43 0.03 1.03 0.06

No Degree 0.23 1.26 1.03 0.61 1.84 1.26

Vocational Degree 0 0

College Degree 0.27 1.31 1.47 -0.26 0.77 -1.10

Employment (Partner)

Not empl. / part-time -0.82 0.44 -4.29 *** -0.38 0.68 -1.53

Full-time employed 0 0

Note: (1) Selection of the sample: cohabiting or married women of the birth cohorts 1961-1980 where the first child wasborn between 1991 and 1994 (2) logistic regression; dependent variable: marital versus cohabiting union (3) Source:Mikrozensus 1997

2. Relative Educational Attainment

The crucial aspect is however less, how the “absolute” labor market position affects

marriage decisions but rather, how the partner’s labor market status matters in relation

to the woman’s labor market status. As argued above, in Germany, the relative gains of

marriage are the higher the more unequal the labor market position of the woman and

her partner. We address this issue by using the woman’s and her partner’s educational

24

attainment as an indicator for his and her labor market status and labor market

prospects.

Table 5 comprises various combinations of the educational attainment of the woman

and her partner. It is worth noting that in East Germany assortative mating is more

common than in the West (for details, see WIRTH 2000). This partially reflects a higher

educational attainment among East German women. About 83 percent of the East

German mothers live with a partner who has the same educational level while in the

West, this only applies to 72 percent.

Table 5: ‘Mating Patterns’ (in percent)

West East

Both college degree 6% 7%

Both vocational degree 56% 74%

Both no degree 10% 2%

Women’s education > Partner’s education 7% 7%

Women’s education < Partner’s education 22% 9%

Sample Size 5,503 727Note: (1) Selection of the sample: cohabiting or married women of the birth cohorts 1961-1980 where

the first child was born between 1991 and 1994; Women in education or with partner in educationwere omitted. (3) Source: Mikrozensus 1997

Table 6 contains the results of a logistic regression model on the probability to be

married versus cohabiting (when the first child is between 3 and 6 years of age). The

key independent variables are the various combinations of the woman’s and the

partner’s educational attainment. The results for West Germany partially fit our

hypothesis. Women who are better educated than their partners are less likely to be

married (compared to couples where both have a vocational degree). Surprisingly, the

results for East Germany do not fit our hypothesis at all. Women with higher

25

educated partners have about the same marriage risks as couples where both have a

vocational degree. Even more surprisingly, women who are better educated than their

partners have the highest (!) marriage risks (compared to all other categories).

Table 6: Logistic Regression Model “Marital Union versus Cohabiting Union”;

Focus of Model: Mating Patterns

West Eastb exp(b) t b exp(b) t

Intercept 2.89 17.99 20.96 *** 1.26 3.53 7.52 ***

Alter 20-24 -0.50 0.61 -1.90 * -1.32 0.27 -4.29 ***

25-28 -0.17 0.84 -0.91 -0.56 0.57 -2.77 ***

29-32 0 0

32-36 0.3 1.35 1.62 -0.19 0.83 -0.62

Education

Both college degree 0.16 1.17 0.41 0.51 1.67 1.27

Both vocational degree 0 0

Both no degree 0.30 1.35 1.07 0.21 1.23 0.37

Women’s education > Partner’s eduction -0.63 0.53 -2.66 *** 1.08 2.94 2.21 **

Women’s education < Partner’s eduction -0.29 0.75 -1.65 * 0.00 1.00 0.01

2nd child 1.07 2.92 6.86 *** 0.40 1.49 2.01 **Note: (1) Selection of the sample: cohabiting or married women of the birth cohorts 1961-1980 where the first child was

born between 1991 and 1994; Women in education or with partner in education were omitted. (2) logistic regression;dependent variable: marital versus cohabiting union (3) Source: Mikrozensus 1997

3. East-West Comparison

Since we estimated two separate models for East and West Germans, we cannot say for

sure whether the differences are statistically different. We only know that more highly

educated women (particularly when they have a less educated partner) have higher

marriage risks in East Germany. But we cannot tell if more highly educated East

German women are also more likely to be married than their West German counterparts.

In order to test whether the East-West differences are statistically significant, we pool

East and West Germans into one sample and estimate a single regression. To allow the

covariates to vary flexibly for East and West Germans, we interact them with a binary

26

variable for “West German”.8 In a first step, we use West German women who are

better educated than their partners as reference category. In a second step, we change

the reference category. The reason for changing the reference category is to test for each

single subgroup if East and West Germans differ.

Table 7 displays the results from the full regression and Table 8 the results from the

model after changing the reference categories. The major result from this analysis is

that, for almost all subcategories, East German women are more likely to be unmarried.

The only exceptions are women who have a higher educational attainment than the male

partner. For this subcategory, East and West Germans do not differ. This result sheds a

new light on the previous findings. While, so far, we would have argued that college

educated East German women (particularly when they are living with a partner with a

lower educational attainment) are more likely to get married, we now have to assess that

this is only true in comparison to their East German counterparts. However, compared

to their West German counterparts, they have the same marriage risks.

8 By interaction models, it is usually understood that two (or more) independent variables aremultiplied and added to the regression. The interpretation of the coefficient is however tedious,particularly when there are more than two variables involved. A more straightforward way is toconstruct “new variable combinations”. E.g. if one wants to interact two binary variables, oneconstructs four new binary variables which contain all possible combinations of the initially twobinary variables. One of the four binary variables is then used as references category.Mathematically, there is no difference between the two procedures.

27

Table 7: Logistic Regression Model “Marital Union versus Cohabiting Union”;

Focus: East-West Differences

b exp(b) t

Intercept 2.09 8.06 4.34 ***

Age

20-24 -0.88 0.42 -4.51 ***

25-28 -0.35 0.71 -2.56 ***

29-32 0

32-36 0.14 1.15 0.87

2nd child 0.81 2.24 6.64 ***

Education

Both no degree & West -0.98 0.38 -1.36

Both no degree & East 1.42 4.12 2.60 ***

Both vocational degree & West -1.10 0.33 -2.27 **

Both vocational degree & East 1.00 2.73 2.06 **

Both college degree & West -0.57 0.57 -0.92

Both college degree & East 1.19 3.29 1.93 *

Women’s education>Partner’s education & West 0

Women’s education > Partner’s education & East 0.38 1.46 0.72

Women’s educationPartner’s education & West 0.38 1.46 0.72

Women’s education>Partner’s education & East 0

Model 2b Both college degree & West 1.76 5.79 3.21 ***

Both college degree & East 0

Model 2c Both vocational degree & West 2.11 8.23 14.51 ***

Both vocational degree & East 0

Model 2d Both no degree & West 2.39 10.92 4.02 ***

Both no degree & East 0

Model 2e Women’s education

28

3 Conclusion

In this paper, we investigated the high rates of non-marital childbearing in East

Germany in the 1990s. “Non-marital childbearing” is a rather complex explanandum,

which unifies a bundle of heterogeneous family forms. In order to analyze non-marital

childbearing, it is indispensable to distinguish single parents from couples who have

children in marital and cohabiting unions. Our primary focus of the analysis was the

cohabiting unions with children, or more specifically, couples who live in cohabiting

unions and are still not married when the youngest child reaches three years of age. We

hypothesized that a high labor market orientation among East German women could

explain the high ratio of cohabiting unions with children (and with it the high ratio of

non-marital childbearing).

For West Germans, we find the expected pattern. A relatively high female labor market

orientation is negatively correlated with being married. For East Germany, however,

we do not find the expected pattern. Quite the opposite is true: East German women

with a college degree are more likely to get married than less educated East German

women. Furthermore, East German women with a relatively higher educational level

than their partners are significantly more likely to be married (than any other reference

category).

How do these results contribute to understanding the high non-marital birth rates in East

Germany? At first sight, it seems puzzling that more highly educated women are more

likely to be married in East Germany. However, studies for other countries (particularly

the US and Sweden) report similar findings (e.g. BLOSSFELD 1995; OPPENHEIMER

1995). In the US-American context it is frequently argued that more highly educated

women contribute more assets to the relationship; getting married insures these assets

(e.g. DUVANDER 1999; WHITE/ROGERS 2000; TEACHMAN ET AL. 2000). It is

29

furthermore argued that with the decline in men’s economic status, women have gained

the role of “family co-providers” (COONTZ 1997: 57). Possibly, these aspects also apply

to East Germany. However, the relative higher marriage risk for more highly educated

women is only valid if one compares college educated women with other women who

hold less than a college degree. Compared to their West German counterparts, they are

less likely to be married.

Against the background of our empirical findings, we have to reject our main

hypothesis: It is not possible to decompose the East-West differences in non-marital

childbearing by differences in the employment behavior of East German women. But

how can we interpret the increasing role of non-marital births then?

� Possibly, the standard variables which are usually associated with greater economicindependence (such as educational attainment or the employment status at the date of

interview) are no valid indicators for East Germany. Due to the relative high provision

rate with public day care, women are in general able to reconcile childrearing and

employment. In contrast to their West German counterparts, who basically do not have

much choice but to withdraw from full-time employment after childbirth, East German

women have the option to choose between the ‚Male Breadwinner Model‘ and the

‚Dual-Earner Model‘ when they have children. In other words, in East Germany

economic independence is not a privilege for more highly educated women. Instead,

the overwhelming majority of women take women’s employment for granted. Against

this background, our final conclusion is that East Germans do not conceptualize

marriage as an institution to support the non-employment of one partner.

30

Acknowledgment

For valuable comments we would like to thank Johannes Huinink, Heike Trappe and the

participants of the conference on the 2nd demographic transition in Europe which was

held in Bad Herrenalb in June 2001.

31

References

Adler, Marina A. (2001): German Unification as a turning point in East Germanwomen’s life course: what happens to the standard female biography? University ofMaryland. Mimeo.

Andersson, Gunnar (1998): Trends in marriage formation in Sweden 1971-1993. In:European Journal of Population 14: 157-178.

Beckmann, Petra/ Engelbrech, Gerhard (1999): Beschäftigungsentwicklung undPerspektiven ostdeutscher Frauen in den 90er Jahren. In: Wiedemann, E. (et al.) (eds.):Die arbeitsmarkt- und beschäftigungspolitische Herausforderung in Ostdeutschland.Nürnberg: Bundesanstalt für Arbeit: 203-226.

Bennett, Neil G./ Blanc, Ann Klimas/ Bloom, David E. (1988): Commitment and themodern union: assessing the link between premarital cohabitation and subsequentmarital stability. In: American Sociological Review 53(1.): 127-138.

Berghahn, Sabine (1992): Frauenrechte im Vereinigungsprozeß. Das Ende derPrivilegien oder sozialer Kahlschlag? In: Faber, C./ Meyer, T (eds.): Unterm neuenKleid der Freiheit: Das Korsett der Einheit. Auswirkungen der deutschen Vereinigungfür Frauen in Ost- und Westdeutschland. Berlin: Ed. Sigma: 63-98.

Blossfeld, Hans-Peter (eds.) (1995): The New Role of Women: Family Formation inModern Societies. Boulder, Colo. (et al.).

Blossfeld, Hans-Peter/ Huinink, Johannes (1991): Human capital investment or normsof role transition? How women’s schooling and career affect the process of familyformation. In: American Journal of Sociology 97(1): 143-168.

Braun, Michael/ Scott, Jacqueline/ Alwin, Duane F. (1994): Economic necessity or self-actualization? Attitudes toward women's labour-force participation in East and WestGermany. In: European Sociological Review 10 (1): 29-47.

Brinkmann, Christian/ Wiedemann, Eberhard (1995): Arbeitsmarktrisiken imostdeutschen Transformationsprozess: Ergebnisse des Arbeitsmarkt-Monitors 1989 bis1994. In: Mitteilungen aus der Arbeitsmarkt- und Berufsforschung 28(3): 323-338.

Bundesministerium für Arbeit und Sozialordnung (BMA) (2000): Übersicht über dasSozialrecht. 6th revised and enlarged edition. Bonn.

Coontz, Stephanie (1997): The Way We Really are. New York.

Council of Europe (1999): Recent Demographic Developments in Europe. Strasbourg.

Cromm, Jürgen (1998): Familienbildung in Deutschland: soziodemographischeProzesse, Theorie, Recht und Politik unter besonderer Berücksichtigung der DDR.Opladen/ Wiesbaden.

Dorbritz, Jürgen (1997): Der demographische Wandel in Ostdeutschland - Verläufe undErklärungsansätze. In: Zeitschrift für Bevölkerungswissenschaft 22: 239-268.

Dorbritz, Jürgen (2000): Europäische Fertilitätsmuster. In: Zeitschrift fürBevölkerungswissenschaft 25: 235-266.

�������, Sonja (2000): The effects of children on married and lone mother’semployment in the United States and (West) Germany. In: European Sociological

32

Review 16(2): 137-157.

Duvander, Ann-Zofie E (1999): The transition from cohabitation to marriage: alongitudinal study of the propensity to marry in Sweden in the early 1990s. In: Journalof Family Issues 20: 698–717.

Emmerling, Dieter/ Riede Thomas (1997): Vierzig Jahre Mikrozensus. In: Wirtschaftund Statistik 3: 160-174.

Engelbrech, Gerhard (1997): Erziehungsurlaub – und was dann? Die Situation vonFrauen bei ihrer Rückkehr auf den Arbeitsmarkt: ein Ost/ West-Vergleich. In: IABKurzbericht 8.

Frerich, Johannes/ Frey, Martin (1993): Handbuch der Geschichte der Sozialpolitik inDeutschland. Band 2: Sozialpolitik in der Deutschen Demokratischen Republik.München (et al.).

Frerich, Johannes/ Frey, Martin (1996): Handbuch der Geschichte der Sozialpolitik inDeutschland. Band 3: Sozialpolitik in Deutschland bis zur Herstellung der deutschenEinheit. 2.nd edition. München: Oldenbourg Verlag.

Gysi, Jutta/ Speigner, Wulfram (1983): Changes in the life patterns of families in theGerman Democratic Republic. Berlin: Institut für Soziologie und Sozialpolitik an derAkademie der Wissenschaft der Deutschen Demokratischen Republik (eds.).

Hoem, Jan (1996) The harmfulness or harmlessness of using and anticpatory regressor:how dangerous is it to use education achieved as of 1990 in the analysis of divorce risksin earlier years? Yearbook of Population Research in Finland 33: 34-43.

Höhn, Charlotte (1992): Population-relevant policies before and after Unification ofGermany. In: Krishnan, P./Tuan, Chi-Hsien/Mahadevan, K. (eds.): Readings inPopulation Research. Dehli, India.

Höhn, Charlotte/ Dorbritz, Jürgen (1995): Zwischen Individualisierung undInstitutionalisierung - Familiendemographische Trends im vereinten Deutschland. In:Nauck, B./ Onnen-Isemann, C. (eds.): Familie im Brennpunkt von Wissenschaft undForschung. Neuwied (et al.): 149-176.

Holst, Elke/ Schupp, Jürgen (1999): Erwerbsbeteiligung und Arbeitszeitwünsche 1993und 1997. West- und Ostdeutschland im Vergleich. In: Glatzer, Wolfgang/ Ostner, Ilona(eds.): Deutschland im Wandel: sozialstrukturelle Analysen. Sonderband der ZeitschriftGegenwartskunde. Opladen.

Huinink, Johannes (1998): Ledige Elternschaft junger Frauen und Männer in Ost undWest. In: Metze, R. (et al.) (eds.): Der Transformationsprozess: Analysen und Befundeaus dem Leipziger Institut für Soziologie. Leipzig: 301-320.

Huinink, Johannes (1999): Die Entscheidung zur Nichtehelichen Lebensgemeinschaftals Lebensform – Ein Vergleich zwischen Ost- und Westdeutschland. In: Klein, T./Lauterbach, W. (Hrsg.): Nichteheliche Lebensgemeinschaften. Opladen: 113-138.

Huinink, Johannes/ Wagner, Michael (1995): Partnerschaft, Ehe und Familie in derDDR. In: Huinink, J. (et al.) (eds.): Kollektiv und Eigensinn. Lebensverläufe in derDDR und danach. Berlin: 145-188.

Kaufmann, Franz-Xaver/ Kuijsten, A./ Schulze, H./ Strohmeier, K. (1997): Family Lifeand Family Policies in Europe. Oxford.

Kreyenfeld, Michaela (2001): Labor Markets, Family Policies and Fertility Decisions –

33

East Germany in the 1990s. Rostock. Ph.D. Thesis.

Kreyenfeld, Michaela (2000): Changes in the timing of first birth in East Germany afterRe-Unification. In: Schmoller’s Jahrbuch - Zeitschrift für Wirtschafts- undSozialwissenschaften 120(2): 169-186.

Mayer, Karl-Ulrich/ Diewald, Martin/ Solga, Heike (1999): Transitions to post-communism in East Germany: Worklife mobility of woman and men between 1989 and1993. In: Acta Sociologica 42(1): 35-53.

Nave-Herz, Rosemarie (1994): Familie heute. Darmstadt.

Obertreis, Gesine (1986): Familienpolitik in der DDR 1946-1980. Opladen: Leske undBudrich.

Oppenheimer, Valerie K. (1995): The role of women’s economic independence inmarriage formation. In: Blossfeld, H.-P. (eds.): The New Role of Women. Boulder (etal.): 236-243.

Peuckert, Rüdiger (1999): Familienformen im sozialen Wandel. 3. Edition. Opladen.

Raley, Kelly R. (2001): Increasing fertility in cohabiting unions. Evidence for thesecond demographic transition in the United States? In: Demography 38(1): 59-66.

Sackmann, Reinhold (2000): Geburtenentscheidung und Lebenslaufpolitik imostdeutschen Transformationsprozess. Zeitschrift für Soziologie der Erziehung undSozialisation: 146-163.

Sainsbury, Diane (1997): Taxation, family responsibilities, and employment. In:Sainsbury, D. (ed.): Gender and Welfare State Regimes. Oxford: Oxford UniversityPress: 185-209.

Schimpl-Neimanns, Bernhard (1998): Analysemöglichkeiten des Mikrozensus. In:ZUMA-Nachrichten 22(42): 91-122.

Schneider, Norbert F./ Matthias-Bleck, Heike (1999): Moderne Familie – altes Recht?In: Busch, F.W. (et al.) (eds.): Aktuelle Forschungsfelder der Familienwissenschaft.Würzburg: 185-210.

Seltzer, Judith (2000): Families formed outside of marriage. In: Journal of Marriage andthe Family 62: 1247-1268.

Statistisches Bundesamt (2000): Bevölkerung und Erwerbstätigkeit. Gebiet undBevölkerung 1998. Fachserie 1, Reihe 1. Stuttgart.

Statistisches Bundesamt (2001a): Altersspezifische Geburtenziffern für Ost- undWestdeutschland für das Jahr 1999. Bereitgestellt durch Dieter Emmerling (StatistischesBundesamt).

Statistisches Bundesamt (2001b): Statistik der Kinder- und Jugendhilfe. VerfügbarePlätze in Tageseinrichtungen für Kinder. Wiesbaden.

Smock, Pamela J. (2000): Cohabiting in the United States. An appraisal of researchthemes, findings, and implications. In: Annual Review of Sociology 26: 1-20.

Teachman, Jay D./ Tedrow, Lucky M./ Crowder, Kyle D. (2000): The changingdemography of America's families. In: Journal of Marriage and Family 62: 1234–1246.

Trappe, Heike (1995): Emanzipation oder Zwang? Frauen in der DDR zwischen Beruf,Familie und Sozialpolitik. Berlin.

34

Tyrell, Hartmann (1985): Literaturbericht. In: BMJFG: NichtehelicheLebensgemeinschaften in der Bundesrepublik Deutschland. Stuttgart: 93-173.

White, Lynn/ Rogers, Stacy J. (2000): Economic circumstances and family outcomes: areview of the 1990s. In: Journal of Marriage and Family 62: 1035–1051.

Wirth, Heike (2000): Bildung, Klassenlage und Partnerwahl. Opladen.

Witte, James C. /Wagner, Gert G. (1995): Declining fertility in East Germany afterunification: a demographic response to socioeconomic change. In: Population andDevelopment Review 21: 387-397.

35

Appendix

Table 9: Description of the Variables

Birth Biography

A woman is assumed to have given birth, if she is labeled as “head of the family” or“partner of the head of the family” and if she is living in the same family with a personwho is labeled as “child in the family”. The age and order of the birth is interfered fromthe age and the number of children in the family, i.e. it is the difference between theyear of birth of the mother and the age of the child.

Family Form

• “Married” refers to all women who report that they are married, divorced or widowedat the time of interview. This is irrespective of if they are living with their partners.

• “Cohababiting” refers to women who are living in a cohabiting union.• “Single” refers to women who do not live with a partner in the same household.

In Figure 3, we use information on the year the respondent got married. It is notpossible to identify if this information relates to a first or higher order marriage. Forsimplicity, we assume that it relates to first order marriages. In contrast to most otherquestions in the Mikrozensus, respondents are free to answer this question. Therefore,the non-response rate for this question is relatively high.

Education

There are three binary variables that indicate the highest degree obtained at the time ofinterview. A distinction is made by:• No degree: This category also includes respondents who only received training on the

job (“Anlernausbildung”).• Vocational degree: “Lehrausbildung”, “Meister”, “Fachschulabschluß”• College degree: University or college degree (“Universitäts”-/

“Fachhoschulabschluß”)

Employment Status

A distinction is made by• Part-time employment (>0 and

36

Employment Status of Partner

We distinguish:• Full-time employed• Not employed or part-time employed• In education or military service

East/ West German

An East German is a respondent who lives in the five new “Länder” or East Berlin in1997. A West German is a respondent who lives in the territories of former WestGermany.