European Journal of Population 13: 269–298, 1997. 269
c
1997 Kluwer Academic Publishers. Printed in the Netherlands.
Wanting a Child without a Firm Commitment to the
Partner: Interpretations and Implications of a
Common Behaviour Pattern among Norwegian
Cohabitants
ØYSTEIN KRAVDAL
Department of Economics, University of Oslo, PB 1095 Blindern, N-0317 Oslo, Norway (Fax:
47-22-855035; E-mail: okravdal@econ.uio.no)
Kravdal, Ø. 1997. Wanting a child without a firm commitment to the partner: Interpretations and
implications of a common behaviour pattern among Norwegian cohabitants. European Journal of
Population.
Abstract. Further knowledge of combined birth and marriage intentions among cohabiting couples
will improve our understanding of the nature of consensual unions and be important from a policy
perspective. According to Norwegian surveys from 1988 and 1996, about 1/3 of the births to cohab-
iting couples are mistimed. Among the remainder, between 1/2 and 3/4 are to couples who at least
have no intention to marry within the next couple of years. The most radical estimate, based on the
most recent survey, is that there are three equally large categories of births to cohabiting couples:
mistimed births, intended births to couples planning marriage, and intended births to couples with
no marriage plans whatever. The 1996 survey also revealed that a clear majority of these couples
who appear to want a child without planning marriage, explain this attitude partly by the less easy
dissolution of a marriage. In other words, their consensual union is indeed considered different from
marriage in terms of commitment and stability and they may have concerns about the quality of the
relationship. There were weak indications that cohabitants with an intended birth in the absence of
marriage plans were less likely than others to consider a parental break-up to be very deleterious for
the child.
Øystein Kravdal. De´sirer un enfant sans engagement ferme du partenaire: interpre´tation et implica-
tions d’une conduite fre´quente parmi les cohabitants norve´giens.
Re´sume´. Des connaissances plus approfondies sur les intentions combine´es de naissance et de
mariage des couples cohabitants, ame´lioreraient notre compre´hension des unions consensuelles.
C’est la` un objectif ne´glige´, mais important, dans les recherches sur les familles et non moindre
dans une optique politique. Selon des enqueˆtes norve´giennes de 1988 a` 1996, environ un tiers des
naissances de couples cohabitants sont non de´sire´es. Parmi les autres, entre la moitie´ et les trois
quarts des naissances surviennent dans des couples sans intention de se marier dans les deux ans a`
venir. Des enqueˆtes plus re´centes montrent qu’il y a trois grandes cate´gories de naissances de couples
cohabitants d’effectif e´quivalent: naissances non de´sire´es, naissances de´sire´es des couples pre´voyant
un mariage, et naissances de´sire´es de couples n’en pre´voyant pas. L’enqueˆte de 1996 re´ve`le aussi
qu’une nette majorite´ de ces couples qui de´sirent un enfant sans le mariage craignent que celui-ci
rende la dissolution de l’union plus difficile. Cela indique la diffe´rence entre union consensuelle et
mariage en terme d’engagement et de stabilite´ et signale le souci au sujet de la qualite´ de l’union. Il y
a de le´ge`res indications selon lesquelles les cohabitants avec une naissance de´sire´e sans mariage sont
moins susceptibles que les autres de conside´rer qu’une se´paration soit tre`s nuisible pour l’enfant.
270 ØYSTEIN KRAVDAL
1. Introduction
Dramatic changes in family behaviour have taken place in a variety of industri-
alized countries during recent decades, but nowhere has the drift away from the
formal marriage been more pronounced than in the Nordic countries. Their posi-
tion as forerunners is most clearly seen with respect to out-of-wedlock fertility and
informal cohabitation (Monnier and Guibert-Lantoine, 1996). For example, more
than 45% of Norwegian babies are now born out-of-wedlock, up from 30% in the
late 1980s and 10% in the mid-1970s (Statistics Norway, 1996). A large majority
of these births are to couples living in consensual unions. This behaviour appears
to be widely accepted. In a nationally representative sample of 28-year old women
in 1988, as many as 60% reported that a ‘consensual union is just as acceptable as
marriage even when the couple have a child’ (unpublished calculations from data
described below). Even among women who themselves had chosen a more tradi-
tional pathway towards parenthood, and were married when they became pregnant
with their first child, the proportion was as high as 43%.1 As yet, the soaring out-of-
wedlock fertility has not provoked much political concern. In a government report
on population trends and policies that was presented as part of the preparations
for the ICPD conference in Cairo, the issue was barely touched (United Nations,
1993).
The above-mentioned government report, as well as some of the media discus-
sion, seems to reflect a notion that consensual unions are almost as stable as
marriages. While few would doubt that informal cohabitation among young child-
less adults tends to be short-lasting, often ending in a break-up rather than marriage,
it is perhaps less obvious that cohabitants who have had a child together face a
particularly high disruption rate. However, recent empirical research has demon-
strated quite convincingly that the large excess fragility is not restricted to the
childless cohabitants. For example, Hoem and Hoem (1992) found in Swedish data
from 1981 that a cohabiting couple with at least one child had a dissolution rate
more than twice as high as an otherwise equal married couple. This excess appears
to have been reduced during the 1980s (Duvander, 1996). Using data from the late
1980s, also some Norwegian studies have shown, more or less explicitely, that
break-ups occur relatively often among currently cohabiting parents, and among
parents who were cohabiting when they had their first child (Blom et al., 1993;
Jensen, 1996).2
In light of this fragility and the possibly harmful effects of a parental split-up
on children’s well-being (e.g. Amato, 1991; Cherlin et al., 1995), the increasing
proportion of births in consensual unions is a highly important research issue –
not least from a policy perspective. A thorough discussion about needs for and
possibilities for influencing the first steps of family-building in order to protect
children, should it ever be initiated, requires a much better empirical platform
and theoretical understanding than currently exist. This study is one step in the
establishment of such knowledge.
WANTING A CHILD WITHOUT A FIRM COMMITMENT TO THE PARTNER 271
For example, one should know more about the extent to which entry into parent-
hood among cohabiting couples is the result of active decison-making (except for
the trivial fact that they at least must have taken the decision not to have an
abortion if the pregnancy was mistimed). It is firmly established that many births
are so-called mistimed, or outright unwanted, even in countries where modern
contraceptive technology is widely used and abortion easily available (e.g. Kost and
Forrest, 1995; Statistics Norway, 1991), and that mistiming occurs more frequently
among the non-married than the married (e.g. Bennett et al., 1995), but less is known
about its prevalence in consensual unions. If it turns out that many births to cohab-
itants are mistimed, one should primarily make efforts to improve contraceptive
use rather than striving to devise a policy meant to influence people’s decisions
about becoming pregnant.
Gaining more knowledge about marriage plans and reasons for the hesitation
to marry is also important, since this gives an indication about the nature of the
consensual unions and the mechanisms behind the statistically well-established
association between cohabitation and disruption. A better understanding of these
mechanisms is crucial for a sound family policy.
Spelling this out more explicitely, marriage itself may prevent disruption. It
is possible that cohabiting couples simply dissolve their unions more often than
the married because of the lower barriers involved in the dissolution process.3
This would be a quite trivial effect of union status that could explain the fragility
of consensual unions, but not that of marriages preceeded by consensual unions.
Axinn and Thornton (1992) may be said to carry this further in arguing that an
informal union tends to weaken the bonds to the partner and to reduce commitment
to the marriage institution more generally, and thus also spur a high dissolution
rate in a subsequent marriage.
On the other hand, there are certain reasons why cohabiting parents are not
married, and for the same reasons they may also be dissolution-prone. In other
words, if they had somehow been forced to marry, children’s experience of parental
disruption would not necessarily have been markedly reduced. Such selection was
suggested to play a dominant role in, for example, a methodologically advanced
study from the United States that focused on the excess divorce risk among previ-
ously cohabiting married couples compared to the directly married (Lillard et al.,
1996), and may be important, though less markedly so, also for the excess fragility
among current cohabitants.
One possible source of selection is individual values, which are likely to be
quite persistent and less relevant as a policy lever. As more thoroughly explained
below, some couples may, for example, prefer cohabitation to marriage because of
purely economic worries about the wedding costs, and this desire may be realized
because they also have a very strong attachment to individual freedom to make
choices regardless of moral authorities. Their intention may not at all be to keep
options more open by not formalizing the union. If the relationship then turns
out to be unsatisfactory, for one reason or another, this liberation from traditional
272 ØYSTEIN KRAVDAL
norms may also be an important co-factor in the decision to split up. An entirely
different possibility is that the couple prefer to remain cohabitants because they are
concerned about the quality of their relationship and consider the informal union
to be more easily dissolved. Such an attitude would serve to verify the existence
of a partly causal effect of marital status on disruption rates, and the doubts about
the quality, which must also be closely linked with actual disruption rates, would
appear as a potentially important selection factor. Such a mechanism is a more
interesting one from a policy perspective, since quality concerns may be expected
to be of a less persistent nature.
In this study, data from the Norwegian Family and Occupation Survey of 1988
and the Statistics Norway Omnibus Surveys of 1996 are used to assess the propor-
tion of mistimed births in consensual unions, and to find out whether the remaining
cohabitants, who deliberately have a child while living in this informal relationship,
are predominantly hesitant towards marriage, or whether they expect marriage to
take place shortly afterwards. Besides, respondents in the Omnibus Survey gave
reasons for not marrying (yet), which are referred and discussed. While many
authors have compared cohabitation with marriage and dating relationships (e.g.
Rindfuss and van den Heuvel, 1990; Santow and Bracher, 1994) – from which
perspective this study would also be interesting – few have attempted to make such
distinctions between different categories within the group of cohabiting parents or
cohabiting couples.
Moreover, for the parents or would-be parents in focus in this study, ideas related
to the child’s situation after a disruption may have a bearing on the decisions. The
available data allowed a check upon whether couples who wanted a child without
intending to marry were less worried than other cohabitants about the consequences
of a disruption upon the child’s well-being. Such a link appears not to have been
addressed previously, but has implications for potential family policy reforms, as
well as for our general understanding of the high prevalence of out-of-wedlock
fertility.
After this discussion of individual considerations behind childbearing in
consensual unions, possible underlying social and cultural forces are pointed out.
The results are reviewed from a policy perspective in the concluding section.
To set the stage before plunging into all these issues, the different pathways
to marriage and parenthood in Norway are illustrated in a simple manner, and the
demographic trends behind the rise in out-of-wedlock fertility are pointed out.4
For simplicity, the focus is on first unions throughout this study, but according to
several checks that have been made, the main results would not have been markedly
different with a broader scope.
2. Data
One of the sources for this analysis was the Norwegian Family and Occupation
Survey of 1988, which contains individual biographies of pregnancies, births,
WANTING A CHILD WITHOUT A FIRM COMMITMENT TO THE PARTNER 273
marital and non-marital unions, educational activity, and employment for a nation-
ally representative sample of 4019 women born in 1945, 1950, 1955, 1960, 1965,
and 1968 (Statistics Norway, 1991). (For an example of previous use of these data,
see Kravdal, 1994). The response rate was 81%. Similar data for men have also
been collected, but were not available for this study in appropriate form.
This survey did not contain data on marriage and childbearing plans at the time
when cohabitants conceived a child, which would have been the most interesting
information from the perspective or this study. However, with due respect to post-
rationalization and the fact that people may change their mind about such issues
quite rapidly, the answers given a few years before or after birth are presumably
reasonably good indicators. The following questions in the Family and Occupation
Survey were relevant: ‘Do you expect to have (more) children?’; if yes, ‘When
would you like to have your next/first child?’; ‘Do you and your cohabiting partner
plan to marry within the next two years?’; ‘When you realized that you were preg-
nant with your first-born child, was the pregnancy wanted but too early, wanted and
well-timed, wanted but too late, or would you say that it was actually unwanted?’5
In this analysis, births reported as ‘unwanted’ have not been distinguished
from those reported as ‘wanted but too early’. Several childless, young women in
the Family and Occupation Survey reported their birth to be ‘unwanted’. However,
given the generally low proportion in the Norwegian population who plan to remain
childless, a more correct description would probably be that they did not intend
to have a child during the next few years, or perhaps were in doubt about having
a child with that partner at all, i.e. that the birth was ‘very strongly mistimed’
(although not so much that it resulted in an abortion, which has been available on
request since the mid-1970s). Also, births that were ‘well-timed’ or ‘too late’ were
grouped together.
The survey was linked with information from the Central Population Register
of Norway about births and marriages during a 3-year period after interview.6
The other data set used in the analysis was the Statistics Norway Omnibus
Survey of October 1996, where there was a national probability sample for both
sexes, but with a response rate that unfortunately only reached 68% (Teigum,
1996). Also this survey included questions about marriage and childbearing plans
and intention status of first birth (restricted to those cohabiting at interview), but
in more detail, and thus supplemented the other survey. The questions were as
follows: ‘When expecting your first child together, did you consider it to be too
early, well-timed, or too late?’; ‘Do you expect to have children together?’; if yes,
‘Do you expect to have children within one year, within 1–3 years, later, or is the
timing uncertain?’; ‘Do you and your partner plan to marry?’; if yes, ‘Have you
thought to marry within 1 year, within 1–3 years, after more than 3 years, or is the
timing uncertain?’. Another important contribution from the Omnibus Survey was
its information about the reasons for hesitation to marry (see Table III for details
about the questions and Kravdal (1997) for a brief discussion of the reliability of
the answers).
274 ØYSTEIN KRAVDAL
All ages between 16 and 79 were represented in the Omnibus Survey, but only
those younger than 43 were considered in this analysis, because this was the highest
age in the Family and Occupation Survey. In tables related to intentions about future
childbearing, an even lower age limit at 39 was set, because this question was not
given to older women.
3. Different routes to parenthood
Figure 1 illustrates the variety in young adults’ life courses, and gives a rough
impression of recent trends in non-marital childbearing. The numbers included are
from the Family and Occupation Survey of 1988. To facilitate comparison of the
two birth cohorts chosen as examples (1945 and 1960), the individual observations
in the survey were censored at age 28. At that age, 88% of both cohorts had entered
their first union. To avoid unneccesary complexity in this introductory example, it
was also censored when a first union was dissolved, which lead to the exclusion of
about 10% of first births in the 1960 cohort, but less than 1% in the 1945 cohort.
With this censoring, 60% of the women in the 1960 cohort were observed to have
become mothers before age 28, compared to 79% in the 1945 cohort.
As many as 47% of the first pregnancies in the 1945 cohort were to single
women, but a large proportion married during pregnancy, leaving only 11% as
unmarried at the time of delivery. (Among these, a large proportion subsequently
married, even quite quickly (not shown), which indicates that they often married
the father of their child). In addition, 1% of the births were to mothers who had
become pregnant while living in a consensual union, and who had not married
during pregnancy.
The picture is, of course, markedly different for the 1960 cohort. Fewer became
pregnant as single women, and within this group fewer married during pregnancy
(46%). On the other hand, a large proportion entered a consensual union during
pregnancy. More often, the first step of family building was cohabitation (63%), and
among these cohabiting women, many became pregnant and gave birth while still
cohabiting. In total, 12% of the first births in this subsample of women younger
than 28 with no previous union dissolution were to single women and 24% to
cohabitants. With respect to pregnancies, 34% were to single and 32% to cohabiting
mothers.
As stated above, out-of-wedlock fertility has continued to increase after the
mid-1980s. According to recent statistics from the Medical Birth Registry (1995),
about 8% of the births (disregarding parity) were to single women in 1995, while
39% were to cohabitants.
The trends in some union formation and pregnancy rates up to 1988 are shown
in Table I as period effects in discrete-time hazard regression models (where it was
not censored at age 28; see notes to the table). In accordance with the numbers
for the 1945 and 1960 cohort shown in Figure 1, a clear movement away from
pregnancies in the single state appears in the estimates. As regards the strong
WANTING A CHILD WITHOUT A FIRM COMMITMENT TO THE PARTNER 275
(a) ‘Pregnancy’ refers to the 9 months before the first live birth. When a pregnancy begins, or a child
is born, at the same time as a change in family status is reported, the latter is assumed to have taken
place first.
(b) 14 were among the 25 who gave birth while cohabiting
(c) 79 were among the 114 with preceding cohabitation
(d) 2 of the 142 who became pregnant we