A Demographic Analysis of Registered
Partnerships (legal same-sex unions):
The Case of Norway
TURID NOACK1,
* ANE SEIERSTAD1
and HARALD
WEEDON-FEKJR2
1
Statistics Norway, P.O.B. 8131 Dep, N-0033 Oslo, Norway; 2
Section of Medical Statistics,
University of Oslo, Norway (*Author for correspondence, e-mail: turid.noack@ssb.no)
Received 25 October 2004; accepted in final form 4 March 2005
Noack T., Seierstad A. and Weedon-Fekjr H., 2005, A Demographic Analysis of Registered
Partnerships (legal same-sex unions): The Case of Norway, European Journal of Population,
21: 89­109.
Abstract. In 1993, Norway became the second country in the world to introduce registered
partnerships. As with same-sex marriages, registered partnerships give the same rights and
duties as marriages for opposite sex couples, with a few exceptions. The present article de-
scribes this `new' demographic event based on Norwegian experiences. Compared to the
number of new cases each year of opposite-sex marriages, registered partnerships amount to
less than 1%. Gay partnerships are more common than lesbian partnerships, but the pro-
portion of female partnerships has increased since the law came into force. In one out of five
partnerships, one or both partners have been previously married to a person of the opposite
sex. At the time they entered into the partnership, 24% of the women and 13% of the men had
at least one child. When examining the registered partnerships entered into in Norway between
1993 and 2001 we estimated the probability of divorce to be 16% after six years. We also
carried out a multivariate analysis of the divorce risk and found that the divorce risk for
female partnerships is double that of the risk for male partnerships. Another high risk group is
partnerships between a Norwegian and a person from a non-Nordic country.
Key words: divorce, Norway, proportional hazards regression, registered partnership, same-
sex marriage
Noack T., Seierstad A. et Weedon-Fekjr H., 2005, Analyse dé mographique des partenariats
enregistré s (unions homosexuelles lé gales). Le cas de la Norve` ge, Revue Europé enne de
Dé mographie, 21: 89­109.
Ré sumé . En 1993, la Norve` ge a é té le deuxie` me pays au monde a` introduire les partenariats
enregistré s. Comme pour les mariages homosexuels, les partenariats enregistré s donnent, a`
quelques exceptions pre` s, les me^ mes droits et devoirs que ceux dé coulant des mariages
hé té rosexuels. Cet article dé crit ce ``nouvel'' é vé nement dé mographique en s'appuyant sur
l'expé rience norvé gienne. Chaque anné e, le nombre de partenariats enregistré s s'é le` ve a` moins
European Journal of Population (2005) 21: 89­109 Ó Springer 2005
DOI 10.1007/s10680-005-3626-z
de 1% du nombre de mariages hé té rosexuels. Les partenariats masculins sont plus fré quents
que les unions lesbiennes, mais la proportion de partenariats fé minins a augmenté depuis les
premie` res anné es d'entré e en vigueur de la loi. Dans un cas sur 5, l'un des partenaires ou les
deux ont é té pré alablement marié s a` une personne du sexe opposé . Au moment ou` ils
concluent le partenariat, 24% des femmes et 13% des hommes ont au moins un enfant. En
prenant en compte les partenariats enregistré s entre 1993 et 2001, le risque de divorce est
estimé a` 16% apre` s 6 ans. Une analyse multivarié e montre un risque de divorce double pour
les unions fé minines comparé a` celui des unions masculines Les couples associant un(e)
Norvé gien(ne) et une personne originaire d'un pays non nordique ont é galement un risque
é levé de sé paration.
Mots clé s: partenariat enregistré , mariage homosexuel, divorce, Norve` ge, ré gression a` risques
proportionnels
1. Introduction
During recent decades, legal recognition of same-sex unions has been
intensely debated and strongly fought for in several countries. In 2004,
particular attention was paid when local authorities in the USA and France
began authorising same-sex marriages, although such unions were not
approved by law. In the end, these marriages were not considered legally
valid, but the ceremonies undoubtedly had a symbolic effect. To date, seven
countries have given same-sex couples legal rights to a registered partnership
or marriage.1
Registered partnership was invented in Denmark, and all the
Nordic countries use that term. As with a same-sex marriage, a registered
partnership has, with few exceptions, all the consequences of an opposite-sex
marriage. There are only minor differences between the various laws in the
Nordic countries.
From a demographic point of view, registered partnerships as well as
same-sex marriages represent a new family type. Although the legal conse-
quences of entering a registered partnership or a same-sex marriage may
differ slightly from country to country, we think that the two phenomena
may be treated as demographically similar events. So far, current statistics
and analyses have been scarce. This article helps provide more knowledge on
this new demographic event. In view of the extensive attention given to
registered partnerships and same-sex marriages both before and after their
legalisation, we may ask how many same-sex couples have actually taken the
opportunity to legalise their relationship. We will also analyse some char-
acteristics of people who have entered into a registered partnership, such as
their age and the age difference between the partners in these cohabitations.
Finally, we will analyse the duration of these cohabitations and present a
multivariate analysis of the divorce risk in registered partnerships. We will
look at the number of divorces during the first years and check whether the
divorce risk is the same for male as for female couples.
TURID NOACK ET AL.90
Our study is based on Norwegian data. Norway seems to be a suitable case
to study for three reasons: first, as early as 1993, Norway became the second
country in the world to introduce registered partnerships, a fact which gives
us the opportunity to study the development over some years. Second, the
Norwegian population statistics provide us with fairly comprehensive
information, i.e. concerning former civil status and the number of children
among those entering into partnerships. Third, Scandinavia has often been
seen as a forerunner when it comes to family change. This contention does, at
least, hold true for legal recognition of same-sex marriages.
2. Background: the legislation of same-sex unions
Although several countries have given some legal recognition of same-sex
couples, many of these regulations have so few legal consequences that they
may not be regarded as corresponding to opposite-sex marriages. A well
known and highly debated case is France. After several years of intense
discussions, same-sex and opposite-sex couples were given the same admis-
sion to contract a `Pacte Civil de Solidarité ' (Civil Solidarity Pact) (PACS).
However, the French law contains very few rights and obligations. As
expressed by Martin and Thé ry (2001, p. 152): ``By signing a Pacs, one party
can consider himself or herself as `almost married' and the other as `almost in
a free union'.''
In some countries, same-sex marriages are legalised at state level, not at
federal level. In 2000, Vermont was the first state in the USA to give legal
recognition to what they call civil union for same-sex couples. Proposals for
constitutional amendments to prohibit same-sex marriages nation-wide, are
evidence that there is considerable opposition in the USA.
At present, legislation on same-sex couples is rapidly being altered in
numerous countries.2
The term `registered partnership' is a new linguistic creation, which is used in
all the Nordic countries. In all these countries, separate acts on registered part-
nerships have been passed and are thus not a part of the Marriage Act. In the
Nordic countries, registered partnership is only available for same-sex couples.
It was hardly a coincidence that the Nordic countries did not want to
incorporate same-sex couples into the Marriage Act. One reason was purely
pragmatic, as it was easier to introduce a new law than to alter the existing
one. Another argument presented in the debate before the Norwegian law
was passed, was a fear that modifying the existing marriage code might be
perceived as a threat towards marriage as a social institution and its unique
position in society. Tactically, it seemed more sensible to propose a new term,
i.e. registered partnership, and a separate act. The Netherlands and Belgium
have chosen a different procedure. In these countries, the rules authorising
A DEMOGRAPHIC ANALYSIS OF REGISTERED PARTNERSHIPS 91
same-sex marriages were incorporated into the existing marriage law in 2001
and 2003, respectively.
Registered partnerships and same-sex marriages give the same rights and
duties as marriages for opposite-sex couples, with a few exceptions
(Waaldijk, 2003). To some degree they are not completely equal to marriage.
There are exceptions in one or more of the following issues: the opportunity
to jointly adopt children, to have medically assisted insemination, to sol-
emnise the partnership and requiring residency before entering into the
partnership. Such exceptions are debated and some exceptions have been
cancelled since the law was first passed.
In Norway, a partnership is formalised by a civil ceremony, much like the
ceremony for heterosexual couples who prefer a civil ceremony to a church
wedding. Couples who enter into partnerships will often make an
announcement in their local newspaper, and many of them invite family and
friends to a large-scale party, very similar to a traditional wedding party.
Registered partnerships have received and still receive quite a lot of general
attention in the media. Several couples with and without children have been
interviewed on the radio, on television, and in newspapers and magazines.
The procedure for dissolving a registered partnership is the same as for
dissolving a marriage. According to the Norwegian law, the main rule is that
one or both partners must apply for a separation licence from the county
governor and can file for divorce after they have been separated for a min-
imum of one year. During the period of separation, they are technically still
registered partners, and cannot therefore enter into a new partnership or
marriage during this period.
In Norway, there is still strong public interest and debate surrounding the
law on registered partnerships. Few want to abolish the opportunity to enter
into a registered partnership, but amendments such as introducing the
opportunity to adopt are controversial.3
Denmark is traditionally seen as a
morally more liberal country than Norway, and the resistance to legal rec-
ognition of same-sex couples seems to be weaker in Denmark than in Nor-
way. A Danish study stressed that few politicians wanted to take on the
burden of opposing the bill on registered partnerships in public (Sland,
1998). An analysis of the proceedings and debate prior to the Norwegian law
being passed also indicated that the politicians had more accepting attitudes
than the population at large (Halvorsen, 1999; Strksen, 2000). Opinion polls
suggested that although a considerable proportion was indifferent, ``there
never seems to have been a general majority in favour of the Act''
(Halvorsen, 1998, p. 213).
Today registered partner, separated partner, divorced partner and surviving
partner are all alternatives of the official civil status in Norway. The juridical
acceptance of a registered partnership will, however, only be valid in coun-
tries with corresponding rules.
TURID NOACK ET AL.92
3. Previous research
3.1. SOME METHODOLOGICAL CHALLENGES OF LESBIAN AND GAY STUDIES
Lesbian and gay studies represent a typical cross-disciplinary research field,
mostly focusing on the lives, identities, and struggle for recognition of
homosexuals. This is also a research field where much work has been pro-
duced outside the universities and some will say ``Lesbian and gay studies still
hover on the borders of full respectability'' (Weeks, 2000, p. 1). Demographic
approaches, such as focusing on the prevalence of same-sex couples and the
duration and the break-up risk of these relationships, are rare. Additionally,
small samples, and even the absence of representative samples, make it dif-
ficult to generalise results from many of the studies, which, after all, have
been conducted. Self-recruited samples, often based on snowball methods,
are common, and respondents are often recruited from organisations of gays
and lesbians or from readers of particular magazines.
The arrival of AIDS increased the demand for epidemiological studies
based on representative samples. More recently, legislative changes and the
growing societal recognition of same-sex partnerships seem to have increased
the interest in more reliable figures, describing prevalence as well as incidence
of same-sex couples. In recent years, studies have been carried out using
census data as well as large-scale surveys (see, for instance, Black et al., 2000,
2002; Badgett, 2001; Jepsen and Jepsen, 2002). Due to the limited size of the
target group, surveys based on representative sampling procedures, will, in
most cases, reveal relatively few respondents having experience from same-
sex relationships of a certain duration. Attempts to solve this problem have
been made by pooling two or more surveys in order to get larger samples (see,
for instance, Badgett, 1995).
The question of homosexual behaviour is also a sensitive topic for large-
scale surveys, and there is obviously a considerable risk of underestimation.
Increased openness and acceptance of homosexuality in general, including
public debates, as we have seen in the fight for legal recognition of same-sex
marriages, may render more acceptability to questions which were considered
unthinkable only a few years ago.
As a family type, same-sex marriages and registered partnerships are
classical demographic events, which in principle are well suited for statistics
and analysis, in the same way as opposite-sex marriages. To make compar-
isons of rates for same-sex marriages, we need an estimate for the risk
population, i.e. the number of homosexuals. More recent research shows,
however, that sexual orientation is not a `once and for all' fact for all indi-
viduals, but may vary in strength and even change over the life course. A
famous Norwegian female athlete who divorced her same-sex partner and
shortly after cohabited and got pregnant with a male partner, put it this way
``I am not a lesbian or bisexual, I am just myself''. She is, in other words, a
A DEMOGRAPHIC ANALYSIS OF REGISTERED PARTNERSHIPS 93
typical example of what is called queer theory, a theory where sexual identity
is not seen as fixed, but may be more fluid and alternating.
Given the assumption that sexual identity may alternate from time to time
and over the life course, the question concerning the number of homosexuals
is difficult to answer and may even be seen as irrelevant. Studies show that the
number of homosexuals may vary depending on the wording of questions
and the issue studied, such as self-identity as gay and lesbian, homosexual
practice at least once in a lifetime, the sex of the majority of partners and so
on. The fact that estimates of gays and lesbians depend on which criteria are
being used, was already evident in Kinsey's famous report. In fact, Kinsey
himself argued strongly against the idea that people can easily be divided in
two classes, homosexuals and heterosexuals. In exploring male homosexuals
in the population, he gave several estimates ranging from 4% to 37%,
depending on whether the criterion was exclusively homosexual throughout a
life or at least some overt homosexual experience. About 10% were more or
less exclusively homosexuals for at least 3 years (Kinsey et al., 1948, pp. 650­
651). Notwithstanding, Kinsey's sampling procedure also had its weakness.
According to several authors, the strong belief in 10% homosexuals ex-
pressed by lay persons as well as among professionals, is a misinterpretation
of what Kinsey in fact meant and said (Lauman et al., 1994; Pedersen and
Kristiansen, 2003). Recent surveys based on representative samples, conclude
10% is an overestimation. The estimations from these surveys are that about
4­9% of men and about 4% of women had at least one same-sex partner
during their life (Spira et al., 1993; Lauman et al., 1994; Black et al., 2000).
Considering these results, a precise risk population for estimating rates for
same-sex couples is beyond reach. Neither can we take for granted that the
figures from one society will be valid in another society.
3.2. LESBIAN AND GAY COUPLES
Our study does not contain information on all same-sex couples living
together: only those who have taken a further step, and entered into a reg-
istered partnership are included. Given that registered partnerships and
same-sex marriages have only existed for a short time, characteristics of these
types of unions are still rather unexplored, and we have not seen studies that
look at the divorce risk in the way this article does.
In recent years, however, studies have been conducted of same-sex couples
living in consensual unions based on census data and large-scale surveys.
Considering our knowledge of the differences between opposite-sex couples
being married or cohabiting, it seems reasonable to assume that same-sex
couples living in consensual unions will not necessarily be representative of
those who have married or entered into a registered partnership. In the
TURID NOACK ET AL.94
absence of comparable data on same-sex marriages, a look at some demo-
graphic characteristics of persons living in same-sex cohabitation may,
however, be of interest.
Generally, studies of homosexual practice as well as homosexual identity
find that men are in the majority. With regard to same-sex couples, Canadian
Census data show that 55% were male couples compared to 51% in the USA
(Turcotte et al., 2003). In the Norwegian Census 2001, the proportion of
male couples was 49%.
US Census data indicate that having children living in the household is
fairly common for female couples, but rare for male couples; 22% and 5%
respectively (Black et al., 2000). In the Canadian Census 2001, 15% of the
lesbian couples and 3% of the gay couples were registered with children in
the household (Turcotte et al., 2003). To be previously married is fairly
common both for women and men living in same-sex consensual unions. Of
those living in same-sex consensual unions, 20% of gays and 30% of lesbians
were previously married. Most of the children living in households with
same-sex couples were probably born while the mother, or more seldom the
father, lived in an opposite-sex relationship (Black et al., 2000).
Same-sex couples are more inclined to live in metropolitan areas than in
other areas (Black et al., 2000; Turcotte et al., 2003). In Canada, the propor-
tion of same-sex couples was three times higher in metropolitan areas than
elsewhere in the country. In the USA as well as in Canada, the concentration in
the metropolitan areas seems to be stronger for male than for female couples.
The inclination to choose a partner with characteristics similar to oneself
is confirmed in several studies of opposite-sex couples. There is, in other
words, a positive assortative mating. Whether same-sex couples practice
positive assortative mating is questionable. Considering the smaller pool of
partners who are accessible for lesbians and gays, it has been supposed that
the partners will be less similar. However, considering some typical demo-
graphic variables, an overall level of positive assortative mating also seems to
be present for same-sex couples (Jepsen and Jepsen, 2002).
To our knowledge, studies of dissolution risks in same-sex consensual
unions are without exception based on small and non-representative samples.
These studies provide a valuable insight into couple dynamics and typical
problems in gay versus lesbian relationships (see for instance Kurdek, 2003).
It seems impossible, however, to draw conclusions from these studies on the
dissolution risk in same-sex consensual unions in general, and more specifi-
cally on the behaviour of female and male couples.
4. Data and methods
The present analysis is based on data from the Central Population Regis-
ter (CPR) in Norway. Every demographic event (birth, death, marriage,
A DEMOGRAPHIC ANALYSIS OF REGISTERED PARTNERSHIPS 95
registered partnership, divorce, change of address, etc.) is recorded in this
register. Based on some decades of use for a variety of statistical and ana-
lytical purposes, our evaluation of the CPR is that it covers both stock and
flow data reasonably well.
Our data comprise 1293 registered partnerships contracted from August
1993 to the end of 2001. At the end of 2001, 118 of these partnerships had
ended in divorce and 212 couples had obtained a legal separation licence.4
Forty partnerships had ended because one of the partners died, of which 85%
were male couples.
The data from the CPR provide information on the date for the start and
the possible cessation of each partnership, and the way in which the possible
cessation took place. The year of birth, sex, civil status, municipality, citi-
zenship and all children (dates of birth) are recorded for each partner.
The following variables have been defined for the analysis; Age, which is
the mean age of the two partners at the time they entered into the registered
partnership. Age difference, which is the age difference between the two
partners in the couple. Sex of the couple, which in this presentation will
sometimes be referred to as gays and lesbians rather than men and women.
Civil status, which since the registered partnership law came into force, also
includes registered partner. Municipality, defined as living in the capital city if
at least one of the two was settled there. Citizenship is used as a proxy
variable for national origin at the time of partnership formation. Children,
which includes all the children of each partner.
In addition, we have constructed a variable called pioneer, which means
that the partnership was contracted within the first 12 months of the law
coming into force.
For the analysis of the data, we have used Kaplan­Meier plots and pro-
portional hazards regression (Cox­regression) models. All the variables are
fixed at what they were at the start of the registered partnership. Time is
measured from entering into the partnership until divorce or censoring.
Censoring occurs if one of the partners dies, if the couple moves abroad or
else at January 1st
2002. The analysis was conducted using the S-PLUS
software package.
5. Registered partnership trends in Norway
Norway had more new partnerships shortly after the law was passed than in
the subsequent seven years. The same pattern was seen in Denmark and
Sweden. In recent years, there has been a slight increase in the number of
partnerships contracted (Figure 1). Compared to the total number of
inhabitants, registered partnerships seem to be slightly more widespread in
Denmark than in Norway and far less than in the Netherlands (Festy, 2001;
Waaldijk, 2001; Digoix et al., 2004).
TURID NOACK ET AL.96
The number of lesbian partners has increased evenly every year (Figure 1).
The high number of male couples in the year the law came into force was
followed by a drop the first 2 years but has later increased. This means that
the proportion of lesbian couples has increased from 27% to 43% in 2003.
The same development in the sex ratios has been observed in Denmark and
Sweden.
The couples we have defined as pioneers, i.e. those who entered into
partnerships within the first 12 months, deviate in different ways from the
subsequent partners (Table 1). With a mean age of 39.8 years, compared to
37.1 years of age for the other partners, the pioneers were significantly older.
It is reasonable to assume that many of the pioneers had waited for a long
time to legalise their relationship, which may have resulted in an older and
also accumulated group. The average age difference between the youngest
and the oldest partner is 7.0 years, with no significant difference between the
pioneer partners and those who entered into partnerships later.
In six out of ten partnerships entered into in the first year, one or both of
the partners were settled in the capital. This dominance has, however,
weakened since then (Table 1). The proportion of partners living in the other
large cities and their surroundings is, however, disproportionately large.
In 19% of the partnerships, one or both partners had previously been
married. In partnerships between one unmarried and one previously married
person, the previously married person is the older of the two in seven out of
ten cases. To be previously married was somewhat more common among the
pioneers than in the subsequent cohorts. Information on previous civil status
0
30
60
90
120
150
20032002200120001999199819971996199519941993
Male
Female
Number
Source: Population Statistics. Unpublished figures.
Figure 1. Registered partnerships contracted in Norway, 1993­2003.
A DEMOGRAPHIC ANALYSIS OF REGISTERED PARTNERSHIPS 97
is, however, missing for a fairly large share of the partners, most of them
foreigners.
Nearly one third of the registered partnerships are between a Norwegian
and a person with another citizenship. For marriages, with the exception of
those where both were foreigners, nearly one fifth were cross­national unions
during the same period. The proportion of cross national partnerships has
been relatively stable from year to year. The foreign partners are mostly from
the USA, Sweden and Thailand. These three countries are responsible for
about a quarter of the cross­national partnerships.
Compared to the cross-national partnerships, the Norwegian­Norwegian
couples are older and the age difference between the partners is smaller
(Table 2). Most remarkable is the difference in the proportion of lesbian
couples, 47% in the Norwegian­Norwegian partnerships and 21% in the
cross-national partnerships.
The Norwegian press has from time to time given the impression that there
is a baby boom among lesbian and gay couples. This assertion is not con-
firmed by our data. Until the end of 2001, only 24 persons (20 women and 4
men) have had `registered partner' as their civil status when they became
biological parents. This amounts to 0.9% of all who had entered into a
partnership.
Although the number of those who have become parents while living in a
registered partnership is small, many more of the partners had children born
before they entered into a partnership. In 8% of the registered partnerships,
one or both partners were parents at the start of the union. However, the
proportion who live with children is smaller. In some cases, the children are
Table 1. Pioneers compared with subsequent partnerships
Characteristics at the time of
entering into the union
Pioneers (August
1993­July 1994)
Partnerships
contracted
August 1994­
December 2001
All
Average age in years 39.8 37.1 37.6
Average age difference in the
partnerships in years
6.5 7.1 7.0
Percentage one or both living in the
capital
60 54 55
Percentage one or both previously
married
22 18 19
Percentage lesbian couples 27 41 38
Percentage cross-national
partnerships
28 33 32
TURID NOACK ET AL.98
grown up, in other cases they live with the other parent. In 2003, a total of 76
children (0­17 years) were living in partnership families.
6. The divorce risk in registered partnerships
6.1. UNIVARIATE ANALYSIS
Until now, we have looked at the characteristics of those who have chosen to
enter into a registered partnership. In the remainder of the paper we use Cox
regression to estimate what these characteristics might mean to the risk of
divorce. As in marriages, divorces in registered partnerships rarely occur
within the first two years after entering into partnership (Figure 2). Those
Table 2. Norwegian­Norwegian couples compared with cross-national couples
Characteristics at the time of entering
into the union
Norwegian­
Norwegian couples
Cross-national
couples
Average age in years 39.1 34.7
Average age difference in the partnerships
in years
5.9 9.2
Percentage one or both living in the capital 53 59*
Percentage lesbian couples 47 21
*13% unknown.
0 2 4 6 8
0.5
0.6
0.7
0.8
0.9
1.0
Proportion not divorced
The thin, light grey lines represent a 95% confidence interval.
1
Years since entering the partnership
Figure 2. Probability that the registered partnership still lasts by duration.
A DEMOGRAPHIC ANALYSIS OF REGISTERED PARTNERSHIPS 99
who decide to divorce so quickly will normally have moved away from each
other at the latest one year after the union was entered into. From the second
year, the amount of divorces steadily increases. Many of the couples have
most probably been living together for a long time before they entered into a
registered partnership, but we have no way of knowing the total duration of
the relationship.
The probability that the partnership has ended in a divorce is estimated to
be 8% after 4 years and 16% after 6 years. There is no sign of a typical year
of crisis.
There were large variations in divorce risks between the different groups of
partners. For one thing, it seemed that the pioneer couples, those who en-
tered into partnerships within the first 12 months of the law coming into
force, had a lower probability of divorce than couples from later years. High
age at the time of entering into a partnership, in this analysis measured by the
couple's average age, also gave more stable relationships compared to those
with a lower average age.
To explore the effect of citizenship at the start of the partnership, we
looked at three different groups. The first group consisted of partnerships
between two Norwegians or between a Norwegian and a Nordic citizen, the
second group was composed of partnerships between a Norwegian and a
citizen from Europe or North America5
, and the third group contained
partnerships between a Norwegian and a citizen of a third world country.
The partnerships in the first group were clearly more stable than in the two
non-Nordic groups. From Figure 3, it could be construed that there also was
0 2 4 6 8
0.5
0.6
0.7
0.8
0.9
1.0
The rest of
Europe and
USA
Proportion not divorced
Years since entering the partnership
Norway and the
Nordic countries
Third world
countries
Figure 3. Probability that the registered partnership still lasts by duration and citizen-
ship.
TURID NOACK ET AL.100
a difference between the two non Nordic-groups, but this difference did not
turn out to be significant in our analysis.
The difference between the groups only becomes clear after the partner-
ship has lasted four years. After such a long time, partners with foreign
citizenship will normally be guaranteed residency in Norway, even if they
divorce. As with marriages, it could be questioned whether such cross-na-
tional partnerships give room for pro forma partnerships, i.e. partnerships
contracted to ensure the other party residency in Norway. Our analysis gives
no evidence of this. If that was the case, we would have expected a very clear
increase in the number of divorces just after four years, but there is no such
sign (Figure 3).
Gay and lesbian relationships are often described as quite different in
nature. For this reason, we decided to look at how the divorce risk depended
on the sex of the couple. As already discussed, there were very few divorces
during the first years, and the divorce risk was almost the same for female
and male partnerships. However, after 3 years, there was an increasing dif-
ference (Figure 4). Six years after a partnership was entered into, the prob-
ability of divorce among the lesbian couples was estimated to be about 21%,
in contrast to the 13% among gay couples.
A comparison between partnerships from the capital and partnership
couples from other parts of the country showed no significant difference.6
However, the groups were too small to analyse this phenomenon properly.
0 2 4 6 8
0.5
0.6
0.7
0.8
0.9
1.0
Lesbian couples
Proportion not divorced
Years since entering the partnership
Gay couples
Figure 4. Probability that the registered partnership still lasts by duration and sex of the
couple.
A DEMOGRAPHIC ANALYSIS OF REGISTERED PARTNERSHIPS 101
The age difference between the youngest and oldest partner showed no sig-
nificant influence on the divorce risk.
6.2. MULTIVARIATE ANALYSIS
Finally, we wanted to see if the effect of the different variables changed when
we included all of them in the same model. We have carried out two analyses.
The first analysis contains all partnerships registered between 1993 and 2001.
Male couples where both partners were Nordic citizens at the start of the
partnership are used as a reference group (Table 3). The other is an analysis
where only the couples consisting of two Norwegian citizens at the time of
the union formation are included (Table 4). In this analysis, the reference
group is a male couple.
In our first analysis, of all partnerships, we found that controlled for other
variables,
average age, sex, citizenship and average age difference significantly
influenced the divorce risk (Table 3).7
We found that the lesbian partners'
risk was estimated to be 1.58 times the risk of the gay partners.
The difference between gay and lesbian partners is even higher among the
Norwegian­Norwegian partners (Table 4). The lesbian partners' risk of di-
vorce was estimated to be 2.17 times the risk of the gay partners in the
Norwegian­Norwegian case. Here we find no significant effect of average age
difference, but average age still seems to be important.
Table 3. Relative risk of divorce. All partnerships included in the analysis
Variable Effect 95% confidence
interval
P-value
Average age (on 10 years scale) 0.36 (0.26­0.49) <0.0001
Women 1.58 (1.28­1.94) <0.0001
Europe and USA 1.22 (0.93­159) 0.16
Third world countries 1.24 (1.03­1.49) 0.03
Unknown country 1.03 (0.86­1.24) 0.72
Average age difference (on 10 years scale) 1.04 (1.01­1.08) 0.004
Non-significant variables*:
Not from the capital 0.99 (0.82­1.20) 0.93
Municipality unknown 0.36 (0.0­35.5) 0.66
Non-pioneers 1.02 (0.79­1.20) 0.83
Earlier married 1.27 (0.99­1.63) 0.06
Unknown/Other 1.06 (0.88­1.28) 0.53
Young children 1.13 (0.83­1.55) 0.45
Old children 1.14 (0.87­1.49) 0.33
Unknown 0.47 (0.32­0.68) <0.001
*Each variable added to a model consisting of the significant variables.
TURID NOACK ET AL.102
Among the Norwegian couples, there are almost 47% lesbian couples,
whereas there are only 21% among the cross-national partnerships. The
mean age difference of the Norwegian couples was only 6 years, compared to
9 years for the cross-national group. This might contribute to the fact that
average age difference is not a significant variable in the second analysis.
Disproportionately many male couples and a high average age seem to
have been decisive factors of the pioneer couples' lower rate of divorce.
Controlling for these factors, we no longer find an independent, significant
effect of being a pioneer.
In both cases, with and without the international partnerships, we tried to
include earlier marriage history as an explanatory variable. For both types of
analyses, there seemed to be an effect of being previously married, in the
sense that it raised the probability of a divorce. However, the data is too
limited to give significant results for this variable.
Finally, we have explored the effect of one or both partners having chil-
dren when they entered into the partnership. However, the fact that the
group of partners with children was small and the lack of information on
children among the partners from other countries made it impossible to
properly test the effect of this variable.
7. Summary and discussion
As a new family type and a new demographic event, the emergence of reg-
istered partnership fits neatly into the theories on individualisation and
de-standardisation in family life (see for instance Corijn and Klijzing, 2001).
Cultural changes and a decrease in normative controls of behaviour seem to
Table 4. Relative risk of divorce. Norwegian­Norwegian partnerships
Variable Effect 95%
confidence
interval
P-value
Average age ( on 10 years scale) 0.33 (0.22­0.49) <0.0001
Women 2.17 (1.31­3.60) 0.003
Non-significant variables*:
Not from the capital 1.18 (0.52­1.40) 0.52
Non-pioneers 0.89 (0.96­1.08) 0.70
Average age difference (on 10 years scale) 1.02 (0.96­1.08) 0.50
Earlier married 1.75 (0.98­3.12) 0.06
Unknown/other 0.83 (0.11­6.05) 0.86
Young children 1.42 (0.64­3.13) 0.39
Old children 1.60 (0.54­4.72) 0.39
*Each variable added to a model consisting of the significant variables.
A DEMOGRAPHIC ANALYSIS OF REGISTERED PARTNERSHIPS 103
have given individuals greater freedom, but also a stronger obligation to
independently decide how to organise their lives (Beck and Beck-Gernsheim,
1995). In a parallel development, we have witnessed an increasing separation
between reproduction and sexuality. Thus, the disparity between same-sex
and opposite-sex relationships is diminished (Giddens, 1992). Some people
may find it ironic that registered partnerships emerge in an era where mar-
riage is losing ground. But it may also be seen as a logical consequence, as the
deviant and discriminated group, the homosexuals, were only given the
privilege to marry when marriage as a social institution had lost much of its
importance in society (Moxnes, 1993).
A decade after the Norwegian law on registered partnerships was passed,
the figures for 2003 showed 27 male and 21 female partnerships contracted
per million inhabitants. This is less than one per cent of all marriages and
registered partnerships contracted that year. The Norwegian figures are small
compared to most other countries that have legalised same-sex marriages. An
overview of homosexual practice in Western Europe shows, however, little
difference between countries (Festy, 2001). This may indicate that the main
reasons for the differences rather are to be found in external circumstances,
such as less openness, or a variation in privileges and rights for married
versus cohabiting couples.
In the last decade, many laws and regulations have been changed in order
to treat married and cohabiting people more equally. Although same-sex
cohabitants in many situations will be treated equally with married people
and those living in a registered partnership, the equalisation has generally not
gone so far as for opposite-sex cohabitants having children together. That
means, that the pluses and minuses in being a registered partner rather than a
mere cohabitant are more relevant than for many heterosexual couples
choosing between cohabitation and marriage. If they enter into a registered
partnership, some of their welfare benefits will be reduced, for instance, be-
cause they are regarded as a couple rather than two separate persons. In
other circumstances, they may benefit from living in a registered partnership.
Which alternative is most advantageous will depend on the situation at any
given time and the degree of risk the person is willing to take. Altogether, a
registered partnership may often emerge as a safer alternative in many ways.
We found that in about one fifth of the partnerships, one or both of the
partners have previously been married. Our data do not, however, establish
whether the change occurred because sexual identity had been suppressed or
whether it changed.
Not unexpectedly, our study showed that cross-national unions are very
common; nearly one third of the registered partnerships comprised a
Norwegian and a foreigner. For most of these Norwegian-foreigner couples,
Norway must have been the only alternative of entering into their partner-
ship since most foreigners who enter into partnerships with Norwegians come
TURID NOACK ET AL.104
from countries where same-sex marriages have not been legalised. As with
marriages, it could be questioned whether such cross-national partnerships
give room for pro forma partnerships. If we look at the divorce patterns after
4 years, which for foreigners is the minimum duration of marriage to be
assured legal residency in the country after a divorce, we do not see any short
term change around 4 years.
The admission to medically assisted insemination, joint adoption and the
well-being of children in same-sex marriages and registered partnerships are
topics that have been highly debated in Norway, as in many other countries.
We found that many of those entering into registered partnerships were
parents, 24% and 13% of the women and the men, respectively. However,
much lower proportions were living with children in the household. In 2003,
about 0.85 per 10,000 of children under age 18 lived with mothers or fathers
who were registered partners. Our estimations also indicate that few of these
children have been born to mothers or fathers with `registered partner' as
their present civil status. Less than one per cent of the partners have become
parents while living in a registered partnership. The baby boom among gays
and lesbians that has sometimes been depicted in the media thus seems to be
closer to a myth than fact.8
Taking into account that most of the partnerships have so far had a
relatively short duration, it seems reasonable to assume that a larger share of
the partners will become parents in the future. In a survey of the living
conditions of gays and lesbians, about one third of the young childless said
that they wanted children (Hegna et al., 1999). The proportion wanting to
become parents, however, seems to be considerably smaller than in the
population in general (Noack and stby, 1996). Analyses of fertility inten-
tions have also shown that these measures are far from accurate estimates of
subsequent fertility, at least at the individual level (Noack and stby, 2002).
When examining the registered partnerships entered into in Norway from
1993 to 2001 we estimated the probability of divorce to be 16% after six
years. However, the divorce pattern varies greatly between subgroups.
We found that lesbian partnerships had a 2.17 times higher risk of ending
in divorce than gay partnerships. Comparisons between gay and lesbian
relationships have been made and discussed in several studies. The idea that
same-sex couples become marriage-like, in the sense that one of the partners
takes on the traditional female role while the other partner takes on the male
role, seems to be more or less outdated (Patterson, 2000). There seems to be
general agreement that there are important differences between lesbian and
gay couples but there is still a dispute about what these differences are.
Earlier research gives no clear answer to the question of which are the most
stable, the gay or the lesbian relationships.
While many previous studies on gays and lesbians are based on small and
not very representative samples, our project avoids this problem since it
A DEMOGRAPHIC ANALYSIS OF REGISTERED PARTNERSHIPS 105
includes the whole population of registered partners, and not just a sample.
When we conclude that gay partners stick together more than lesbian
partners, we are therefore on quite safe ground. However, these findings will
not necessarily hold true for lesbian and gay relationships in general. To be
able to comment on that we would need to know more about those who
choose to enter into a registered partnership, and those who choose to
cohabit, or not to cohabit at all.
Another high risk group we found in our study was the group of part-
nerships between a Norwegian and a non-Nordic person. The estimated
divorce risk was 1.2 times higher than in partnerships between two Norwe-
gians or between one Norwegian and one from another Nordic country.
Cultural differences are not necessarily the only reason why the non-Nordic
partnerships end in divorce more often. The fact that one of the partners
might not have a residence permit in Norway could mean that these couples
have less time and opportunity to test out their relationship before they enter
into the partnership compared to couples where both partners are already
living in the country. As long as the possibility of entering into a registered
partnership is limited to a small number of countries, most foreign partners
will be compelled to have residence in Norway, at least as long as they want
to live in a country where their partnership is valid as a way of living to-
gether. Unlike married couples in the same situation, they do not have the
same freedom to consider where to settle.
We also found that high age difference was another risk factor for divorce.
Entering into a registered partnership at a young age also had a negative
effect. Such effects are also well known in opposite-sex divorce studies.
Those who choose to enter into a registered partnership are not necessarily
representative of all couples in same-sex relationships. The behaviour of
partners in a registered partnership is not necessarily similar to the behaviour
of gays and lesbians living in a consensual union or those who are only
lovers.
We have concluded that there are some quite considerable differences in
the risk of divorce between different groups of partners. To do further
analyses on what causes these differences, one would need other types of data
in addition to what the registers can tell us. A demographic analysis can,
however, be a starting point for formulating new hypotheses and carrying out
more detailed analyses.
The maximal duration of time for the partnerships in our study is eight
and a half years. It is not obvious whether the partners will behave in the
same way that we have described, in the years to come. From studies on
marriages, we know that the risk of divorce strongly decreases with time, and
as the couple becomes older.
In the analysis, we have used the term `pioneer couples' to refer to the
partnership couples who entered into their registered partnership within the
TURID NOACK ET AL.106
first 12 months of the law coming into force. In a sense, it could be claimed
that we have only been dealing with pioneer couples, i.e. partnerships entered
into in the early days of the law. Many factors could change as more people
live in registered partnerships.
In a follow-up study, we compare the experiences with registered part-
nerships both in Norway and Sweden (Andersson et al., 2004). So far, the
similarities seem to be greater than the differences.
Notes
1
Registered partnerships became possible in Denmark in 1989, followed by Norway (1993),
Sweden (1995), Iceland (1996) and Finland (2002). The Netherlands introduced registered part-
nerships in 1998 and same-sex marriages in 2001. Belgium legalized same sex marriages in 2003.
2
An updated overview is provided by Waaldijk (2003). The web site of the International
Lesbian and Gay Association (ILGA) also is considered a reliable source for information
(http://www.ilga.org/index.asp).
3
In Norway, only stepchild adoption is allowed for couples living in registered partnerships.
This means the right to adopt a partner's child. Registered partners are not entitled to adopt
jointly as married couples are. Since 2003, registered couples in Sweden have been entitled to
adopt jointly. The traditional ``donor'' countries seem, however, to be restrictive to let same
sex couples adopt children from their country.
4
As the main rule, partners have to register as being legally separated for a minimum period
of one year before being granted a divorce.
5
The actual countries are the non-Nordic countries of Europe, USA, Canada and Australia.
6
Couples where at least one of the partners was living in the capital at the time of entering
into the partnership.
7
Controlled variables were average age, average age difference, citizenship, earlier marriages,
municipality and parenthood at the start of the partnership.
8
We do not, however, have any information on parenting in lesbian and gay cohabitations.
Assuming that the proportion of children born in cohabitation is approximately the same as
for heterosexuals, the number of children born in lesbian or gay unions (partnerships or
cohabitations) may be more than doubled, but a figure still far from what we associate with
the term ``baby boom''.
References
Andersson, G., Noack, T., Seierstad, A. and Weedon-Fekjr, H., 2004. The demographics of
same-sex ``marriages'' in Norway and Sweden', Max Planck Institute for Demographic
Research, Rostock, (MPIDR Working Paper WP 2004-018).
Badgett, L., 1995. `The wage effects on sexual orientation discrimination', Industrial and Labor
Relations Review 48: 726­739.
Badgett, L., 2001. Money, myths, and change. The Economic Lives of Lesbian and Gay Men.
Chicago: University Press of Chicago.
Black, D., Gates, G., Sanders, S. and Taylor, L., 2000. `Demographics of the gay and lesbian
population in the United States: Evidence from available systematic data sources',
Demography 37: 139­154.
Black, D., Gates, G., Sanders, S. and Taylor, L., 2002. `Why do gay men live in San Fran-
cisco?', Journal of Urban Economics 51: 54­76.
A DEMOGRAPHIC ANALYSIS OF REGISTERED PARTNERSHIPS 107
Beck, U. and Beck-Gernsheim, E., 1995. The Normal Chaos of Love. Cambridge: Polity Press.
Corijn, M. and Klijzing, E., 2001. Transition to Adulthood in Europe. Dordrecht: Kluwer
Academic Publishers.
Digoix, M., Festy, P. and Waaldijk, K., 2004. Same-sex couples and heteronormativity. Paper
presented at the PAA 2004 Annual Meeting, Boston, April 1­3, 2004.
Festy, P., 2001. `The ``Civil Solidarity Pact'' (PACS) in France: An impossible evaluation',
Population & Sociétés. June No. 369: 1­4. Institut national d'é tudes dé mographiques,
Paris.
Giddens, A., 1992. The Transformation of Intimacy. Oxford: Polity Press.
Halvorsen, R., 1998. `The ambiguity of lesbian and gay marriages: Change and continuity in
the symbolic order', in J. Lo¨ fstro¨ m (ed), Scandinavian Homosexualities: Essays on Gay and
Lesbian Studies. New York: Harrington Park Press, 207­231.
Halvorsen, R., 1999. `Partnerskapslovens tvetydighet', Tidsskrift for samfunnsforskning 3:
310­331.
Hegna, K., Kristiansen, H.W. and Moseng, B.U., 1999. Leveka° r og livskvalitet blant lesbiske
kvinner og homofile menn. Rapport 1/99, Norsk institutt for forskning om oppvekst,
velferd og aldring, Oslo.
Jepsen, L. and Jepsen, C., 2002. `An empirical analysis of the matching patterns of same-sex
and opposite-sex couples', Demography 39: 435­453.
Kinsey, A. C., Pomeroy, W.B. and Martin, C.E., 1948. Sexual Behaviour in the Human Male.
Philadelphia/London: W.B. Saunders.
Kurdek, L. A., 2003. `Differences between gay and lesbian cohabiting couples', Journal of
Social and Personal Relationships 20: 411­436.
Lauman, E., Gagnon, J. M., Michael, R. T. and Michals, S., 1994. The Social Organization of
Sexuality: Sexual Practice in the United States. Chicago: University of Chicago Press.
Martin C. and Thé ry I. 2001. `The PACS and Marriage and Cohabitation in France', Inter-
national Journal of Law, Policy and the Family 15: 135­158.
Moxnes, K., 1993. `Partnerskapsloven - et uttrykk for parforholdets endrede karakter',
Lvetann 3: 30­31.
Noack, T. and stby, L., 1996. Fertility and Family Surveys in Countries of the ECE Region.
Standard Country Report Norway. Economic Studies No 10a, United Nations Economic
Commission for Europe. Geneva/New York: United Nations.
Noack, T. and stby, L., 2002. `Free to choose ­ but unable to stick with it?', in E. Klijzing
and M. Corijn (eds), Fertility and partnership in Europe: Findings and Lessons from
Comparative Research. Vol. II. pp. 103­116 Geneva/New York: United Nations.
Patterson, C., 2000. `Family relationships of lesbian and gay men', Journal of Marriage and the
Family 62: 1052­1069.
Pedersen, W. and Kristiansen, H. W., 2003. `A° gjre det, a° fle det og a° vre det.
Homoseksualitet i det seinmoderne', Tidsskrift for samfunnsforskning 44: 3­37.
Sland, B., 1998. `A queer nation? The passage of the gay and lesbian partnership legislation
in Denmark, 1989', Social Politics 5, Spring: 48­69.
Spira, A., Leridon, H. and Gray, R., 1993. Biomedical and Demographic Determinants of
Reproduction. Oxford: Clarendon.
Strksen, I., 2000. Homofili og politikk. En komparativ analyse av verdiendringer og men-
ingskoalisjoner bak vedtaket av partnerskapsloven i Danmark og Norge. Hovedoppgave
ved Institutt for sammenlignende politikk. Universitetet i Bergen.
Turcotte, P., Renaud, V. and Cunningham, R., 2003. Same-sex relationships and sexual ori-
entation in Canada: data, concepts, and methodological issues. Paper presented at the
2003 PAA annual meeting Minneapolis, May 2003.
TURID NOACK ET AL.108
Waaldijk, K., 2001. Small change: How the road to same-sex marriage got paved in the
Netherlands, in R. Wintermute and M. Andens (eds), Legal recognition of same-sex
partnerships: A study of national, European and international law. pp. 437­446, Oxford:
Hart publishing.
Waaldijk, K., 2003. Major legal consequences and procedures of civil marriage, registered
partnership and informal cohabitation for different and same-sex partners in nine Euro-
pean countries. Work in progress, version 17 September 2003. Leiden: E-M. Meijers
Institute of Legal Studies, Universiteit Leiden.
Weeks, J., 2000. `The challenge of lesbian and gay studies' in T. Sandfort, J. Schuyf, J. W.
Duyvendak and J. Weeks (eds), Lesbian and Gay Studies. An Introductory, Interdisciplinary
Approach. pp. 1­13. London: Sage Publications.
A DEMOGRAPHIC ANALYSIS OF REGISTERED PARTNERSHIPS 109