Newsletter issues
Combined index

Household and Family Trends in Europe: from Convergence to Divergence

The International Year of the Family 1994 together with the World Population Conference have drawn greater attention to the changing family. On the level of the European Union interest in the family was intensified, as the first steps taken towards the creation of a European household and family statistics and the strengthening of the European Observatory on National Family Policies (EONFP) show. In addition, other international organisations, such as the Council of Europe or the United Nations Economic Commission for Europe (ECE), have intensified their activities in the field of population and family research. In this context the question may be posed how household and family structures in Europe evolve. Not only for the European Union but also for sociological research in general the question of divergence or convergence of national household and family structures and of national demographic developments is of great importance.

Growing convergence of the "old" diminishing household and family forms

Until the 1960s a growing similarity of patterns in household and family structures could be observed in the industrialised countries of Western Europe. Since that time some family patterns have been showing a tendency to diverge in the European countries. The aim of this contribution is to verify empirically the thesis of the succession of convergence and divergence of family and household forms.

Diminution of households not yet finished

Today the concepts of household and family are virtually identical. One basic feature of the development of households is the diminution of households. From a historical perspective the diminution of households is a result of the nuclearisation (the disappearance of persons not belonging to the nuclear family, such as servants, farm labourers and relatives), the fertility decline and the solitarisation, i.e. the growing tendency to live alone. Another example for the process of the diminution of the household is the decreasing mean household size. For most European countries a household included between 4 and 6 persons around 1850 and about 2 to 3.4 persons in 1990. In addition, this process is highlighted by the decline of "bigger" households. Thus over half of all households in all European countries had more than four members around 1850, while this figure ranged between 20% to 50% in 1990. This ubiquitous trend towards small living units becomes even more evident if one looks at the development of households with five and more persons. Around 1850 the shares in Europe ranged from 30% to 70%, in 1990 from 5% to 15% in the majority of all west European countries (Graph 1).

Continuing nuclearisation

Nuclearisation in the sense of dissolution of family and household forms which transcend the nuclear family is proceeding in Europe. In 1990, the shares of extended family households in relation to all family households varied from about 2% in West Germany to about 15% in Greece. But meanwhile a lot of countries have reached the lower limit. Thus the share of extended family households in West Germany has stagnated since 1976 at roughly 2%; in Belgium and Hungary as well a further decline in the decade from 1980 to 1990 could not be observed. Switzerland and Italy still show significant decreases.

This continuing nuclearisation can also observed regarding the development of households with three and more generations. But this way of presenting data according to generations is only possible for few countries. In the western industrialised countries such as Germany, the United States and Canada the shares of households with three and more generations are meanwhile (1990) below 3 percentage points. As late as 1957 the share in West Germany was three times higher (9%). With 16%, Albania on the other hand shows a still predominantly preindustrial pattern of the generation structure.

Thus the model of the nuclear family has become the dominant family type in all industrialised European countries. This is in contrast to, for example, the industrialised or industrialising nations of Asia such as Japan and Korea (but China, too) where the share of households with three and more generations varies from 15% to 20%, but shows a strong tendency to decline.

Another sign for a trend towards the nuclear family is the growing share of married people in percent of the total population, a tendency that lasted for several decades. This trend however only continued until the 1960s; since that time a reversal of the trend could be observed: the share of married persons in the total population is declining again. This reversal of the trend is only partly a new development away from the nuclear family. Socio-structural and demographic reasons, such as. the postponement of the age at first marriage due to a higher participation of women in education and employment and an increase in the number of divorces, are also responsible for this development. The invention and diffusion of "new" forms of living could be further reasons for the decreasing married population.

Growing divergence of "new" increasing household and family forms

Since the 1960s some new developments in the field of living forms have accelerated, although these forms pre-existed in historical terms, but never became dominant phenomena. These new developments have not threatened the model of the nuclear family in its substance, but one could speculate if they could substitute the nuclear family in the distant future.

Growing solitarisation

One of the decisive new developments is the very strong increase in the number of people living alone, defined as one-person households by household statistics. In relation to the total number of private households this share varied from 5% to 27% in Europe; in 1990 the share of the same countries ranged from 14% to 40%. Albania and Turkey have not yet participated in this development. The gulf between these countries has thus become substantially bigger.

If the number of one-person-households is put in relation to the total population, this share varied between 1.5 % and 10% in 1960; in 1990 this share ranged between approximately 4% and 18%. This means that the divergence is growing.

Quantitative decline of the significance of the married couple in favour of the one-parent family Measured by all households with a family nucleus, that is the family households so to speak, the quantitative importance of the married couple with and without children has been declining in many European countries since the 1960s. This development is a reflection of the growth of one-parent families. In most European countries the overwhelming majority of all family nuclei still consists of a married couple. In 1980 their share ranged from about 85% to about 95% in the majority of countries. Until 1990 the differences between countries grew, accompanied by a general decrease of this share. Although the decrease, for example, of the share of married couples of all households with family nuclei was very moderate, the decreases witnessed in Hungary, the former German Democratic Republic (GDR) and Belgium were significant.

The relative quantitative decrease in the number of married couples is the result of a growth in the number of one-parent families. This increase can be observed in all European countries, in fact with a growing divergence. The growth rates of the 80s are significantly higher than those of the 60s and 70s. In 1990, the share of incomplete family households varied from about 5% to 15%. The overwhelming majority of all one-parent families in all European countries is formed by mothers. Their share varies from about 75% to 90%. The development of this share does not reveal a clear trend. But there are signs in some non-European countries, e.g. the United States, for an increase in the share of men.

The reasons for this shift in familial forms of living from the married couple to the one-parent family differ significantly from country to country; however, the majority is probably a result of an increase in divorce and the rising social acceptance of births out-of-wedlock and the social position of the lone parent. The traditional reason, widowed women with young children, probably plays an increasingly minor role.

Cohabiting couples on the increase

Since the end of the 1970s a growing tendency towards cohabitation, i.e. persons of different sex living together, has been observed. According to Kathleen Kiernan and Valerie Estaugh (1993) two types of cohabitation can be distinguished. One form is "nubile cohabitation", whereby young people in their twenties or early thirties live together as a preliminary stage or as a substitute for marriage. Parallel to the growing number of divorces the type of "post-marital cohabitation" has become increasingly important, partly as a substitute for remarriage, partly as the first stage. Comparative studies have shown that in all countries for which data are available there has been a growing trend since the early 1980s. Of young people below the age of 25 living together, more and more do not marry but live as a cohabiting couple. In the higher age brackets where marriages start occurring the share of women living in cohabitation is getting smaller. However, a rising trend can be observed here as well. Problems of comparability arise with data on cohabitation because they overwhelmingly stem form surveys which are very different as regards the number of respondents, concepts, assessed variables and time of investigation. If one wants to compare these surveys printed analyses and relative figures are often the only accessible data. Among European countries there exists a significant variation in the incidence of cohabitation. Those countries for which data are available can be subdivided into three groups: countries where cohabitation has established itself, countries where it is developing as an important living arrangement, and those countries where cohabitation has remained unknown until now (or was not yet detected).

(1) Countries where cohabitation established itself as socially accepted behaviour are Sweden, Denmark and Iceland. In the other Scandinavian or Nordic countries Norway and Finland the number of cohabiting couples has reached a high level. Even before 1960 living together without being married was not impossible in the Nordic countries (as compared to the Mediterranean countries, e.g.). Thus good preconditions for this behavioural pattern already existed. The majority of children today are no longer born within marriage in these countries, but in cohabitation.

(2) The second group of countries where cohabitation slowly emerges as a form of living arrangement includes Austria, Finland, Norway, Great Britain, France, the Netherlands, Switzerland and Germany. Within this group two subgroups can be distinguished. The rate of children born out-of-wedlock is high in (a) Austria, Finland, France, Great Britain and Norway, but rather low in (b) the Netherlands, Germany and Switzerland. These observations suggest the following interpretation: in Switzerland, the Netherlands and Germany cohabitation means that childless partners live together and marry when children arrive. This is not true for Austria, Finland, Germany, Great Britain and Norway; in any case, births of children occur much more frequently within cohabitation, although the cohabiting couples often marry afterwards (see also Haskey 1992).

(3) The third group with no or undetected cohabitation is made up mainly by the Mediterranean countries, probably Ireland and some east and south-east European countries. Growing divergence of out-of-wedlock births

In some Nordic countries the number of children born outside marriage has risen tremendously. The share of births out-of-wedlock (in % of all live births) is particularly high in Iceland, Sweden, Denmark and Norway, but it is also remarkable in France, Finland, Great Britain and Austria (Graph 2). On the other hand, the share of children born out-of-wedlock in Mediterranean countries and Germany, Switzerland, the Netherlands and Luxembourg is rather small if south, east and southeast European countries are excluded whose out-of-wedlock birth rate is traditionally very low. The low shares of these Central European countries may partly be attributed to institutional factors. For example, family law in Germany grants fewer rights to the father than to the mother if a child is born out-of-wedlock. Therefore it probably promotes marriage in the case of pregnancy. In the same way, the German personal tax system promotes marriage by suggesting the so-called tax splitting for married couples which ceteris paribus reduces the tax load. In the Nordic countries most of the children born outside marriage are born within cohabitation and not, for example, to a "single" mother.

Number of families without children rising

Since the 1960s, an increase in the number of childless married couples has been observed in many European countries. This increase obviously has different growth rates, indicating a growing divergence between countries. In 1990 the share of married couples without children in percent of all one-family households increased from about 20% in Ireland to about 45% in Switzerland. In Belgium and Italy apparently the shares did not change.

The continuing fertility decline, the postponement of the age at first marriage and the age at the birth of the first child could be reasons for this increase; furthermore, the share of married couples remaining without children for a lifetime probably increased.

Families with children by number of children

Different family types by number of children shall be depicted in a cross-sectional analysis. Family nuclei with children may be subdivided into three main types, namely married couples with children, fathers with children and mothers with children (Graph 3). In the twelve countries of the European Union about 80% of all family nuclei, on the average, were married couples with children in 1991. High shares of married couples with children are to be found in Greece (89%), Portugal (87%) and Italy (86%). The lowest share of married couples with children is to be found in the United Kingdom and Denmark (78% each) and in Belgium (79%). The number of fathers with children amounts to less than 5% in all twelve EU countries. The lowest shares are to be found in the Mediterranean countries Portugal (1.8%), Greece and Italy (each 2.2%). Marked differences can therefore be observed in EU countries regarding mothers with children. Their share is lowest in the Mediterranean countries and highest in Denmark, Belgium and the United Kingdom.

The number of children reveals noticeable differences between the European countries (Graph 4). The presentation according to households can only roughly be interpreted as being differences in fertility, because only children present in the household are counted and not the total number of births. The system of having one child only seems to be most widespread in Germany, Portugal, Luxembourg and Belgium. Families with many children - here families with three and more children - are very frequently found in Ireland, a country that plays a special role within the EU. France, the Netherlands and Belgium also lie above the average.

Single fathers show a relatively high variability - probably because of their low occurrence, but here fathers with one child are dominant. The reasons for being a single father are probably very different from country to country (being divorced, widowed).

About two thirds of all single mothers have one child and about one in five single mothers has two children (Graph 5). In Ireland, the number of children, as is the number of lone parents with many children, is far above the average (EUROSTAT 1994; see also Duchêne/Eggerickx 1994).

The North-South gradient in Europe

The question of territorial patterns in household and family structures cannot be resolved easily because the grouping of countries partly differs a great deal from indicator to indicator. However, a rough schematisation reveals the picture of a North-South gradient. If one uses the mean size of the households as the simplest indicator, Germany, the Scandinavian countries and the United Kingdom could be grouped together as countries with very small households. The Mediterranean countries Spain, Portugal and Greece, but also Ireland - apparently the European peripheries - belong to the cluster of countries with the largest households. Ireland is therefore the big exception in the North-South gradient. The majority of the continental countries belong to those lying between the extremes. The Mediterranean country Italy rather belongs to the continental group than to the peripheral group.

If the countries of eastern Europe and southeastern Europe are added, the picture becomes much more complex. The former COMECON countries - Yugoslavia included - show completely different household and family structures, and have never formed a unity in this respect. Thus the former German Democratic Republic belongs to the Nordic group, Poland, Yugoslavia and Romania to the southern European group, and Hungary and Bulgaria to the continental group.

In principle these territorial patterns also appear with respect to other indicators of the household and family structures, such as extended families, one-person households, one-parent families, cohabitation and births out-of-wedlock. Clustering, i.e. using a plurality of indicators, could be a valuable future research task.

Some covariations

In the following section some interrelations between the household and family structures and the demographic and socio-structural covariates will be presented. One example for the analysis of a covariation of traditional household and family structures is the analysis of the connection between the share of extended family households with another variable. If one postulates the hypothesis that the share of extended family households in a country depends heavily on the share of the agricultural sector, as this form of household is to be found mainly in family farms, the result would be what is depicted in Graph 6. Thus a rather clear positive relation between both variables emerges for those EU countries for which data are available: the higher the share of employment in the agricultural sector, the higher the share of extended family households. Furthermore it becomes evident that there are two main country clusters: on the one hand the highly industrialised countries of the continent and northern Europe, on the other hand the Mediterranean countries including Ireland with a still large agricultural sector.

In a second example the covariation of a "new" family type - namely the share of one-parent families - with one further variable is analysed. In this case it would be plausible to say that the number of one-parent families is decisively influenced by the number of divorces. According to Graph 7 a positive relation between both variables exists although this relation is not totally clear. The group of the Mediterranean countries (for Spain and Portugal no data are available; Ireland does not have a divorce law as yet; regarding Italy it is only divorces, but not legal separations - which are much more frequent - that are included) again forms a separate cluster with very low divorce rates. The reasons for the formation of one-parent families are much more of a traditional nature (widowhood) in these countries.

In a further example the relation between the mean household size and the overall fertility rate is analysed (Graph 8). In this case mainly two country clusters exist. On the one hand the highly industrialised countries of the continent and northern Europe, on the other hand the four Mediterranean countries Greece, Italy, Portugal and Spain. Ireland and Sweden are outliers in this respect. The high fertility rate found in Sweden is exceptional compared to the other Nordic countries, too. The fertility rate in the former GDR stood at 1.52 before the reunification in 1990 and fell to a low level because of the "demographic revolution" caused by economic uncertainty. The very low level of fertility observed in the Mediterranean countries in combination with the relatively large households suggests that in these countries household size is not only determined by the number of children - this means by the nuclear family; instead - as we have seen above - other household members, such as grandparents, often co-reside in the household in these countries.


Main sources for this contribution are parts on household and family statistics taken from the population censuses of the European countries. The large number of titles cannot be cited here. The demographic data mainly stem from Council of Europe 1994. Data on employment in agriculture are taken from EUROSTAT 1993.


Council of Europe 1994: Recent Demographic Developments in Europe, 1994. Strasbourg: Council of Europe Press (annual publication).

Duchêne, Josiane; Eggerickx, Thierry 1994: Haushalte und Familien in den Ländern der Europäischen Gemeinschaft 1990/91. Soziales Europa. Die Europäische Union und die Familie. No. 1, 1994, 28-37.

EUROSTAT 1993: Employment and Unemployment. Aggregates, 1980-1991. Luxembourg: Office des publications officielles des Communautés européennes (Theme 3, Series C).

EUROSTAT 1994: Households and Families in the European Union. Rapid reports. Population and social conditions. Luxembourg (François Begeot).

Haskey, John 1992: Patterns of marriage, divorce, and cohabitation in the different countries of Europe. Population Trends no. 69, 27-36.

Kiernan, Kathleen E.; Valerie Estaugh 1993: Cohabitation. Extra-marital Childbearing and Social Policy. London: Family Policy Studies Centre (FPSC) (=Occasional Paper 17).

Dr. Franz Rothenbacher
University of Mannheim, MZES, D-68131 Mannheim
Phone: (+49) (0)621 - 292 1738

EURODATA Newsletter No.1, p.3-9