Routes and determinants of leaving home: the city of Gothenburg, 1915–1943

Abstract

This paper deals with the home-leaving of young adults in Gothenburg, Sweden, in the period 1915–1943. We used individual-level panel data from population registers and poll-tax records, taking a competing risk design for the analysis of the determinants of leaving home to marry, or for non-familial living. We found a transitional, marriage-driven pattern of leaving home that neither fits the old context of life cycle service, nor the alternative modern routes out of the parental home into unmarried householdship. Young adults typically stayed at home until they married, although some moved out to temporary non-familial living first. Non-familial living consisted mainly of lodging in another household, but working outside it, which in a way was a forerunner of the modern pattern, in sharp contrast with the remnant of preindustrial times: the flow from rural areas into Gothenburg of teenage women immigrants to become residential domestic servants. Interestingly, we found that the main determinants of home-leaving in studies of modern-day populations were equally important in the population of Gothenburg in 1915–1943. For both young men and young women, having their own resources (employment, earnings) was positively associated with the likelihood of leaving the parental home. We also found clear gender differences. A higher level of human capital of the father was associated with later home-leaving to marry for sons, and earlier leaving for non-familial living for daughters. Lower levels of household income, or the presence of minor siblings or a widowed parent were push factors for non-familial living for daughters. We found no similar push factors for sons.

Keywords:

• Marriage
• transition to adulthood
• age norms
• intergenerational transfers
• interwar period

Notes

1. With remaining in the parental home as the base risk, scholars have, for example, used the following optional living arrangements associated with leaving home: marriage, premarital residential independency, and residential semi-autonomy in dormitories and barracks (Goldscheider & DaVanzo, 1989); marriage, premarital residential independency (Avery et al., 1992); family independence (marriage/cohabitation); non-familial living (Goldscheider et al., 1993); marriage, independent living (Buck & Scott, 1993); and living with or without a partner (Mulder & Clark, 2000).

2. A net flow of resources from the parents to the children is supported by findings showing that leaving the parental home increases the likelihood of poverty for young adults (Aassve, Davia, Iacovou, & Mazzuco, 2007), also in present-day Sweden and the other Nordic countries (Kauppinen et al., 2014).

3. There are other historical, longitudinal demographic databases covering urban populations in Sweden in the early twentieth century. The POPLINK database, held by the Centre for Demographic and Ageing Research, Umeå University, covers the towns of Umeå and Skellefteå. The Roteman archives, held by the Stockholm City Archives, include data on the population of Stockholm between 1878 and 1926. Neither of these databases contains earnings variables, which are crucial for this study. The Scanian Economic Demographic Database, held by the Centre for Economic Demography, Lund University, is currently adding the population of the town of Landskrona to the database, including earnings variables, but the data are not available yet.

4. Initially, separate income registers were kept, but from the early 1920s, income records were integrated into the poll-tax registers.

5. Earnings are annual taxable earnings per year for each person in the household, distributed by earnings from employment, self-employment, capital, and property. The limit for taxable income was 800 kronor in 1915 and 1919, and 600 kronor from 1923 onwards.

6. Including seven in-migrating children who probably moved into Gothenburg with their families.

7. Seven people who migrated in to Gothenburg as children were excluded from the estimate of the means.

8. Two-thirds of these people were observable in the panel years of 1915 and 1919, before the central register of the poll-tax office that recorded migration was developed.

9. In our data, we observed the place of residence and civil status of an individual every fourth year. We knew, for instance, the age of a person in the first panel year after a change in place of residence or civil status, but we cannot say exactly which year the change took place, which is why we have refrained from estimating the mean or median age at the time of leaving home, or at marriage. In addition, we cannot distinguish whether a move from the parental home occurred at the same time, or slightly before, or after, a marriage. We have to allow for the possibility that the decision to marry, and the decision to leave the parental home, were influenced by different factors. In our analyses, we therefore separate moves from the parental home that occurred together with marriage from other moves.

10. In the poll-tax records, the occupational title and taxable earnings were noted in separate columns, the occupational title relating to the situation in December and the earnings relating to the entire year. Individuals with stable jobs and taxable earnings typically appeared with both an occupational title and an earnings record. Younger people with earnings below the level of taxability appeared with an occupational title but without any earnings record. In a similar way, older individuals with unstable employment appeared in the poll-tax records with an earnings record but without any occupational title. Note that no recorded earnings did not necessarily mean that a person had no earnings at all, but that the earnings were below the limit of taxability.

11. As mentioned in Section 4, above, 34 males and 23 females in the early leaver group were brought up in Gothenburg. Adding them to the non-familial living outcome [2(a) in Table 1] would not change the pattern we found in the study population. Leaving the parental home to marry was still more frequent than leaving the parental home for non-familial living, especially for women.

12. The estimation includes both sample persons who moved out of the parental home for non-familial living and the group of early leavers (not in the regression sample).

13. For Sweden as a whole, age at first marriage increased from the 1910s, and peaked in the 1930s (Statistics Sweden, 1999, p. 100).

14. Adding those 57 sample persons who were brought up in Gothenburg, and who had left the parental home at the age of 16–19 (not included in the regression sample), would lower this age by 2 years.

15. The results regarding both the total household income and the parent’s occupational class were robust, excluding the other variable from the model (results not shown).