The Emergence of the Dutch Housewife Revised. How Shifts in Local Labour Market Structures Shaped Dutch Unmarried Women's Labour Force Participation, 1812–1929

e-ISSN: 2352-6343 DOI article: https://doi.org/10.51964/hlcs9581 The article can be downloaded from here. © 2021, Boter This open-access work is licensed under a Creative Commons Attribution 4.0 International License, which permits use, reproduction & distribution in any medium for non-commercial purposes, provided the original author(s) and source are given credit. See http://creativecommons.org/licenses/.

HISTORICAL LIFE COURSE STUDIES VOLUME 10, SPECIAL ISSUE 3 (2021), 130-134, published 31-03-2021 The Emergence of the Dutch Housewife Revised Most studies on the long-term development of female labour force participation argue that social norms and rising wages were key drivers. However, the majority of these conclusions apply to married women. Instead, this study zooms in on unmarried women. Based on nearly 2 million marriage records that have been digitised by Kees Mandemakers' LINKS project, it shows that there were large regional differences in the levels of labour force participation that were closely connected to local job opportunities. This research concludes that even though social norms and income levels were indeed important, local sectoral employment shares were the key driver of Dutch unmarried women's work during the long-19th century.

Corinne Boter
Utrecht University How Shifts in Local Labour Market Structures Shaped Dutch Unmarried Women's Labour Force Participation, 1812Participation, -1929 Social norms regarding domesticity have long been the dominant explanation for declining Dutch female labour force participation (FLFP) rates during the long-19th century (van Poppel, van Dalen, & Walhout, 2009). However, these norms principally affected married women, as unmarried women did not have their own households to take care of. Furthermore, scholars of recent studies find that local labour market structures were a more important driver of FLFP than social norms (Boter & Woltjer, 2020;You, 2020). To truly understand the trajectory of women's work, it is therefore necessary to study unmarried women separately and to explore demand for labour, in relation to supply, as a factor of influence.
Based on Dutch marriage records from the period 1812-1929, this paper 1 reconstructs unmarried women's labour force participation (UFLFP) and explores how and why it changed over time. From 1811 onwards, vital events (birth, marriage, and death) were registered in a uniform way in all municipalities, which makes marriage records exceptionally fit for long-term, comparative research. These records have been digitised by the LINKS project, developed by professor Kees Mandemakers (International Institute of Social History), which aims to reconstruct family lives by linking vital events. UFLFP is calculated as the share of brides with a listed occupation in their marriage record. Based on recent collaborative work with Pieter Woltjer (2020), it has become clear that UFLFP decreased during the long-19th century, but that there were vast differences between municipalities, which cannot be explained by social norms. We found that a (much) higher than average share of the brides listed an occupation in municipalities in which industries with a high demand for female labour dominated the labour market. This leads to the conclusion that local labour market structures were the most important determinant of UFLFP. More generally, shifting labour market structures can therefore explain an important part of the overall decline of UFLFP in the Netherlands.
During the 19th century, FLFP rates in Western Europe declined significantly, which is usually explained by the spread of the domesticity ideal from the higher to the lower societal classes. Households increasingly desired to consume 'non-market produced' goods, such as a clean and cosy home and healthy food, and to achieve this, women withdrew from the labour market to become housewives. However, the inclusion of demand-side factors in research on women's labour history has gained ground during the past decade. Based on the 1881 census enumerators' books, Xuesheng You has shown that everywhere in England and Wales, married women's LFP was lower than that of unmarried women, but that "relative spatial patterns of female LFPRs [labour force participation rates] were largely the same across different marital groups" (You 2020, p. 112). He has found regional pockets with high FLFP rates in, among others, Lancashire (cotton industry) and Yorkshire (wool and worsted industry), where demand for female labour was exceptionally high. He concludes that demand for women's labour was the most important driver of FLFP because "[…] the supply-side conditions, though with variations, must have been more homogenous across the country for each marital group" (You, 2020, p. 115).
The domesticity norm has also been proposed as an explanation for the decreasing Dutch UFLFP rates during the long-19th century. Frans van Poppel et al. (2009) have found that the share of brides that listed an occupation in their marriage record decreased from more than half in the first half of the 19th century, to around 15% in the 1920s. They reason that because women increasingly quit their jobs before marriage, thus before they were supposed to take care of their own households, the domesticity norm must have been the most important incentive for women to withdraw from the labour market.
However, there are two reasons to question this line of reasoning. First, the domesticity norm could also have a positive effect on UFLFP. For instance, working as a domestic servant before marriage could teach young girls about the ins and outs of good housewifery, an excellent preparation for married life. Moreover, working in general could provide young girls with a small amount of capital for the foundation of a new household. Second, as van Poppel et al. also note, there were differences between the Dutch provinces, which they explain by differences in wealth and labourers' skill levels (van Poppel et al., 2009, p. 114). Moreover, the marriage records actually reveal that UFLFP not only differed between, but also within provinces. In some municipalities, an exceptionally high share of the brides listed an occupation. Figure 1 shows the percentage of brides with a listed occupation in the period 1890-1929 at the municipal level. Even though on a national level, UFLFP had already decreased to less than 20% by 1900, there were several municipalities with UFLFP rates of more than 50%. Crucially, municipalities with particularly high and low UFLFP rates were often situated close together. In line with You's conclusions about England and Wales, I argue that social norms could never have been this regionally dependent and that instead, we have to look at local demand for labour to explain regional differences. Source: Boter and Woltjer (2020, p. 15). NB: The highlights of Emmen and Enschede were added for the present paper.
There are two pockets of high UFLFP highlighted in Figure 1 to which I will pay closer attention. In these regions, female labour was demanded for two different industries. First, in the municipality of Enschede (situated in the larger region of Twente), the textile industry dominated the labour market from halfway through the 19th century onwards. In 1830, the Southern provinces -present-day Belgium -seceded from the Netherlands, after which the import of Flemish cotton cloth was immediately restricted. To keep

LOCAL DEMAND FOR FEMALE LABOURERS
exports of Dutch cotton yarn and cloth flowing, the cotton textile industry was relocated to Twente. 2 This boosted demand for labour and the mechanization of the textile industry in the decades to follow would only reinforce this development. As in other countries, an important share of the textile labour force was female. However, demand for female labour fluctuated over time as a result of gender-based divisions of labour in factories, where most of the machines in use were exclusively operated by either men or women. 3 For instance, in spinning, throstles were operated by women and self-actors by men. Because the number of self-actors increased during the second half of the 19th century, the demand for male labourers increased concurrently. However, in the early-20th century, the ring spinning machine was introduced, designed to be operated by women (Boter & Woltjer, 2020, p. 7). Consequently, the demand for female workers increased more rapidly relative to male labourers.
Second, in Emmen (southeast Drenthe), the peat industry was extremely important and demanded significant female labour. Peat production -the extraction and drying of peat to be used as fuel or building material -increased dramatically during the period 1850-1925 in the northern provinces of Groningen, Friesland, Drenthe, and Overijssel. Drenthe was by far the most important province in terms of output, being responsible for almost half of the total peat production in the period 1617-1950, 45% of which was produced in the southeast (Gerding, 1995, pp. 271-273). The production process was heavily segregated along gender lines. Women played an especially important role in the drying process and the stuffing of dried peat on ships for transportation while men were responsible for cutting the peat out of the ground. Since the drying process required relatively substantial labour inputs, demand for female labour was high (Gerding, 1995, p. 353). Moreover, permanent peat labourers (as opposed to seasonal labourers) were commonly hired as one family unit, meaning that every household member participated. The vital role of women workers in the textile and peat industries is clearly reflected by the high share of brides with an occupation in Enschede and Emmen (see Figure 1).
Zooming out to nation-wide economic developments tells us more about why overall UFLFP rates declined during the 19th and early-20th centuries. The marriage records show that during this period, the employment shares of the sectors with a positive effect on UFLFP decreased, notably the textile industry and private services. At the same time, those sectors with a negative effect, such as wholesale and retail trade, increased. 4 The occupational censuses of the period 1849-1909 paint a similar picture, which shows that the marriage records reliably reflect national economic structures. The share of the total Dutch labour force working in the textile and apparel industries decreased from 9.1% in 1849 to 7.3% in 1909, while in trade this share grew from 6.8% to 9.6%. Thus, shifting sectoral employment shares explain an important part of the decreasing UFLFP.
Still, there were also important supply-side factors at play. First, the occupational status of the groom was an important determinant of the chance that a bride would state an occupation. 5 In the total sample of records from 1812-1929, only 14.3% of the women marrying a man with an upper-class occupation recorded an occupation, as opposed to 50.7% of the future wives of unskilled farm workers. In other words, the higher the occupational status of the groom, the lower the likelihood of a bride's occupational mention on the marriage record. Furthermore, rising levels of GDP per capita negatively affected UFLFP because the necessity for women to earn wages decreased as a result of increasing welfare. Second, social norms regarding women's work certainly played an important role. Qualitative evidence has shown that by the end of the 19th century, most people thought married women should not be working away from home and that a good housewife was crucial for a proper lifestyle. Still, as mentioned before, the effects of these norms on how people thought about unmarried women's labour could be negative as well as positive (Boter & Woltjer, 2020, pp. 7-8).
2 A large portion of the Dutch cotton yarn and cloth was exported to Java by the Netherlands Trading Society. 3 Mechanisation of the Dutch textile industry only truly took off during the second half of the 19th century and spinning mechanised decades before weaving. 4 See Boter and Woltjer (2020) for more information about the categorisation. 5 We used HISCLASS to classify all occupational titles into one out of seven categories.
This paper has argued that the vast regional differences in UFLFP suggest that local demand for labour was the most important determinant of UFLFP. Social norms cannot explain these differences because they were relatively homogenous across the country. Indeed, the labour markets in two of the municipalities with exceptionally high UFLFP rates (Enschede and Emmen) were dominated by industries with high demand for women workers. The textile industry traditionally employed large numbers of women, although demand for labour depended partially on the type of machinery in use. Women also played a vital role in the heavily gender-segregated peat industry. The relative importance of such sectors with a high demand for female labour decreased during the long-19th century, which can explain a considerable part of the overall decline in UFLFP rates.
Shifting sectoral employment shares might also explain part of the decreasing (U)FLFP in other Western European countries. Because civil records are available for most European countries from at least the mid-19th century onwards, and because their contents are relatively uniform through time and space, they allow for temporal and international comparative analysis of the drivers of FLFP. Furthermore, Kees Mandemakers' LINKS project offers unique opportunities to link information on birth-, marriage-, and death certificates to approach the history of (U)FLFP from a life-course and household perspective. Such research is an important next step for understanding the role of supply-and demand-side factors in shaping the historical trajectory of women's work.