Three myths about old age before modernity – and why historians should care

ABSTRACT

Despite the increasing challenges twenty-first-century societies face in accommodating older adults, many misconceptions about old age before modernity continue to exist. These are rarely expressed in ‘formal theory’ produced by academics, but rather persist in ‘lay theory’ which is anchored in collective memory and popular culture. Such taken-for-granted beliefs are nevertheless influential, first because they determine which questions academics do and do not ask about old age in history, and second because they present society with a grand narrative about ageing in the past, present, and future. We provide researchers with a research agenda. We do so by identifying three myths about old age: 1) people usually lived short lives and continued in their line of work until they dropped dead without having ‘retired’ from work; 2) men and women who did grow old simply moved in with their next of kin; 3) pensions played no role during old age. We debunk these based on a case study of the Nordic Countries before c. 1850 and hope to encourage research into myths about old age before modernity in other European regions.

KEYWORDS:

• Old age
• retirement
• family support
• pensions
• history

1.

The 2019 Hollywood horror movie Midsommar contains a gruesome scene depicting ättestupa: older men and women climb up a high cliff and jump to their deaths. Such suicide to prevent yourself from becoming a burden to the community is a well-known Nordic myth. The historian Birgitta Odén (1921–2016) connected it to the push for better care of older adults during the formative stages of the Swedish welfare state. She explains how the author Ivar Lo-Johansson (1927–1990) used ättestupa to convince the Swedish general public of a long history of neglect of old people and the necessity to correct this by creating a social welfare system (Odén, 1996, pp. 11–12). Thus, a myth about old age made public opinion ready for universal welfare (Gaunt, 1996, pp. 50–53). In contrast, this tale about suicide has also recently been used to argue against the welfare state. In 2023, the economist Yusuke Narita used the ättestupa scene from Midsommar to start a discussion about how to deal with Japan’s ageing population. Although Narita denied having suggested older persons should commit mass suicide, his critics have claimed that his using the image of elderly men and women jumping off cliffs is no innocent metaphor intended to initiate a discussion on pension reforms, but an idea that resonates with younger Japanese who object to making social transfers to the one-third of the population who receive old age pensions (Rich & Hida, 2023).

One of the tasks of historians is to prevent historical myths from being used in contemporary debates such as the one on how to care for ageing populations (Lagerlöf Nilsson et al., 2022, p. 166). Earlier scholarship used such terms as ‘beliefs’ or ‘myths’ and ‘mythology’ to describe these misconceptions. Thus, Jill Quadagno already remarked that ‘beliefs about the aged do not conform to their positions in society’ (1982, p. 22). Such myths can be considered as exponents of ‘lay theory’: ‘a structured and coherent system of beliefs, rules, and concepts’ that usually serves to provide a perception of truth. It differs from a ‘formal theory’ which aims to establish truth. The two usually interact: formal theory can inform or inspire the construction of lay theory, but it is equally true that ‘lay theories can inspire the construction of formal theories whether explicitly, unconsciously, or automatically’ (Levy et al., 2006, pp. 5–6). In scholarship, this is perhaps best visible where historians of the modern era construct their research by making assumptions about the premodern era.¹ One prominent example is the ‘fairy-tale’ that the more distant past was ‘a golden age of ageing’ (Ottaway, 2004, p. 1). Outside academia, false beliefs emerge in popular culture where they can play ‘a significant role in (re)producing social norms’ (Haring et al., 2023, p. 9) and can contribute to polarization because ‘beliefs’ ‘often divide people into different camps’ (Zedelius & Schooler, 2017, p. 72). It follows that it is important to uncover any unconscious or automatic assumptions that may be driving decision-making, both among policy-makers and also among scholars who may base the questions they ask on false beliefs. Therefore, in this paper, we try to identify and debunk three myths about old age before modernity that are rooted in lay theory but are also likely to influence formal theory.

The three myths we propose in this paper concern the concept of ‘functional age’, which defines as ‘old’ ‘those who were infirm, frail, and suffering incapacities of body or mind to the extent that they could no longer fully support or take care themselves, and who also gave the appearance of being old’ (Roebuck, 1979, p. 417). This means that we apply a ‘functional line of reasoning based on a biological perspective’ (Bulder, 1993, p. 13) and focus on functionality rather than perception.² Even though these myths largely exist in the realm of ‘lay theory’, our goal is not to study current false beliefs: a more precise study of lay theories about old age before modernity is a desideratum left to the field of cultural studies.³ We have chosen to rather to use ‘myths’ as a starting point of a discussion that might help scholars becoming aware of misconceptions that inform research and indicate research gaps they might address in future studies. Two of our myths were already identified in the 1980s by Jill Quadagno⁴ and we add a third based on a ‘research gap’ identified by financial historians (Boyer & Schmidle, 2009, pp. 255–256; Morris, 2005, p. 172; p. 148; Zuijderduijn, 2024, pp. 87–90). The first myth that we propose is that people usually lived short lives and continued in their line of work until they dropped dead without ‘retiring’ from work. The second myth is that men and women who did grow old simply moved in with their next of kin. The third myth is that pensions did not play a role during old age. We will address these three myths and demonstrate that pre-modern people were likely to grow old and face later-life impairments that affected their labour input – and already did so years before they eventually passed away; that they could not necessarily rely on family; and that they therefore used various financial techniques to prepare for their final years.

Our main purpose is to show how the three myths hold up to empirical research and to provide suggestions for future research. We focus on the Nordic countries before c. 1850⁵ because the welfare state in these countries has inspired old age research which provides a starting point for our discussion.⁶ In addition, since the Nordic countries pioneered national statistics, this allows us to paint a picture of elderly men and women in the past. Our aim is not to provide a complete historiographical overview of old age in the Nordic countries, but rather to explain which faulty images of old men and women in the past should be corrected and how this can help to improve the general understanding of old age in the past. To this end, we combine historiography, source editions, and primary sources from the Nordic countries, as well as limited data from other regions. Even though we focus on the Nordics, we hope to encourage research identifying the interaction between ‘lay’ and ‘formal’ theory in other European regions as well. More concretely, we hope to demonstrate that when it comes to old age, the past is not the entirely foreign country it is often made out to be. Rather, pre-modern ageing posed similar challenges to the ones we face today.

2. The first myth: Working until you dropped dead

In 1763, the Swedish doctor David von Schultzenheim (1732–1823) observed that ‘many people in their fifties suffer from old-age impairments’. He continued by noting that ‘most of them only pass away in their seventies or eighties’ (Odén, 1996, p. 91). If the doctor is correct, many pre-modern individuals experienced an ageing process that was accompanied by the development of physical and mental conditions that threatened their livelihood. Did Von Schultzenheim exaggerate when he suggested many of his contemporaries became unfit for work decades before they eventually died? In his influential book Growing Public: Social Spending and Economic Growth since the Eighteenth Century, Peter Lindert claims that ‘few people survived to be elderly’ and goes on to explain that this is one of the main reasons why public pension systems did not emerge until modern times (2009, p. 7).⁷ In doing so, Lindert subscribes to the belief that life was short before modernity. Historians of old age have long resisted this image and pointed out that it was not at all unusual to grow old. They often cite a misinterpretation of life expectancy figures (Ottaway, 2004, pp. 21–22). Today, Western European newborns can be expected to reach the age of 80, but a century ago life expectancy at birth was not much more than 45 years. If we go back to the eighteenth century, the unique Swedish data collected by Tabellverket, the first national bureau of statistics in the world, show that newborn boys and girls in Sweden and Finland would on average live to reach their thirties (Kannisto et al., 1999, p. 10). Chilling figures for sure, but they do not tell us much about old age because low life expectancy at birth was almost exclusively caused by staggeringly high child mortality. Anyone who survived childhood could reasonably expect to reach 60, which is often regarded as the threshold of pre-modern old age (Jåstad, 2011, pp. 31–32; Ottaway, 2004, pp. 18–21). Thus, in pre-modern Finland, youths aged 15 could expect to live to close to 60 (Kannisto et al., 1999, p. 5); in Sweden, 30-year-old men could expect to live about another 33 years and reach 63, and women could expect to live another 35 years to the age of 65. At the age of 50, men could expect to live another 18 years (reaching 68), and women could count on another 20 years (reaching 70). These figures are lower than Von Schultzenheim’s ‘seventies or eighties’, but we should keep in mind that he wrote about people in their fifties (not at the age of 50): those who were already in their mid to late fifties would likely have celebrated their 70th birthday.⁸

Growing very old was not at all unheard of. When eighteenth-century scholars thought of extreme old age, they were not thinking of men and women who exceeded 60 or 70, but rather of those who exceeded 100 years: the sociologist Peter Laslett wrote about a ‘cult of centenarians’ in the early-modern period (1999, p. 25). Thus, the Danish author Bolle Willum Luxdorph (1716–1788), a resident of Copenhagen, collected images of the very old and published on the subject. He focused on people older than 80, probably because the book he was writing, Index tabularurum pictarum et cælatarum qvæ Longævos repræsentant (1783), was intended as a gift to a friend on his 80th birthday (Kjaergaard, 1995). To gather data, he approached clergy in Denmark and Norway and asked them to indicate octogenarians living in their districts. His Norwegian contacts reported fewer than 30, his Danish contacts less than 10. This was not at all impressive, but surely part of Luxdorph’s problem was that he did not only ask the clergy to indicate people who claimed to be that old but also to produce evidence, using baptismal records to confirm the octogenarians’ age. Although such records were available in Denmark and Norway – they had been kept since the seventeenth century – it seems that few clergymen were willing and able to carry out such an investigation, and this must be why Luxdorph’s study yielded fewer old individuals than might have been expected. A Danish census taken in 1787 clearly shows that Luxdorph’s hometown of Copenhagen already counted 363 octogenarians, that another 576 lived in other Danish cities, and in the countryside 3,789 people had celebrated at least 80 birthdays.⁹

Copenhagen was not unique in harbouring many octogenarians: in Stockholm similar numbers are observed. In 1763, when Von Schultzenheim made the claim cited above, there were 284 people aged over 80 in the Swedish capital. When we use 60 as the threshold for old age, between 6% and 8% of Stockholm’s population can be classified as ‘old’. Elsewhere in Sweden, the share of old adults was even greater: during the first Swedish census of 1750, no fewer than 171,949 Swedes older than 60 were counted, accounting for 9.6% of the total population (Historisk statistik, 1969, p. 68). The share of elderly individuals continued to be between 8% and 10% until 1890, when an increase set in that eventually would lead to the 25.7% of Swedes who were over the age of 60 in 2021. Like today, females were more prominent among the older population: in 1750, 11,0% of the women were over 60, as opposed to 8,1% of the men. As Table 1 indicates, this was not only the case in Sweden but common throughout the premodern Nordic Countries: usually 8% to 10% of the population was older than 60.

Table 1. Share of males and females over the age of 60.

	% males >60	% females >60
Denmark 1787	8.0	9.3
Denmark 1800	8.9	9.9
Finland 1751	7.7	9.4
Finland 1800	5.9	6.7
Iceland 1703	5.4	8.8
Iceland 1800	8.9	12.7
Norway 1800	7.9	9.8
Sweden 1750	8.1	11.0
Sweden 1800	7.8	9.7

Source: Gille (1949), pp. 20–21.All observations are based on national censuses.

2.1. All observations are based on national censuses

Von Schultzenheim was probably correct when he claimed many people grew old, but what about his second claim – that many already suffered from old-age deficiencies in their 50s? This is much more difficult to verify because pre-modern sources hardly ever give an indication of the physical and mental condition of inhabitants and the ability to make a living. To investigate how functionality developed over the life course, we use income tax records. We have collected data from the 1800 Stockholm mantalslängder,¹⁰ registers which, among many other things, recorded whether people had paid income tax. There was a low threshold for doing so: even low-income servants and day labourers paid income tax.¹¹ To date, we have randomly sampled 11 out of 556 Stockholm blocks (see Appendix). These blocks were home in total to 1,467 individuals, 493 of whom were men over the age of 15. We focus on the share of them paying any income tax.¹²

Figure 1 divides men into 5-year and 10-year age brackets and into manual and non-manual workers.¹³ Until the 41–50 age bracket, more than 90% of the manual and non-manual workers earned enough to meet the threshold for the income tax. In the following age brackets the share of taxpayers drops below 90%, first among manual workers and later also among non-manual workers. In the age bracket of men over 61, only 20% of manual workers met the income tax threshold, and 67% of the non-manual workers. We interpret this not as old men retiring altogether, but rather that ill health forced them to work less, or to transfer to lower-paying jobs, causing them to drop below the threshold for paying income tax. Income tax data thus suggest that the working capacity of manual labourers began to decline after they turned 50. Future research should provide more precision about the development of income over the life cycle, and how the ageing process impacted workers in town and countryside in the Nordic Countries. But the notion that 50 was a turning point for many manual workers resonates with other sources: in Sweden, in the later eighteenth century, the poor aged over 50 were forbidden to move to Stockholm, Gothenburg, or Malmö because it was believed their working days were over by then. And when a universal old-age pension system was discussed in the mid-19th century, it was suggested that servants could begin to draw from the fund after they had turned 50 (Högman, 1999, pp. 43–44).

Figure 1. Proportion paying income tax: manual and non-manual workers.

Source: Mantalslängder Sample of 493 men over the age of 15 in 11 Stockholm blocks in 1800 Including active soldiers and state officials who did not pay taxes (see Appendix).

The contemporary doctor Von Schultzenheim claimed that old-age impairments manifested themselves around the age of 50; income tax data suggest he may have been correct with respect to manual labourers. Non-manual labourers managed to stay above the threshold for taxation for about 10 more years. This may have been the result of them earning more than manual labourers. But it may also have been the result of non-manual labourers not suffering old-age impairments to the same degree as manual workers did. The latter were particularly likely to develop labour-related health issues as they performed heavy work that is nowadays largely mechanized. The obvious point of comparison is twenty-first-century workers in physically demanding jobs who struggle to continue in their line of work once they are in their fifties¹⁴; it is difficult to imagine that pre-modern manual labourers working before the onset of labour regulation and modern medicine lasted any longer.¹⁵

To make matters worse, pre-modern people are likely to have developed various impairments – old-age deficiencies – at a faster pace than today. Exposure to a poor diet, disease, and an unhygienic environment ‘in utero, in infancy, and in later developmental ages’ leads to ‘scarring,’ with long-lasting health effects that manifest themselves during adulthood and contribute to poor health later in life (Fogel, 1994, pp. 90–91). Thus, T. Bengtsson and Broström (2009) have demonstrated that early life exposure to disease led to adverse health and labour capacity in older ages in South Sweden in the nineteenth century (2009, p. 1589). There can be little doubt that before improvements in sanitation and hygiene, and the development of vaccination programmes, ‘scarring’ also occurred frequently during early childhood. We can estimate the extent of ‘scarring’ by looking at the infant mortality rate, which measures the number of deaths in the first year of life, and thus get an idea of the health environment children were born into. In the pre-modern Nordic countries, at least one in five babies died before turning one.¹⁶ These figures were even higher in cities, where children grew up in a particularly adverse environment (S. Edvinsson, 1992, p. 188). However, regardless of whether they were born in an urban or rural environment, pre-modern infants were much more likely to suffer ‘scarring’ than children born today; as a result, the former were likely to develop impairments in later life stages.

Unhealthy living conditions had an impact not only on the young but also on the adult population, and this brings us to another reason why pre-modern populations may have developed impairments at a faster pace than populations today. The pre-modern world harboured many contagious diseases, such as smallpox, cholera, and tuberculosis, that not only killed many thousands but also caused impairments. Take tuberculosis, for instance: today a quarter of the world’s population is infected and at risk of suffering impairments such as respiratory impairment (in 20.7% of the cases), musculoskeletal impairment (17.1%), and mental health disorders (23.1%) (Alene et al., 2021, p. 3). This and other infectious diseases contribute to impairments and lower the age at which the average individual can no longer make a living. We can safely assume the same was true for the past. According to James Riley, historical diseases causing lasting damage in tissue or cells include ‘rheumatic fever, typhoid fever, diphtheria, influenza, syphilis, tuberculosis, and others’. About these diseases, he writes: ‘people who survived [sickness] often lost the ability to function normally, but that loss of ability did not reduce their life expectation in proportion’ (1989, p. 36).¹⁷ To put it another way: survivors of disease lost healthy life years without losing equally many life years – just as Von Schultzenheim already claimed in the eighteenth century. The notion that old age impairments may have come earlier in history may also be deduced from Daniel Schäfer, who writes about ageing in the twenty-first century that ‘the average decline of bodily functions shifts to later life stages or slows down when compared to earlier generations’ (2022, pp. 6–7). Thus, many generations ago loss of functionality may have come earlier than today. How much earlier is almost impossible to say at the present state of research because large health surveys of old people were not taken until recently. An exception is an inquiry of United States army veterans born in or shortly before 1845, and surveyed in 1910, which suggests that these men were in a worse shape than their counterparts surveyed in 1983, causing Robert Fogel to conclude that ‘chronic conditions were far more prevalent throughout the life cycle for those who reached 65 before World War I’ (1994, p. 32). Although this hints at a relatively early loss of functionality in the past, more research is needed to reconstruct the premodern ageing process.

3. The second myth: Loving families

Before Sweden announced cutbacks in its pension system in the mid to late 1990s, its government conducted several extensive inquiries into the pension system and how to reform it. Interestingly, the first of these included a ‘background’ in which the investigators provided a summary of the ‘historical developments’ that led to the welfare state, which suggests that historically the family was the most secure old age provision (Allmän pension, 39). Although probably largely correct, the statement also resembles the lay theory of a ‘golden age of the elderly’ when the old were assured they could live with the young once they could no longer support themselves (Horden & Smith, 1998, pp. 2–4).¹⁸ This myth about old age is very much alive, in particular in the conservative idea that there are family traditions that we have lost not too long ago, and can therefore easily retrieve, to compensate for welfare cuts. Thus, Sweden’s Christian Democrats, for instance, seem to believe it to be ‘natural and positive’ to provide voluntarily care (anhörigstöd) for elderly relatives, and the liberal-conservative party Moderaterna likewise suggest more family support is required (Stenberg, 2022, pp. 76, 81, 102). Since the Enlightenment, liberals and conservatives have suggested that family safety nets were strong in pre-modern societies. Thus, Adam Smith (1723–1790) already claimed that in traditional societies the elderly were more respected than in modern societies (Högman, 1999, p. 10). Likewise, Fredrick Le Play (1806–1882) contrasted a lost world of loving families with the much colder kin networks of the industrial age and suggested his contemporaries should rediscover family traditions (Janssens, 2002, pp. 2–3). Only in the 1960s did scholars gradually begin to rectify this image of strong pre-modern family ties, beginning with Peter Laslett in his book The World We Have Lost. He demonstrated that pre-modern England was not dominated at all by extended family households where grandparents, children and grandchildren supported each other. Before the Industrial Revolution, people in England usually lived in the same nuclear family households we are used to today. Laslett therefore identified the extended family household as one of the ‘misbeliefs about our ancestors’ (2021, p. 81).

One reason why this second myth about old age has nevertheless prevailed is because our view of the past is blurred by a temporary spike in the number of extended families. Steven Ruggles has demonstrated that the number of households with extended kin in England and the US doubled during early industrialization (1750–1870) and peaked in the second half of the nineteenth century. Subsequently, extended households gradually became less common until they were back to pre-industrial levels by the 1970s (1987, pp. 4–6). The ‘golden age’ where old parents were welcomed into the homes of their children only existed for a brief period. According to Ruggles, extended families could emerge because of three factors. First, rising incomes allowed people to take care of ageing relatives. Second, the population grew during the first stage of the demographic transition, when mortality declined, while fertility remained high. During this stage, which occurred in the nineteenth century in most Western countries, ‘the potential for the formation of extended families is unusually high’. Third, this period saw the emergence of new values with respect to family life (Ruggles, 1987, pp. 133–134, 206–207). By dictating that women should not work but focus on providing care, the emerging breadwinner-homemaker model helped to create a pool of caregivers. This model could be quite persistent: in Sweden the so-called ‘housewife era’ lasted until the 1970s (R. Edvinsson & Edvinsson, 2017). The three factors fostering the establishment of extended families were unique to the era of industrialization. They helped create the ideal conditions for providing family assistance: an increasingly wealthy population with a relatively large number of relatives, of whom half specialized in homemaking and could take care of old relatives. This is the image we see when we look back at our own family histories, but to think our grandparents and great-grandparents had normal families would be a mistake: they lived in highly unusual times.

Whether independently-living old men and women were much worse off because they lacked caregivers, is difficult to establish. Historians have rightfully criticized Laslett’s equation of nuclear families with a lack of family assistance: relatives living in separate households may very well have helped each other out. Beatrice Moring summarizes the available evidence on family relations in the pre-modern Nordic countries by indicating that even though nuclear family households were relatively common, this does not necessarily ‘support the idea of a less family-orientated Northern Europe’ (2015, pp. 23–24, 29, 40). We agree that household structures cannot tell us much about the availability of care: old parents and children can live in separate nuclear family households and yet support each other on a daily basis. But we do believe that it is important to problematize family assistance and to point out that notions of a past golden age of ageing are too optimistic: a variety of factors could prevent old men and women from being cared for by kin and may have forced them to make alternative preparations for old age (section 4).

First, many old men and women never had any children. According to Regina Toepfer, ‘childlessness was more widespread in medieval and early modern Europe than it is today’, especially in cities (2022, p. 4). Many urban women never married and – given the taboo on illegitimacy – unmarried women were not very likely to have given birth. Second, even men and women who had had children could not be sure their offspring would survive long enough to provide eldercare. As explained above, about one in five infants died before reaching their first birthday, and another 1 in 10 did not reach the age of 10 (T. Bengtsson & Lundh, 1999, pp. 4–5). To make matters worse, there are indications that infant mortality clustered in certain families, exposing parents to the risk of growing old without any living children (S. Edvinsson et al., 2005). Leonardo Fusè’s study of the countryside of pre-industrial Sundsvall, in Sweden, reveals that 10.6% of people never had any children, and of those who did have at least one child, 4.6% were childless by the age of 60. Between one in six and one in seven entered old age without children. Partly as a result, more than one in four old men and women did not live with any of their offspring at the age of 60, and by the age of 85 this was one in two. Fusè’s analysis also shows many 60-year-old people did not have a son or daughter living in their village (slightly more than one in five) or in the parish (slightly less than one in five) (2008, pp. 67–68, 72, 74). This suggests a third factor to consider: children might have migrated, both within and outside the country. Daughters could have found work as maidservants in towns, and sons in the army or navy, and as a result, they would have been unable to offer day-to-day support. Fourth, intergenerational support required relations between parents and children to be good. In Sweden and Finland quite a few sons and daughters appear to have been brought to trial for not supporting their elderly parents (639 trials in Sweden and Finland between 1745 and 1754) (Liliequist, 2014, pp. 20–21). This is remarkable because violence against parents was a capital offence; yet every year dozens of parents brought charges that could result in the execution of their own flesh and blood. Clearly, family relations could sour. Fifth, even if adult children were alive, living nearby and willing to take care of their parents, they also needed to be able to do so. Today, many adults face the challenge of ‘sandwiching’: they are in their child-bearing and -rearing years when their parents enter old age and require help, and are therefore overburdened with caring for both their own young children and old parents. Pre-modern families experienced the same problem, especially where marriage was postponed and the age at first marriage was between 25 and 30 (Lindberg et al., 2017, p. 161; Lundh, 2003, pp. 41–43). In the words of Peter Laslett:

Where marriage was late, as it had been for so long in north-western Europe, the parents of married couples began to lose their children’s earnings and to lose each other in widowhood at the very point in their children’s life-course where these children in turn may have been particularly hard pressed because of the arrival of their own offspring. Simultaneously bodily weaknesses increased in these elders, and employability declined. (1988a, p. 169)

We should add that subsequent research has suggested that the age at first marriage in Northern Europe was even higher than in north-western Europe (England and the Low Countries) and that therefore sandwiching likely was an even greater problem in the Nordic countries (Ogilvie & Dennison, 2014, p. 670). Sixth, it has been suggested that social norms prescribing that children leave the parental household – neolocality – may have prevented intergenerational support because ageing parents were expected to be self-sufficient. According to Laslett, this put them at risk of ‘nuclear hardship’ (1988a). This idea of norms standing in the way of family support has also been questioned: research has demonstrated strong economic ties between parents and their adult children (Moring, 2015, pp. 23–24, 37).

If, for whatever reason, children were unable to provide support, perhaps siblings would have been available? Such supporters had to be alive: estimates for premodern England suggest that it was not very common to have a living brother or sister at the age of 66 and there is little reason to suspect this was very different in the Nordic Countries (Laslett, 1988b, p. 16). Furthermore, not all siblings would have been equally likely to be of help: younger brothers and sisters were probably more likely to offer support than older siblings suffering from old-age impairments. Siblings also had to live nearby: Birgitta Odén has pointed out that many immigrants in late-medieval Stockholm lacked a kin network (1987, pp. 32, 38). This continued to be so until modern times: even as late as 1890, almost two-thirds of the female industrial workers were not born in the city (Moring, 2015, p. 38). Nevertheless, it seems that seniors with siblings in their vicinity could often rely on them for some form of support: in the Swedish town of Sundsvall, it was not uncommon for unmarried men and women over 50 to live with their siblings (Högman, 1999, pp. 135–136).

Up to this point we have indicated the factors standing in the way of intergenerational support: not everyone could rely on a family safety net. Of course, this does not mean that no one could turn to relatives and in fact, in some regions it was common for old adults to live in three-generation households. This type of cohabitation was particularly common in the countryside: in Iceland (1801), 9.9%–46.2% of the households were three-generation households (Gunnlaugsson, 1988, p. 72). Cohabitation of parents, their children, and grandchildren was also quite common in rural Sweden (19%–41% of the households) (Gaunt, 1976, p. 53). In towns, extended family households were less common, though: in Stockholm, in 1800, only 17% of the old adults lived with relatives; the vast majority of these (89%) were women.¹⁹ More than four out of five old Stockholmers did not move in with next of kin. Few of them would have been able to depend on charity: only 10% of the capital’s old men and women received public support. This would probably also not have been desirable. To spend one’s final years living on poor relief was something social middling groups tried to avoid as much as possible, first because this carried a stigma of social decline, and second because the elderly poor risked ending up living in workhouses (Högman, 1999, pp. 142–143).

Given that poor relief was undesirable for most, and support from family and friends uncertain, how could seniors make ends meet? Cristina Carlsson Wetterberg, discussing old women in Malmö around 1900, indicates that they relied on a combination of pensions, wages, and assistance from poor relief institutions and next of kin (1988, p. 246; Boyer & Schmidle, 2009, pp. 255–256). For the pre-modern period, historians have suggested old women and men relied on a ‘mixed economy of welfare’, which included support coming from ‘individuals and families, neighbours and communities, mutual-aid organisations, charities and commercial organisations’ (Harris & Brigden, 2007, pp. 1, 8). In the historiography, the latter have mostly been identified with life insurance companies that emerged in modern times and that could provide a surviving spouse with a steady income. Similar constructions also existed in the pre-modern period, but these tend to have been overlooked. In fact, there were various ways to create steady income streams that served as pensions.

4. The third myth: Pensions did not exist

If people could not assume they would be taken care of by family during old age, and living off poor relief was undesirable, it might have been wise to try to secure a steady income. Old age pensions are understudied because of our inability to look beyond today’s labour versus leisure dichotomy. Pensions allow us to retire from the labour force and enter a ‘third age’ of leisure (Ehmer, 2015, p. 135). Such a sharp break between working life and retirement usually did not exist before the welfare state: most labourers could at most hope that earnings they lost because of old-age impairments could be compensated with a pension they drew from capital. Even a small pension equivalent to several weeks’ wages may have prevented impoverishment during the final years, provided that seniors continued to work as much as possible and could perhaps find some additional support from family and friends. We therefore propose defining the pre-modern pension as a regular payment in money or in kind made during a person’s old age financed by capital which that person has previously acquired.²⁰ This is a broad definition that avoids the assumption of a sharp break between working life and retirement, and references to modern institutions such as the state and pension funds, and thus allows us to study how capital could be used to keep living standards at the desired level before modernity. Only returns on capital invested in real estate or financial instruments provide ageing individuals with a fixed income regardless of how old they get; accumulated savings can be depleted and therefore do not offer protection against the possibility of outliving your accumulated capital.²¹ To get a fixed income equivalent to an annual income from work – as is common for modern pensions – requires a substantial investment, and therefore Peter Lindert believes that public pension systems did not emerge until recently because ’average working incomes were too low’ to allow people to save for full pensions (2009, p. 7). However, when we abandon the assumption that pre-modern seniors were looking to retire completely, and instead may have sufficed with small pensions supplementing other sources of income, saving for old age may have been attainable.

A second reason why pre-modern pensions have been overlooked is their association with saving. Since Max Weber published his The Protestant Ethic and the Spirit of Capitalism in 1905, saving has been considered ‘modern’ by social scientists, so it is often assumed that pre-modern people did not have the mentality required for saving. Yet there can be little doubt that people in the medieval and early-modern Nordic countries set money aside. Coin hoards from the Viking era have been associated, among other things, with life-cycle saving in preparation for marriage (Gruszcynski, 2019, pp. 170–171), and more in general ‘to create future spending power’ (Casey, 1986, p. 53). Such archaeological evidence does not reveal whether people also saved for old age, and if so, which social groups did so. The best way to study saving is through the administration of savings banks, such as Falu stads sparbank, established in 1825, which received deposits from copper miners and maids, who managed to save substantial sums (Lilja, 2004, pp. 218–220). But since saving banks emerged slowly in the Nordic Countries, it is difficult to study saving behaviour before modernity. Few pre-modern savers who wanted to profit from compound interest, or wanted to create a pension, could deposit their savings with banks. Instead, they might have invested directly in financial instruments, such as Swedish government bonds, which were held by ‘peasants and other low ranking persons’ (Ericsson & Winton, 2018, p. 9). Opportunities to invest in such financial instruments may, however, have been limited until Sweden issued large numbers of promissory notes (Riksgälds) and the Danish Crown sold more than 1,000 life annuities, both around 1800 (Winton, 2015, p. 17).

Another way to save for old age was to participate in early pension funds. Starting in the seventeenth century, the Swedish and Finnish military’s Vadstena krigsmanshuskassa (1647–1865) and Kungliga Amiralitetskrigsmanskassan (1642–1737) provided old age pensions. True life-cycle saving emerged when the former developed the Arméns pensionskassa (1757–1937), where officers contributed 6% of their wages to the fund (Eriksson, 2008; Petersson, 2014). In Denmark–Norway, royal servants contributed to the Almindelige Enkekassan (general widows’ pension fund) established in 1775, which provided widows with 30–50% of the annual wages of the deceased husband. Between 1775 and 1845, almost 15,000 men contributed to this pension fund. They were mostly state employees, but also civilians looking for saving schemes that helped them to arrange old age insurance (Fogsgaard, 2005, pp. 15–16, 63).

Lower social groups may have saved for old age by lending out money at interest. To do so was very common: the pre-modern world was thoroughly ‘financialized’, and the total sum of money lent and borrowed by people in all strata of society was comparable to the capital handled by banks in modern times (Hoffman et al., 2019, p. 1). This was also the case in the Nordic countries. Finnish probate inventories show that people were part of peer-to-peer lending networks where large and small savers used promissory notes to invest their savings (Dermineur 2021). In Gothenburg and Kalmar, such investments were made in wide credit networks exceeding family groups (Hallén, 2015; Lindgren, 2002). In the countryside, networks were smaller and savers often lent to relatives (Köll, 1983); this was even done by the landless, who could be responsible for up to 40% of the savings that were invested. This type of lending usually yielded interest at 5–6% (Lindgren, 2017; Perlinge, 2005) and may thus have helped provide small pensions. Nordic research on informal saving links up with claims that a rural working family in England could set aside money to prepare for difficult phases of the life cycle, such as old age. Opportunities to do so existed well before modern times: they were greatest in the fifteenth century, when living standards were high and families did not have to spend their entire income on necessities, and again in the eighteenth century (Horrell et al., 2022). Although more research is needed, it seems that the scope for saving for old age by means of lending out on credit also existed in the pre-modern Nordic countries.²² Such loans may also have forged social ties between creditor and debtor, creating ‘quasi-kinship’ relationships that could also be important during old age for seniors lacking a strong family safety net (Dermineur, 2022, p. 58).

Up to now we have discussed how financial institutions and financial instruments could help to create pensions. Real estate could also be used to create returns on capital that provided an income during old age. Gregory Clark suggested that landowners in England in the Middle Ages could expect returns of 10%, and as a result, one acre sufficed for a comfortable old age; however, returns fell to about 5% in the centuries that followed (2007, p. 189). Even though returns on land were not as high in the medieval Nordic countries (Dahlbäck, 1981, p. 68; Söderberg, 2019), renting out real estate and (partly) living off rents was quite common. This was also the case in towns: in Stockholm, about one-fifth of the population owned real estate in a comprehensive wealth tax survey compiled in 1715,²³ and in 1800 7% of the older population reported rental income from housing.²⁴ Given the high rental incomes reported, it is likely that our sources only cover income from renting out entire houses and apartments and that they do not include income from lodgers. Taking in lodgers, who would either pay for their stay, but could also contribute by doing chores, was quite common, especially during old age. Lodgers were very common until recently: around 1900, they were to be found in half of the working-class homes in Stockholm and in 60% of the houses of widows in Helsinki (Moring, 2015, p. 39, 2016, pp. 4–5).²⁵

We will now turn to pensions in kind. In the Nordic countries and elsewhere in pre-modern Europe, real estate was widely used to negotiate retirement contracts. These were formal arrangements for old age: property owners used their real estate to secure food and lodging for their remaining years. In the contracts, the older generation agreed to hand over real estate to the younger generation on condition of lifelong food and lodging; if the latter was not provided, the elderly men and women could ask a court of law to nullify the contract and have the real estate returned to them. Retirement contracts existed throughout the Nordic countries (Blom, 2000, pp. 39–40; Dribe & Lundh, 2005, pp. 182–183; Gaunt, 1983, p. 251; Kristensen, 1965, pp. 241–242; Lindberg et al., 2017, pp. 160–161; Lundh & Olsson, 2002; Pettersen, 2007, pp. 75–76; Stephensen, 2002, pp. 2, 12). These already emerged in the Middle Ages: a Finnish example of syytinki/sytning dates from 1319, and retirement contracts continued to be used there for centuries. In fact, nineteenth-century Finnish politicians claimed that such contracts caused ‘laziness and avoiding work’ and therefore suggested a minimum retirement age (Hannikainen & Vauhkonen, 2012, pp. 18, 20–21). This was, of course, largely political rhetoric: in most cases, retirement contracts concerned parents handing over the family farm to a child while creating a ‘family insurance strategy’ that guaranteed everyone a decent living (Moring, 2003, p. 83). By recording a contract, seniors protected themselves against possible future conflicts, for instance when they grew so old they could no longer help out.

Leisure may have been a more important motive when retirement contracts were agreed with non-kin. Property owners without offspring, or without offspring willing and able to take over the farm and take care of parents, could offer real estate to non-kin in exchange for food and lodging. In Iceland, this was only allowed if relatives were unable to take care of kin and poor relief institutions also could not help out (Stephensen, 2002, p. 12). Yet agreements between unrelated parties were not uncommon: in South Sweden, 22% of land transfers between non-kin recorded in 1766–1799 included retirement contracts (Dribe & Lundh, 2005, pp. 174–175; Ågren, 2009, pp.116–117). And although most attention has gone to retirement contracts in the countryside, these existed in towns as well. Birgitta Odén discusses the Stockholm widow Lucia Clementsdotter who in 1415 handed over real estate to Johan Sigridsson and his wife Kristina on condition she would receive ‘welfare with food and beer and a room to lie in as long as I live’ (1987, pp. 40–44).

Perhaps, the clearest example of financing a leisurely old age consists of old men and women entering a corrody contract which gave them the right to spend their final years in a hospital or other institution. After handing over real estate or paying an entry sum, they received the right to lifelong food and lodging in a bed or bed box in a hall, or in a small apartment. Religious institutions in Stockholm admitted ingifter or provent and Odén explains that ‘monasteries came to function as retirement homes for those who could pay for their upkeep’ (1987, pp. 36–37). This was an old tradition: corrodians were already being admitted to Bergen, Norway’s Munkeliv monastery and Trondheim’s Nidaros Cathedral chapter before the Black Death (Pettersen, pp. 35–43). Corrodians also existed elsewhere in the Nordic countries: Danish monastic hospitals sold corrodies (proventkjøp) (Brøgger Daugaard, 1830, pp. 321, 383;, pp. 282; Ehlers, 1898, pp. 1–2) and the Finnish Nådendal nunnery admitted several corrodians immediately after its foundation in 1438 (Leinberg, 1890, p. 310).

Corrodies could be quite affordable because institutions offered different types of retirement for a range of prices. In 1638, Stockholm’s Danviken’s hospital sold 16 fribröder contracts (Unger, p. 113) and a sample collected for this article²⁶ shows that the institution sold 10 fribröd contracts in 1668, usually to an older man and his wife or a mother and her child, but sometimes even for two siblings. These fribröd contracts provided lifelong food and lodging without expectation that the recipients would perform any work at the hospital (Unger, 1996, pp. 75, 208). Fribröder thus received a full pension. On average, they paid the equivalent of 340-day wages²⁷ of an unskilled male worker, and these corrodies were purchased by the better-off, such as a former alderman, a farmer, and the widows of a councilman and a tailor. Another 58 corrodies were fattigbröd contracts, which provided food and lodging of a lower quality on condition that the recipient also helped out performing labour duties (Unger, 1996, pp. 75, 117, 208). These corrodies were less expensive at around 68-day wages of an unskilled labourer.²⁸ Fattigbröd contracts did not allow a ‘third age’ of leisure but did prevent a fall in living standards during old age.

5. Conclusion

Population ageing is likely to lead to debates about social transfers. To prevent policy-makers from dusting off myths about old age, it is important to identify and debunk false beliefs through new research. We have formulated three myths about old age before modernity and have tried to indicate why these rest on false assumptions. We see a transitional period between full-time work and death during which people were forced to gradually reduce their labour input: although some retired fully, most only partly stopped working or took on less demanding tasks. Not everyone could expect to receive sufficient support from their next of kin. To complement declining earnings and prevent impoverishment, various techniques were available to the older generation that let them use their capital to secure pensions that were the equivalent of weeks, months, or even a full year’s work.

We have suggested a functional approach to old age before modernity and believe that future research should focus on how labour, family, and finance may have contributed to maintaining living standards during the final life stage. With respect to labour, we still know far too little about the end of premodern careers: at what age were farms and workshops handed over to the young and what did doing so mean for the old? Concerning the family safety net, a main challenge is to find out whether independently-living older adults could muster support from family and friends.²⁹ And with respect to finance, it is important to find out more about the financial techniques and investment opportunities that were available to the general population and could be used to create pensions.³⁰ We have tried to illustrate these issues by focusing on the premodern Nordic Countries, but believe that our research agenda might also be helpful for scholars of other areas. The share of the population that reached ages over 60 is unlikely to have been unique to the North of Europe,³¹ which means that people elsewhere also faced the question of how to prepare for old age. The role of family during the final life stage may have differed across regions: David Reher (1998) has argued for pronounced differences between the more ‘individualistic’ North and more ‘collective’ South of Europe. Likewise, the availability of financial instruments allowing for the creation of pensions is likely to have varied between regions and in town and countryside. Such differences lend themselves to comparative research designs that can help to improve our understanding of the variables that shaped the position of old adults in premodern society.

Ageing is a universal experience that we share with our ancestors. Like us, they were likely to experience growing older and spending their final years as old adults. Like us, they might have anticipated, or dreaded, the process of ageing. And like us, they might have cared for old relatives when they were young, and depended on young relatives when they were old. The history of old age provides us with a mirror to look back to make sense of our own lives, but it also influences how we look forward to the coming decades of population ageing. The three myths we have formulated about old age before modernity are part of lay theory and as such not harmless but connected to current discourses about ageing. Lay theory also influences formal theory by guiding the questions scholars ask, or do not ask, about old men and women in the past. We hope to have demonstrated that there are still questions to be answered – and myths to be debunked.

Acknowledgments

We thank two anonymous referees, Helene Castenbrandt, Tim Riswick, and audiences in Gothenburg, Malmö, Lund, Utrecht, and Rome for their comments on the paper.

Disclosure statement

No potential conflict of interest was reported by the author(s).

Additional information

Funding

Jaco Zuijderduijn acknowledges the support by Vetenskapsrådet [2017-01857].

Notes

1. Cf. Burgess’ use of modernisation theory to create a distinction between the ‘premodern’ and ‘modern’ in his influential book on ageing, published in 1960 (Burgess, 1960, p. 3).

2. Other concepts of old age include ‘cultural age’ (how did individuals perceive their ageing bodies and minds?) and ‘social age’ (how did others look at old men and women?) (Thane, 2000, pp. 4–5). Cf. Examples of books on old age starting from a cultural history perspective include (Ottaway, 2004; Troyansky, 1989).

3. Eg. Jones and Batchelor (2015) about the representation of old adults in popular culture.

4. ‘ … retirement is not solely a phenomenon of modern society … ’, and ‘ … the extended family household of traditional society is not universal but varies significantly over time and space … ’ (Quadagno, 1982, p. 22).

5. We use indicators for economic development as an indicator for modernization: after c. 1850 Nordic economies began to modernize and converge with other European economies (Ljungberg & Schön, 2013, p. 106).

6. For instance, present research by Ulrika Lagerlöf Nilsson and others in the AgeCAP network, and the ’Äldre i samhället -förr, nu och i framtiden’ project in Lund, Gothenburg and Uppsala’ in the 1980s.

7. The other reasons he mentions are increasing wealth and the rise of democracies (Lindert, 2009, p. 7).

8. Life expectancy (additional years) in 1750–1790 for men: 18.16 years (at age 50) and 10.02 years (at age 65). For women 19.59 (at age 50) and 10.51 (at age 65) (https://www.scb.se/en/finding-statistics/statistics-by-subject-area/population/population-composition/population-statistics/pong/tables-and-graphs/births-and-deaths/life-expectancy-17512023/; accessed 20-5-2024), Debiasi (2020), The historical origins, 22.

9. Tabeller over Folkemængden (1857) indicates that in 1801, 416 octogenarians were counted in Copenhagen, 628 in other Danish cities, and 6,102 in the countryside (which in 1801 included territories in Norway, the Faroe Islands and Iceland: Johansen, 2004, p. 26).

10. Överståthållarämbetet för uppbördsärenden, År 1800, SE/SSA/0031/06/G 1 BA/G 1 BA:26/4 (1800).

11. By 1800 there was no specific tax threshold, but people paid income tax (inkomst bevillning) until they left an occupation or simply could no longer pay as their earnings had dropped below subsistence level (Rydin, 1882, pp. 233–234).

12. At present the number of female income tax payers is too low to provide a reliable image.

13. We used HISCO classification to divide workers into manual and non-manual workers (https://datasets.iisg.amsterdam/dataverse/HISCO). Cf (Van Leeuwen et al., 2002).

14. Cf. the development of physical and mental conditions over the life cycle (Schäfer, 2022): pp. 2–4.

15. Cf. Lars Edgren’s study of Malmö craftsmen, who usually retired years before they passed away (1982, p. 41).

16. Infant mortality rates were 21% in the Denmark countryside (1740–1801), c. 20% in Sweden and Finland (1750–1800), and c. 35% in Iceland (1838–1850) (T. Bengtsson & Lundh, 1999, p. 3; Johansson, 2002, p. 36; Tomasson, 1977, p. 413).

17. Losing the ability to function normally during old age is for instance demonstrated by osteoarchaeological research into skeletons of Danish premodern populations, which finds support for the idea that older individuals suffered from ‘non-life-threatening degeneration’ (Parker, 2019, pp. 230–231).

18. Cf Blaikie (1999, pp. 31–33), who writes about a ‘popular mythology’ suggesting that old adults could rely on support from their offspring.

19. Data taken from the mantalslängder analysing the entire sample of 1,468 individuals, and the subgroup who were over 60 (114 individuals) (see appendix). A recent analysis of the Swedish town of Linköping between 1750 to 1800 shows that only few households consisted of more than two generations and the mean household size was low, decreasing over time from 5,4 in 1750 to 3,2 in 1800 (D. Lindström, 2020, p. 237).

20. This is an adaptation of the definition given by Bach: ‘a regular payment made during a person’s retirement from an investment fund to which that person or their employer has contributed during their working life’ (Bach, 2003, p. 144).

21. For instance, in Sweden it was possible to purchase care for a period of a year by making annual payments to an institution (Lindberg et al., 2017) but this exposed the recipient to the risk of running out of savings.

22. Cf. early-modern agricultural workers’ ability to save: Ågren, 2009, p. 107.

23. Tax records for Stockholm in 1715 suggest that 26% of workers, 40% of artisans and middle-income groups, and 54% of merchants owned some form of real estate (E. Bengtsson et al., 2022, p. 165).

24. Calculated using the sample of mantalslängder (see appendix).

25. Lodging was also common throughout the Nordic countryside: (J. Lindström, 2019; Moring, 2016, pp. 3–4; Jåstad, 2011, 51.

26. Riksarkivet, Justitiekollegiets och förmyndarkammarens protokoll 1668.

27. A day wage was 28 öre kopparmynt (Söderberg, 2010, p. 473). 32 öre was one daler kopparmynt so to get 300 daler kopparmynt (R. Edvinsson, 2010, p. 98) equal to the least expensive fribröd contract one would need to work ((300 × 32)/28) = 343 days.

28. 20 daler silvermynt for a fattigbröd contract corresponds to about 60 daler kopparmynt, to obtain 60 daler kopparmynt one needs to work ((60 × 32)/28) = 68 days.

29. One approach is discussed in Verbeke (2021).

30. Especially informal networks allowing for peer-to-peer lending, as discussed in Hoffman et al. (2019), might be analysed from the perspective of creating provisions for old age

31. Cf. comparable shares of people over 60 in premodern England (Wrigley et al., 1997, pp. 614–615) and eighteenth-century France and Italy (Mols, 1955, p. 214; Troyansky, 1989, pp. 8–9). Additional data for population over 65 in Burgess (1960, p. 465).

Appendices

Appendix. Sample of tax records (mantalslängd) of 1800

The 1800 mantalslängder cover 556 blocks in Stockholm. We created a sample using the random command in Excel, providing us with a new block at the top of the Excel spreadsheet every time we ran the command. We did this 11 times to get roughly 2% of the blocks, which resulted in us having a random sample of blocks that should reflect the city at large. Indeed, this generated blocks all over Stockholm: the first six mentioned in Table A1 were from a large area north of Lake Malar, then another two from the island now best known as Gamla stan, and finally three from Södermalm South of Lake Malar.

The social composition of the 11 blocks varied. Torkan and Wintervägen were not only inhabited by labourers but also by soldiers in the local garrison who were exempt from paying taxes and therefore did not have to declare income (Lundsjö, 1975, p. 47). Even though they did not pay, active soldiers have been included in Figure 1 because they worked and likely earned money. State employees were also exempted from income taxation and are likewise categorized as working in Figure 1 unless they were indicated as retired. The two districts in the old town (Erisichton and Cadmus) were among the more affluent: based on income taxes paid, Cadmus was the wealthiest block in Stockholm, populated by high-ranking state officials, merchants, and craftsmen, together with their servants and family. Another block, Åkermanen, was predominantly inhabited by a few wealthy farmers’ households, low-income farm workers, and servants. In contrast, Sparfven, Timsten and Kattryggen were home households where masters and apprentices worked in various crafts; median taxes paid in these blocks are among the upper half of the distribution. At the very low end, we have Ormen Mindre, Braxen, and Bonde sonen, which were primarily inhabited by manual workers, journeymen, farmworkers, spinners, and a few lower ranking merchants; the latter paid substantial median taxes, but their titles do not indicate very high social status. The findings from these 11 blocks provide an overview of the general development of how earnings developed as Stockholm males grew older.

Table A1. Blocks sampled of the 1800 mantalslängd..

Block	District/Parish	Number of observations	Median total tax per male head of household	Mean total tax per male head of household
Wintervägen större	Johannes	63	1.36	1.15
Torkan	Johannes	348	0*	1.63
Sparfven	Jakob	146	1.84	2.24
Åkermannen	Kungsholmen	75	1.46	7.71
Timsten	Kungsholmen	26	1.52	2.03
Braxen	Klara nedre	145	2.45	3.25
Erisichton	Staden södra och västra	151	2.44	4.25
Cadmus	Staden södra och västra	55	15.22	14.66
Kattryggen	Maria västra	133	2.77	6.58
Ormen Mindre	Maria västra	199	3.21	6.51
Bonden sonen	Katarina södra	103	2	3.4

Source: Riksarkivet, SE/SSA/0031/06/G 1 BA/G 1 BA:26 (Överståthållarämbetet för uppbördsärenden, År 1800).

We used the districts/parishes mentioned in the source. Districts/parishes are ordered from North to South.

* Median taxes paid is zero because of the large number of soldiers who did not have to declare taxes.

Ideally, we would be able to compare our sample to a robust estimate for the social structure of Stockholm in 1800. To get an impression of the social structure, we used a registry transcribed by the Stockholm city archive containing the occupational titles of all of the approximately 27.000 adult men in the mantalslängd in 1800. We merged the occupational titles with the HISCO codes which were created for the 1880s Swedish census. This is not an ideal approach because of differences in spelling and other shortcomings of merging the 1800 sample and the 1880 HISCO codes. The main issue we faced was that even though usually for master craftsmen apart from their occupational rank (mäster) their precise occupational sector is given (e.g. shoemaker). However, for many apprentices and journeymen (resp. lärling and gesäll), occupational sectors are missing, and, hence, we are unable to assign HISCO codes to them using the merge command in STATA. As a result, for 1800 we were able to automatically assign HISCO codes to 61% of all males older than 15; the remaining 39% could not be coded using HISCO. The alternative, coding manually, would have been too time consuming for the 26.909 males older than 15 in the 1800 mantalslängd. We did manually code the 481 individuals in our sample of 11 blocks (Table 1). Our inability to code many apprentices and journeymen in the 1800 mantalslängd probably explains why the share of manual labourers is lower there than in our sample of 11 blocks (Table A2): most people organized in the craft guild system were manual labourers. Although future research is necessary to improve robustness checks, for now, we suggest that our sample may be reasonably reflecting Stockholm’s male population over the age of 15.

Table A2. Manual and non-manual male labourers over 15 in the sample and in the 1800 mantalslängd.

	Sample of 11 blocks	Stockholm
Non-manual labourers	20.2%	38.2%
Manual labourers	79.8%	62.2%
Sample size	481	16.410

Source: Riksarkivet, SE/SSA/0031/06/G 1 BA/G 1 BA:26 (Överståthållarämbetet för uppbördsärenden, År 1800); Mantalsregister 1800–1884: https://sok.stadsarkivet.stockholm.se/Databas/mantalsregister-1800–1884/Sok?sidindex=0.