2.1. Income inequality
This PhD thesis investigates the long-run patterns of inequality through the lens of income inequality, homogamy and social mobility. The first theme considers income inequality, a subject warranting increasing attention during the last few decades; however, as Lindert and Williamson (2016a) point out, not until now has a new wave of studies substantially advanced the frontiers of research. They also highlight that we need to explore new data from varying periods and countries to find plausible explanations for inequality patterns since regression models are almost useless due to the scarce amount of available data (see the literature reviews in e.g. Morrisson 2000; Atkinson, Piketty, and Saez 2011;
Atkinson and Piketty 2010; Scheidel 2017; Lindert and Williamson 2016b, a; Piketty 2014). Also, many researchers highlight the importance of conducting thorough case analyses to identify such plausible explanations (e.g. Acemoglu and Robinson 2015).
Thus far, research on income inequality in Finland has concentrated on the latter part of the 20th century, and efforts to find explanations for income inequality patterns have focused largely on the 1990s or early 2000s (e.g. Riihelä and Tuomala 2019; Riihelä and Suoniemi 2017; Riihelä, Sullström, and Suoniemi 2008). Fortunately, several studies have assessed income inequality over longer historical time spans: Jäntti et al. (2010) and Jäntti (2006) have investigated income inequality starting from the early 1920s, Hjerppe and Lefgren (1974) have in a pioneering study observed income inequality patterns between the years 1881 and 1967. Moreover, Roine and Waldenström (2015) have estimated top income shares between the 1860s and 1880s. Therefore, there are some studies to build on when exploring income inequality patterns starting from the year 1865. In addition, several studies have examined shorter time spans or have adopted a region-specific focus (e.g. Nummela 1990;
Nummela and Laitinen 1987). Moreover, a vast body of literature exists in the social history field on economic and social inequality, but it covers more specific variables contributing to inequality, such as ethnicity, region, social class or poverty (e.g. Pulma 2012; Peltonen 1992; Haatanen 1968; Siipi 1967; Häkkinen and Peltola 2005).
However, the long-run patterns of income inequality remain unknown since prior studies are methodologically inconsistent and diffuse. For instance, it is not clear how income inequality developed in the late-19th century since Roine and Waldenströms’ (2015) results are not comparable to the estimations made by Jäntti and colleagues (Jäntti 2006; Jäntti et al. 2010). In addition, no estimations exist for the years between the early 1880s and the year 1920. Moreover, some identified
19
patterns found are difficult to explain from an historical standpoint, which demands further elaboration.
Explanations for such patterns are almost completely lacking and the historical context is missing.
Therefore, the research questions regarding income inequality (articles I and II) are as follows:
1. Article I) How did income inequality develop in Finland between 1865 and 1934, and how does it compare with patterns in other countries?
2. Article II) What were the trends and levels of income inequality in Finland between the years 1865 and 2019, and how do the trends compare with those in other countries?
3. Articles I and II) What are the central determinants affecting income inequality patterns?
What does the Finnish historical context reveal when considering the possible drivers of inequality?
Thus, in articles I and II we constructed an income inequality series and knowledge on how income inequality developed between the years 1865 and 2019 by collecting and utilising income taxes and modern income distribution statistics. Distributions of the whole population were utilised when possible, but in many cases only top income shares were used due to a lack of data from the lowest income ranks.1 To construct the income inequality series, we used a common and recently developed method for reconstructing distributions (Blanchet, Fournier, and Piketty 2017). In addition, cross-country comparisons were also made. Furthermore, the articles contribute to discussions on the determinants of income inequality in the long run. The determinants include theories connected with development patterns, such as the Kuznets curve (Kuznets 1955) and the inequality possibility frontier (Milanovic, Lindert, and Williamson 2011); with shocks, such as famines, wars and recessions (Alfani and Gráda 2017; Piketty 2014; Scheidel 2017; Morelli and Atkinson 2015); with wealth, such as the R-G theory (Piketty 2014), capital and labour shares (Bengtsson and Waldenström 2018), and wealth concentration (Piketty and Zucman 2014); with institutions, such as redistribution through taxes and social transfers (Piketty, Saez, and Zucman 2018; Jäntti et al. 2010; Atkinson 2015), government spending (Lindert 2004; Atkinson and Bourguignon 2015; Atkinson 2015), and unions (Förster and Tóth 2015; Farber et al. 2021); as well as with globalisation and technological change (Acemoglu 2002;
Bartels 2019; Atkinson and Bourguignon 2015; Roine, Vlachos, and Waldenström 2009).
1 Empirical evidence suggests that top income shares are good substitutes for other income inequality measures, such as the Gini coefficients (Morelli and Atkinson 2015).
20
2.2. Homogamy
Homogamy in marriage patterns means that spouses have a similar social background in terms of, for instance, family occupation or level of education. The research on homogamy lacks consistent studies on long-run patterns, though. Trends identified in the US suggest that educational homogamy first decreased in 1940–1960, followed by an increase between the years 1960 and 2003 (Schwartz and Mare 2005). Moreover, some evidence indicates that educational homogamy has strengthened in many countries during the last few decades (Katrňák and Manea 2020). Interestingly, educational homogamy in Finland today is relatively modest in comparison with other countries (Mäenpää 2015).
Although the expansion of stratification studies during recent decades has shed light on this complex phenomenon, the roles of modernisation, individual preferences and structural factors in homogamy are still largely unknown (see the review articles: van Leeuwen and Maas 2019, 2010, 2005).
Moreover, studies focusing on historical perspectives of homogamy in Finland are relatively scarce.
Several researchers have studied homogamy and marriage in Finland in the 18th and the 19th centuries, identifying strict marriage patterns in pre-industrial society, especially in the landowning class, with the main reasoning for marriage being economic rather than love (e.g. Moring 1999, 2009; Silvasti 2001). Moreover, some scholars have focused on the history of marriage and sexuality in general (e.g.
Nieminen 1993). Other studies highlight the role of marriage as an institution that maintained or even strengthened social divisions in Finland. Such studies also discuss how the impact of family eventually declined in favour of individual decisions about marriage in the 20th century due in no small part to changes in the modern legal system (e.g. Pylkkänen 1991; Miettinen 2012; Kietäväinen-Sirén 2015). In addition, a valuable study by Jallinoja (2017) examines statuses and dynasties in the upper stratum by analysing marriage patterns and social mobility in varying status hierarchies.
However, proper statistical analyses of historical homogamy patterns between occupational ranks in Finland are still quite rare (see the pioneering work by Häkkinen 2018). Therefore, the research questions presented in article III are as follows:
4. How much homogamy, based on social background, was there in Finland between 1700 and 1910, and how did it change over time?
5. How can we explain the variations in homogamy?
In the article, we study social homogamy in first marriages in detail by comparing the parental occupations of both spouses, in other words, their social origins. We selected this approach not only because it is common practice in the field, but also for practical reasons: the wife was often assigned her husband’s occupation in marriage (e.g. van Leeuwen and Maas 2019). Our empirical work relies
21
on data from population and church registers, which consist of data mainly from southern and western parts of Finland. We constructed variables and regression models to explore the roles of modernisation, individual preferences and structural factors in marriage patterns (see van Leeuwen and Maas 2019, 2010, 2005; Bull 2005; Dribe and Lundh 2005).
2.3. Social mobility
A common view found in the literature is that, at least in market economies, social mobility has remained more or less stable through history. Although differences occur between countries and time periods, scholars suggest social mobility was generally rather stagnant (e.g. Erikson and Goldthorpe 1992; van Leeuwen and Maas 2010; Xie and Killewald 2013; Maas and van Leeuwen 2002). Moreover, such studies are inconclusive about the extent to which industrialisation and modernisation affected social mobility (Kaelble 1981; Wiebke, Ineke, and Marco 2015; Maas and van Leeuwen 2002; Dribe, Helgertz, and van de Putte 2015; Lipset and Bendix 1959; Kaelble 1986). One explanation for the
‘constant flux’ in market economies as compared to non-market economies is the level of efficiency and competition prevalent in market economies, which evens out the playing field and increases opportunities (Featherman, Jones, and Hauser 1975; Friedman 1962). Moreover, stagnant social mobility is also explained by families’ cultural capital or even by genetic inheritance (Clark, Leigh, and Pottenger 2020; Clark and Cummins 2015; Clark et al. 2015; Georg 2004).
Strikingly, these theories and arguments rely on only a relatively minor amount of empirical evidence and often lack long-run time series. Often the studies account for only a brief period, a restricted region or only part of the population, such as elite families (Clark and Cummins 2015; Clark et al. 2015; van Leeuwen and Maas 2010). During the past decade, however, some studies have started to question the prevailing wisdom after finding considerable changes in social mobility patterns in history (Long and Ferrie 2013b, 2013a; Modalsli 2017; Song et al. 2020; Pérez 2019). Since the longer term trends remain mostly unknown, revealing the long-run patterns of social mobility in Finland constitutes a significant contribution to the current body of literature on the topic. Numerous empirical studies exist from Finland that consider the intergenerational transmission of socio-economic status
22
(including education,2 health,3 incomes,4 family formation5 and disadvantage6), however comprehensive long-run studies are quite limited, and existing studies mainly concentrate on the period after the 1950 census (e.g. Erola and Moisio 2006; Pöyliö and Erola 2015; Erola 2009; Häkkinen 2018;
Sirniö, Martikainen, and Moisio 2011; Jarvela 1983). Thus, the research questions presented in article IV are as follows:
6. How did parent-son social mobility patterns develop in Finland, c. 1700–2000?
7. How do these patterns compare with those in other countries?
8. What were the main drivers of social mobility in the long run?
Article IV explores the intergenerational persistence of occupation in Finland from the early 18th century to the 21st century. In other words, it investigates the influence of parents on their sons’
occupational attainment (relative mobility) and how often sons pursued the same occupation as their parents (absolute mobility). Moreover, it compares Finnish data with data from countries, where comparable studies exist: the US, the UK and Norway. To our knowledge, our article is the first to extend the time frame of long-run social mobility literature to cover the 18th century utilising common empirical methods. Moreover, it covers the longest time span ever studied, if the studies on elite families by Clark et al. (2015) are not taken into account. Lastly, recent papers suggest that intergenerational income persistence varies significantly between regions and is heavily influenced by the same forces that determine income inequality (Corak 2013; Solon 2004; Jäntti and Jenkins 2015;
Chetty, Hendren, Kline, and Saez 2014). This proposition, however, does not imply that social mobility is influenced by the same forces. Thus, one of our interests is to discuss development and institutional (mainly educational) as well as income inequality patterns together with social mobility trends.
2 See Erola (2009); Hertz et al. (2008); (2016); Kivinen, Hedman, and Kaipainen (2007).
3 See Kröger et al. (2018); Remes, Martikainen, and Valkonen (2010); Tarkiainen et al. (2014).
4 See Airio and Niemelä (2009); Jäntti et al. (2006); Jäntti and Jenkins (2015); Pekkala and Lucas (2007).
5 Raab et al. (2014); See Erola, Härkönen, and Dronkers (2012); Nisén et al. (2014).
6 Social assistance (e.g. Ilmakunnas 2018), school dropouts and unemployment (e.g. Vauhkonen et al. 2017), and labour market outcomes (e.g. Sirniö, Kauppinen, and Martikainen 2017b).
23