CONCEPTUAL BACKGROUND - Patterns of Inequality: Income Inequality, Homogamy and Social Mobility

5.1. Income inequality

Next, I discuss the conceptual background of this PhD thesis by exploring the theoretical frameworks and the central empirical evidence while considering the long-run patterns of inequality, i.e. income inequality, homogamy and social mobility. First, I discuss the central theories and empirical findings while considering the determinants of income inequality patterns, which are numerous, though they can be divided into the five following factors: 1) development, 2) shocks, 3) wealth, 4) institutions and 5) globalisation and technological change (Table 2).

The process of economic development was already being characterised as a determinant of long-run inequality patterns more than half a century ago by Kuznets (1955). The famous Kuznets curve theory implies that inequality will rise during the initial phase of industrialisation, whereas it will decline after the initial phase when society reaches a certain stage of development. During this phase, the level of productivity differs between urbanised/industrialised and rural regions, which creates pressures for wider income gaps. Also, the weight of the unequal urban regions will rise in comparison with the less unequal rural regions in the distribution, which results in a higher income gap as a default.

The Kuznets curve hypothesis is closely connected with the inequality possibility frontier, which characterises the first increasing part of the curve. Presented already by Kuznets (1955) and further elaborated by Milanovic, Lindert, and Williamson (2011), their arguments state that when average income in a society is close to the subsistence level, it constrains inequality. The idea is that an extreme level of inequality is not possible since the poorest people would struggle to make a living and perhaps would starve to death. Thus, a low average income level produces an inequality possibility frontier, whereas this limitation eases along with a rise in incomes. Moreover, some scholars argue that societies will confront these types of Kuznets curves repeatedly over the course of history due to globalisation, technological change and other forces that rapidly transform societies (Milanovic 2016; Lakner and Milanovic 2016). However, prior empirical studies have yielded quite controversial findings with respect to the Kuznets curve hypothesis (Ryckbosch 2015; Rossi, Toniolo, and Vecchi 2001; Atkinson and Piketty 2010).

Scheidel (2017) has suggested a contrary hypothesis that the only factor that can forcefully interrupt increasing inequality is a period of turbulent shocks and violence. Indeed, shocks have recently begun receiving key attention by scholars observing the decline in inequality during the first part of the 20^th century (see massive empirical work examining the top income shares in Piketty 2014;

Atkinson and Piketty 2010; Piketty and Saez 2006; Atkinson, Piketty, and Saez 2011; Atkinson and Piketty 2007). Top income shares decreased considerably due to the two world wars and their aftermath, but evidence also suggests that shocks played a significant role in the long-run history of income inequality as well (Piketty 2014). For instance, several recent long-run studies have examined famines (Alfani and Gráda 2017), wars (Scheidel 2017) and financial crises (Morelli and Atkinson 2015). However, as Morelli and Atkinson (2015) have highlighted, relatively few studies exist on different types of shocks and crises, and therefore, the impact of varying crises (e.g. banking crises, financial crises, inflation) on income inequality is not well-known.

I. Development

a. Kuznets curve (Kuznets 1955)

b. Inequality possibility frontier (Milanovic, Lindert, and Williamson 2011;

Kuznets 1955) II. Shocks

a. Famine (Alfani and Gráda 2017) b. Wars (Scheidel 2017; Piketty 2014)

c. Other crises (see, for example Morelli and Atkinson 2015) III. Wealth

a. R-G theory (Piketty 2014)

b. Capital and labour shares (Bengtsson and Waldenström 2018) c. Wealth concentration (Piketty and Zucman 2014)

IV. Institutions

a. Redistribution through taxes and social transfers (Piketty, Saez, and Zucman 2018; Jäntti et al. 2010)

b. Government spending, including education (Lindert 2004; Atkinson and Bourguignon 2015)

c. Unions (Förster and Tóth 2015; Farber et al. 2021) d. Other (e.g. Atkinson 2015)

V. Globalisation and technological change

a. Skill-biased technological change (e.g. Acemoglu 2002).

b. Globalisation (Roine, Vlachos, and Waldenström 2009; Bartels 2019) Table 2. The main determinants of income inequality patterns

Source: see text.

Moreover, researchers have established empirical as well as theoretical frameworks for evaluating wealth and capital incomes as a determinant of income inequality. In fact, there is strong empirical evidence that capital shares are one of the most important factors driving income inequality, especially for incomes at the top (Bengtsson and Waldenström 2018; Jäntti et al. 2010). This is not surprising since wealth is much more concentrated at the top in comparison with incomes. Therefore, the changes in concentrations of wealth have played a crucial role in income inequality patterns, and in many cases income and wealth inequalities share a common trend (Piketty 2014; Piketty and Saez 2006; Saez and Zucman 2020). The R-G theory proposed by Piketty (2014) has given rise to much debate among scholars (see Raval 2018; Mankiw 2015; Góes 2016). The main idea of the theory is to compare the return on capital (r) and economic growth (g): in this framework, income inequality rises if r is higher compared with g (and vice versa). In this case, the incomes derived from capital increase faster than labour incomes. Piketty (2014), together with several extensive empirical studies on GDP (the Maddison database, see Bolt and van Zanden 2020), has shown that the annual rate of return on capital (r) has been roughly 4–5 per cent, whereas the annual growth rate of world output (g) was close to zero before industrialisation. Remarkably, the growth rate trend (g) did not exceed the rate of return (r) until the first part of the 20^th century, when capitalists suffered capital losses, destruction and taxes. Moreover, the recent substantial empirical work done by Jordà et al. (2019) on the rate of return in varying asset classes confirms this previous finding. The rate of return was always higher than economic growth throughout the industrial period (1870–2015), except for the war years and their immediate aftermath. In fact, the gap between r and g was roughly 5% in the pre-World War II era, whereas it diminished to only 2% in the 1970s and has risen again to roughly 3–4% today.

Arguably, institutions play a large role in shaping income differences, however at the same time they are perhaps the most complex determinant and poorly understood. Institutions comprise a vast and diverse factor when measuring income inequality, but they can be divided into four subcategories. First, the role of the public sector in distribution is quite direct in modern welfare states through taxes and social transfers (Jäntti et al. 2010; Piketty, Saez, and Zucman 2018). In many rich developed countries, disposable income inequality (after taxes and transfers paid) is at least 40 per cent below market income inequality when measured by the Gini coefficient. Overall, the greatest redistribution of income occurs in the Nordic countries and Benelux countries as well as in Ireland, Germany, France and Austria (Morelli, Thompson, and Smeeding 2015). Moreover, in the longer run the imposing of higher income taxes on those in the top earning brackets hampers their accumulation of wealth significantly (Jäntti et al. 2010; Roine and Waldenström 2015). Second, the collected taxes enhance government spending efforts, which could increase the overall prosperity of the country and level off the income gaps, but eventually this is not always the case (Lindert and Williamson 2016b;

Lindert 2004). Government spending on human capital accumulation (especially education) is one of the most important means of diminishing the gaps between skilled and unskilled workers (skill-premium) (Atkinson and Bourguignon 2015). Third, prior studies suggest that unions have a significant role in equalising wage distribution, constituting a significant force in diminishing top income shares as well as capital shares (Förster and Tóth 2015; Farber et al. 2021). Fourth, the details are crucial regarding how institutions work and their impact on inequality (see e.g. the Finnish comprehensive school reform in Pekkarinen, Uusitalo, and Kerr 2009). Indeed, the Nordic model has been quite successful at adopting institutions that have supported economic growth as well as reduced inequality, however little is still known about what institutions and policies to adopt for the future.

Lastly, a wide variety of institutions can be highly extractive or inclusive in nature, which greatly affects inequality (Atkinson 2015; Acemoglu and Robinson 2012; Roikonen, Ojala, and Eloranta 2021; Koivunen, Ojala, and Holmén 2021; Bengtsson, Olsson, and Svensson 2021). Therefore, there is a great need for further research that investigates various institutions and studies their impact on inequality (Förster and Tóth 2015, p. 1801; Salverda and Checchi 2015).

During the last few decades, some researchers have pointed out that increasing globalisation and technological change have significantly contributed to inequality. In theory, increasing trade and global competition will increase the gaps between skilled and unskilled workers, especially in developed countries. Moreover, current skill-biased technological change together with globalisation increases demand for highly skilled labour, with rapidly increasing wages being paid to them and not to low-skilled workers (Acemoglu 2002). In addition, recent rapid technological progress has created global big tech companies that have a certain monopolistic power, which creates ‘rents’ and increases their profits and their employees’ wages. Also, rapid technological change can often create superstar effects (winner take all) (Koenig 2019; Kurz 2017). This phenomenon can eventually skew the prevailing social norms regarding earnings due to herd behaviour (Piketty and Saez 2003, p. 35).

Empirical evidence, however, is inconclusive on whether globalisation and technological change have played a significant role in recent income inequality trends (Roine and Waldenström 2015; Roine, Vlachos, and Waldenström 2009; Rossi, Toniolo, and Vecchi 2001). For example, it is difficult to explain why massive differences in income inequality levels and trends persist between countries. As argued by Atkinson and Bourguignon (2015), the ‘race’ between education and globalisation/technology is one of the reasons for differences between countries. This race can be observed within a simple supply (education) and demand (globalisation/technology) framework.

Moreover, the theory is far from conclusive since contrary empirical evidence exists that more open market economies will cause severe problems with local monopolies, which diminishes top income shares (Bartels 2019; Roine, Vlachos, and Waldenström 2009). In addition, some scholars have

argued that there is less room for (gender) discrimination in highly competitive markets since companies suffers greatly if they do not hire the most talented and skilled labour (Ponthieux and Meurs 2015). Therefore, it seems that skill-biased technological change and globalisation should be investigated together with labour market characteristics and educational systems rather than as an individual determinant.

5.2. Homogamy

The theories and the empirical foundation for considering the factors that influence assortative marrying patterns are relatively well-established, and we can characterise three main forces: 1) third-party influence, 2) individual preference and 3) structural constraints (see Table 3). Third-third-party influence, including parents, peers and institutions like religious authorities (i.e. the Church), restrained and tried to limit heterogamous marriages. Indeed, family and community had supreme control in many pre-industrial societies as opposed to individual preferences (Moring 1999; van Leeuwen and Maas 2010; Saarimäki 2010). While contractual marriages did not exist in Finland, marriages were regulated through official and unofficial norms and traditions as well as through regulations by the state, the Church, local community and family (Häkkinen 2018, p. 110–115).

Moreover, there was always the possibility that parents would disinherit their children if the children did not take their parents’ opinion into account when selecting a spouse (Moring 1999). The power of parents to influence spouse selection diminished if spouses migrated to a distant region or if their parent(s) were no longer alive at the time of the marriage (Sherkat 2004; Marco and Ineke 2002).

According to modernisation theory, industrialisation and modernisation gave new possibilities to earn a living and increased personal autonomy, which also opened new marriage possibilities. Along with this development, the need for family support (elderly and children) and their role in the marrying process diminished since states adopted new social policies and adult mortality rates declined. In addition, the abandonment of old traditions in the marriage markets reduced the direct role of the Church and the local community (van Leeuwen and Maas 2010, 2019).

The role of individual preference in choosing a spouse is somewhat unclear, however it has been argued that people tend to marry homogamously for cultural reasons. Spouses with the same social background largely share similar preferences and tastes, and they may then adapt to marital life more easily (Kalmijn 1994; Dribe and Lundh 2005; van Leeuwen and Maas 2005). Also, evidence from biographical and proverbial texts from Finland suggests a belief that spouses should have similar status in the lower social classes for the marriage to be prosperous and happy. In other words, cultural knowledge and a shared understanding about rural poverty impact marriage as well as the

intergenerational persistence of poverty (Stark 2011). In contrast, other evidence suggests that spouses preferred partners from an esteemed and prosperous background (Kalmijn 1994; Dribe and Lundh 2005; van Leeuwen and Maas 2005). According to Häkkinen (2018), the common age for marriage was roughly 21–25 for men and 19–23 for women, however the numbers could vary significantly in individual cases. The legal age to marry was 15 for women and 21 for men according to legislation passed in 1734. Although, the third-party influence was strong and it was obligatory to obtain parents’ permission to marry, the individual preferences of spouses were taken into account and potential spouses could meet at local dances or evening plays (Moring 1999, p. 175).

Nevertheless, according to the romantic love hypothesis, the importance of love become more evident in marriage patterns only during the industrialisation phase and later as a result of 20^th-century modernisation processes (van Leeuwen and Maas 2010, 2019).

Furthermore, the structural factors impacting marriage markets (the pool of spousal candidates) play a great role. Structures occur when, in practice, the contexts for meeting one another favour individuals from a similar background (Matthijs 1998). Matthijs and Henk (2001) and others have noted that various institutions shape the odds of meeting someone from a similar background and therefore impact the persons with whom we form personal relationships. The odds of meeting dissimilar people become greater in modern society, which diminishes the possibility for homogamous marriages. Such settings include, for example, more diversified work, school and neighbourhood settings, family networks and voluntary associations (Matthijs and Henk 2001). Thus, encouraging and establishing more diverse meeting places for individuals to meet and form relationships increases the likelihood for heterogamous marriages. Evidence exists, for example, that a diverse labour market (with more women participating), greater levels of migration and mobility (e.g. improved transportation and communication) as well as expanded schooling opportunities all enhance the pool of possible spouses and create more possibilities for heterogamous marriages (van Leeuwen and Maas 2010, 2019, 2005; Bras and Kok 2005; Bras 2004). On the other hand, the poor at times experienced direct marriage restrictions, the purpose of which was to prevent them from reproducing (Seiler 2019).

According to cross-cutting theory, people belong not only social classes but to many other groups as well, such as various local, ethnic, language and religious entities. The presence of large numbers of persons with multiple social origins broadens the spectrum of origin and makes marrying more difficult from a strictly homogamous standpoint (van Leeuwen and Maas 2019, p. 4). This fact has been empirically documented in the religiously diverse country of Hungary between 1870 and 1950 (Lippényi et al. 2019). It is also possible that varying marriage strategies occur in differing social groups. For instance, diminishing social homogamy during the modernisation period in

Belgium (1821–1913) was similarly accompanied by strengthening linguistic boundaries between Dutch and French-speaking communities (Lippényi et al. 2019).

Table 3. Theories of assortative mating in the past.

Source: see text.

Most studies state that marriage homogamy has remained surprisingly stagnant throughout history, but certain other studies have highlighted the role of modernisation in facilitating heterogamous marriages (van Leeuwen and Maas 2010; Dribe, Eriksson, and Scalone 2018; Dribe and Lundh 2009;

Bull 2005). The differing results might indicate significant differences between various social statuses and dynasties (elite vs. common statuses) (Jallinoja 2017). Nevertheless, marriage patterns in many areas remain unknown, and only a limited number of empirical works have examined the long-run trends. This seemingly reflects the situation of there being so few income inequality studies looking at society before the 21^st century or so few social mobility studies in general (see the review articles:

van Leeuwen and Maas 2010; Morrisson 2000).

5.3. Social mobility

There is a common understanding among scholars that social mobility is more or less stable over time, although some differences do exist between countries and time periods (e.g. Erikson and Goldthorpe 1992; van Leeuwen and Maas 2010; Xie and Killewald 2013; Maas and van Leeuwen 2002). Although, social mobility is seemingly higher in modern times compared with the Middle Ages or Antiquity, the literature is inconclusive on whether industrialisation and modernisation are significant factors for changes in social mobility (Kaelble 1981; Wiebke, Ineke, and Marco 2015;

Maas and van Leeuwen 2002; Dribe, Helgertz, and van de Putte 2015; van Leeuwen and Maas 2010).

This finding was long ago proposed by Lipset and Zetterberg (1959), who found similar levels of mobility among different social classes in industrialised societies (Table 4). Several decades later Featherman, Jones, and Hauser (1975) published results from the US and Australia supporting this claim.

I. Third-party influences: modernisation theory II. Individual preferences: the romantic love hypothesis

III. Structural constraints: institutions and the theory of cross-cutting circles

The common belief is that there is no room for significant changes in social mobility in capitalist market economies. These arguments rely on common characteristics of a market economy, like the fact that managers will only hire the most talented and motivated people due to market forces, i.e. the need for efficiency (see the review article by van Leeuwen and Maas 2010). In addition, Friedman (1962), has stated that market forces level out the playing field and increase social mobility in comparison with non-market economies (see also Featherman, Jones, and Hauser 1975; Becker and Tomes 1986). Similarly, other researchers argue that gender inequality will diminish along with the advent of the free market economy (Ponthieux and Meurs 2015). Recently, however, Basu (2017) has presented a theory suggesting that markets create ‘a focal point’ that enhances discrimination.

Furthermore, the British Industrial revolution, more precisely the destruction of hand spinning, has been presented as one example of a malign process in the market economy, one which led to large-scale female unemployment (Humphries and Schneider 2021). Humphries and Schneider (2021) argue that it played a crucial part in the formation of and expanding ideas about the family, contributing to the emergence of the male breadwinner family model. Therefore, the Basu’s theory as well as Humphries and Schneider’s example suggest that free markets and technological development can be quite unequal and discriminatory in nature as well (see also the discussions on possible credit constraints and poverty traps by Piketty (2000)). However, the exact role of market forces in relation to other possible factors remains unclear in this process.

Table 4. The determinants of social mobility

The finding that social mobility generally, though, does not vary much over time has been supported in studies by Clark and colleagues, who used elite surnames as a source (Clark, Leigh, and Pottenger 2020; Clark and Cummins 2015; Clark et al. 2015). They conducted the studies for a number of time periods and in such countries as the US, the UK, China, Australia and Sweden, establishing that social mobility is constant due mainly to genetic reasons and the fact that the correlation between generations is roughly 0.75, which they term ‘the law of social mobility’. In other words, assuming an intergenerational correlation of 0.75, it takes roughly ten generations for elite families to

I. Social mobility is more or less stable over time in market economies

a) Market forces (Lipset and Zetterberg 1959; Featherman, Jones, and Hauser 1975) b) Genes (Clark et al. 2015)

II. Families’ cultural capital (Georg 2004)

III. The same determinants explain equal opportunities (income persistence) as explain income inequality (Corak 2013; Solon 2004)

experience enough social mobility that they would be classified as middle-class families. However, as noted by Pérez (2019), these methodologies capture only the persistence of elite status across surname groups, and therefore, they differ quite a bit from regional or country-level estimates.

Often, it is impossible to separate the impact of genes from a family’s cultural capital. While the existing literature on cultural capital among families is vast, the strong persistence of such capital is due to one simple fact: family is fundamentally the most important factor for transmitting human and material capital to the next generation (Georg 2004; Jallinoja 2017). In principle, families provide their descendants with the foundation for success or failure in life, such as skills, social networks, orientations and preferences, values and attitudes as well as jobs, properties and cash (e.g. Corak and Piraino 2011; Georg 2004; Stark 2018). Indeed, acquired family commodities (education, occupation or incomes) are enhanced or diminished by family culture, which consists of skills, talents, learning

In document Patterns of Inequality: Income Inequality, Homogamy and Social Mobility in Finland, c. 1700–2019 (sivua 46-57)