Support for Democracy in the Age of Rising Inequality and Population Aging

Abstract

This article argues that advanced democracies face a dual crisis of rising inequality and population aging, which jointly put pressure on welfare regime sustainability. Particularly vulnerable are young people and the elderly in the context of these structural changes. As a result, young people and senior citizens in advanced democracies are increasingly focused on economic equality, which in turn undermines their support for democracy in its current form that prioritizes liberal economic policies. Survey data collected between 2010 and 2020 support the argument. Regardless of the political issues on which they differ, young and old appear to share similar views on what constitutes an ideal democracy.

Introduction

The world is facing a democratic crisis. Even prior to Brexit and the US presidential elections in 2016, there were already debates in academic and policy circles on a claimed new democratic recession (Plattner, 2015). The rise of populist leaders in advanced democracies showed that democratic regression was not only a phenomenon in unconsolidated democracies; it had spread to the core of the democratic camp (Diamond, 2020; Fukuyama, 2020). Expert surveys, such as the V-Dem project and Freedom House, have called attention to such a negative trend since 2017, and both institutions’ 2021 annual reports continue to cast a bad omen on democracy (Alizada et al., 2021; Repucci & Slipowitz, 2021). At the same time, economic inequality has also raised concerns among policymakers and scholars, especially after the 2008 financial crisis. Economists and political scientists have explored the causes of economic inequality and its link with the rise of populism in the West (Inglehart, 2016b; Piketty, 2014). Studies have also shown that economic inequality negatively affects people’s support for democracy across regions (Andersen, 2012; Huang & Zuo, 2022; Krieckhaus et al., 2014; Wu & Chang, 2019).

However, this decline in support for democracy is not equally distributed across the population. Foa and Mounk (2016) analyzed popular surveys and found that youth in Western democracies showed the lowest support for democracy. On the other hand, support for democracy is found to increase as one ages, but support begins to decrease when one approaches death (Lechler & Sunde, 2019). Based on these findings, it appears that the relationship between age and support for democracy is inverted U-shaped, with the very young and the very old showing a low level of support. Inglehart (2016b) attributed this low democratic support to existential insecurity, arguing that rising inequality has helped fuel older people’s support for authoritarian political candidates and young people’s dissatisfaction with democracy. While the youth’s dissatisfaction with democracy is understandable since they have suffered from unemployment and stagnant wages, the older generations were viewed as reaping the benefits of past economic growth and enjoying generous social welfare benefits, adding to the economic burden on the younger generations (Gullette, 2000; Tepe & Vanhuysse, 2009). If the elderly are the winners of the current economic order, they should be supportive of democracy, not less. Thus, the phenomenon of declining support for democracy among both the young and the old in Western democracies is a puzzle that needs to be answered.

This article argues that with population aging (i.e. a large proportion of dependent elderly and a decreasing number of young people), the policies responsible for rising inequality after the 1980s puts further pressures on welfare regime sustainability. Youth and the elderly in particular are vulnerable as a result of these structural changes, with the young suffering from stagnant wages and high unemployment rates and the old facing benefit cuts and reduced savings returns. There is a need for welfare policy and protection for both age cohorts to raise their living standards. However, the current political system, which prioritizes liberal economic policy, is in the opposite direction. Thus, as youth and the elderly experience existential insecurity deeply, they tend to demand economic equality profusely, and the demand dampens support for democracy in its present form, particularly in Western countries (which have largely also aged societies). The argument is supported by survey data collected between 2010 and 2020. The findings show that both young and old citizens believe that equal income distribution is one of the essential characteristics of democracy, and such belief negatively impacts democratic support. The causal chain, however, only appears in aged societies, which explains why this loss of faith in democracy seems to be more salient and alarming in Western countries.

According to previous studies, young people and the elderly support democracy less than adults. On the other hand, existential insecurity dampens democratic support. This study provides empirical evidence for a theory that explains how the age cohorts-demand for economic equality-democratic support nexus is conditioned by population aging, a factor which has been overlooked in previous studies. Thus, despite their polarizing opinions about politics and the fierce debate about intergenerational injustice in the West, the young and old are actually united in their views on what constitutes an ideal democracy.

Support for Democracy and Economic Inequality

Democracy as an ideal regime type is losing its appeal in both new and old democracies (Chu et al., 2020; Foa & Mounk, 2016). This legitimacy crisis is widely attributed to poor governance (Chu et al., 2008; Fernandez & Kuenzi, 2010; Fukuyama, 2015; Huang et al., 2008; Kotzian, 2010; Linde, 2011; Linz & Stepan, 1996; Magalhães, 2014). Among the long list of governance dysfunction, economic inequality has increasingly become a serious concern. Economic inequality is not only about income and wealth gaps between different social classes; it also has serious consequences for political equality and regime stability (Karl, 2000; Muller, 1988; Rueschemeyer, 2004). The world has seen a steady rise in economic inequality over the past decades (Alvaredo et al., 2017; Piketty, 2014), a period coinciding with dwindling support for democracy. Empirical studies have found that economic inequality, measured by the Gini coefficient, drives down support for democracy at the individual level (Andersen, 2012; Krieckhaus et al., 2014). When replaced with a subjective measure of inequality (i.e. perception of inequality), the negative effect of perceived inequality on democratic support becomes even more salient than the effect of the Gini coefficient (Wu & Chang, 2019).

However, support for democracy is not equally distributed across the population. For instance, scholars found that the effects of modernization, such as education, urbanization, and high income, prevent people from losing faith in democracy, even though people with these characteristics tend to be very critical about government performance (Huang et al., 2008). By contrast, Foa and Mounk’s (2016, 2017, 2019) analysis of data from the World Values Survey (WVS) and several country polls found that millennials in advanced democracies expressed the lowest levels of democratic support and were willing to consider non-democratic alternatives. The finding is striking since millennials are more likely to have characteristics associated with modernization, which theoretically should strengthen their democratic orientations. However, it has been found that experiencing democracy was associated with support for democracy; youth's low support for democracy would increase as they learned to practice it (Kotschy & Sunde, 2022). The study by Voeten (2016), on the other hand, found that only youth in the United States had a lower level of democratic support, which indicates that Foa and Mounk’s findings were primarily influenced by U.S. data. Various scholars have also examined longitudinal survey data and found no decline in democratic support among European youth (Wuttke et al., 2020; Zilinsky, 2019). As a result, the question of whether young people in industrialized countries support democracy less remains a matter of debate when examining data from several sources over time. Young people, however, are not the only age cohort that shows declining support for democracy. Also analyzing WVS data, Lechler and Sunde (2019, 861, 866) found that older citizens with “expected proximity to death” tended to have a lower level of support for democracy compared to adults.

Together, existing research suggests that young and old cohorts may hold similar negative views of democracy. On the other hand, economic inequality, whether measured objectively or subjectively, drives down democratic support. If rising inequality is responsible for declining support for democracy, why young and old cohorts are especially affected by it remains unanswered. Consequently, this article introduces another structural factor: population aging and elaborates on how this condition strengthens young and older cohorts’ concerns about economic inequality, which in turn negatively affects their support for democracy.

Population Aging, Rising Inequality and the Demand for Equal Distribution

This article argues that the low levels of democratic support among the young and old are due to their demands for economic equality. However, this age cohorts-demand for economic equality-democratic support only appears in societies with an aging population. Population aging, known as an age structure where there is an increasing proportion of dependent elderly and a decreasing number of descendants, has become the feature of developed countries in the past decades. The phenomenon requires state policies to manage the problems arising from the transition from different age structures, especially dramatic fiscal reforms to maintain welfare regimes sustainability (Cincotta, 2017; Goldstone & Diamond, 2020, 869). The structural and institutional changes that widened economic inequality further put pressure on the state to provide welfare measures to lift living standards for those in need. In developed countries, signs of declining support for democracy are thus a result of the double crises of rising inequality and population aging.

Rising inequality in the past decades is jointly caused by structural and institutional factors. On the structural front, globalization has moved many manufacturing jobs offshore from developed to developing countries, forcing domestic workers into lower-paying jobs (Freeman, 2011, 580; Stiglitz, 2007, 68). Coupled with changes in technology, the demand for less-skilled workers has fallen, while workers who can acquire high-tech skills have been rewarded (Goos et al., 2011). The results of these structural changes are a high unemployment rate and depressed salaries for less-skilled workers, which have exacerbated economic inequality in developed countries. The 2008 financial crisis further pushed up the unemployment rate, with the youth unemployment rate hovering at around 20 percent in Europe (Caliendo & Schmidl, 2016, 2).

Most industrialized economies embraced neoliberalism in the 1980s. A series of institutional changes, including labor market deregulation, privatization of public services, and low taxation for corporations and the rich have also contributed to widening wealth gaps in the past decades (Brown, 2017, 92–93). In order to make welfare systems sustainable amid limits on public expenditure imposed under neoliberalism, many countries had to cut retirement benefits or end benefit generosity and prolong working years (Jackson, 2002; Tepe & Vanhuysse, 2009). Policies implemented after the 2008 financial crisis further worsened inequality. Austerity measures in some countries have hit the poor and underprivileged the most, as they disproportionately relied on the public provision to get by after the crisis (Brown, 2017, 102). At the same time, low-interest rates and quantitative easing designed to stimulate growth have led to new bubbles in the making, reflected in skyrocketing house prices in some countries (Islam & Verick, 2011, 49).

Different mechanisms leading to inequality have varied effects across the population in aged societies. For young people, the main drivers of inequality include high unemployment rates, stagnant wages, and rising house prices (Christophers, 2017; Saunders, 2016). Though older citizens might benefit from booming house prices, most are likely to be harmed by austerity measures and low interest rates. As longevity increases and retirees increasingly depend on government pensions as their major source of income, benefit cuts make poverty more likely to occur among older pensioners and those living alone (Tanton et al., 2009; Williamson & Smeeding, 2004). In countries where pensions remain intact, low interest rates deprive older people of extra incomes from their savings when inflation gradually eats away the stocks (Schwab, 2010). As young and old increasingly need government assistance, budget constraints often lead different generations to compete for welfare provisions and subsidies, causing intergenerational conflicts (Higgs & Gilleard, 2015; Rowlingson et al., 2017). Both youth and older people suffer severe consequences from rising inequality and population aging, and the experiences likely lead them to demand a more equitable distribution. The demand then undermines both age cohorts’ commitment to democracy in its present form, which prioritizes liberal economic policy. In contrast, demographically young societies may not have a sharp difference in the demand for economic equality across age cohorts because the youth do not feel particularly deprived by a large elderly population, even when rising inequality is a common problem across the globe. The proposed nexus between age cohorts, demand for economic equality, and democratic support will be tested by the following hypothesis.

Hypothesis: In aged societies, youth and elders have a higher level of demand for equal income distribution, which then drives down their support for democracy.

Data and Method

This article relies on the World Values Survey Wave 6 and Wave 7 (WVS VI 2010–2014, VII 2017–2020) to test the hypothesis (Haerpher et al., 2020; Inglehart et al., 2014). The dataset includes 67 country cases at different stages of economic and political development. The two waves of the WVS are ideal for testing the hypothesis since most people’s concerns about economic inequality were invigorated by social movements such as Occupy Wall Street in 2010 and the rise of populist leaders in 2016. These phenomena were rooted in the measures taken at the time of the 2008 financial crisis—rescuing banks and stimulating the economy with quantitative easing— which exacerbated economic inequality. In the two survey waves conducted after the Great Recession, resentment should show up if certain groups are hurt by rising inequality. There are other surveys that ask similar questions as the WVS and are of good quality. However, we are interested in how population age structure affects political attitudes of certain age groups, and the coverage of the WVS offers sufficient cases for examining the effect of population aging in comparison to surveys constrained to particular regions with countries likely characterized by similar age structures.

This article argues that the youth and elders are likely to have a higher demand for economic equality due to their living circumstances. Though the financial crisis might create a period effect of weakening support for democracy across all age cohorts, the old and young are more likely to suffer amid rising inequality and population aging as explained in the previous section. For this analysis, we divide ages into seven groups based on intervals of approximately 10 years, then take cohort and cohort squared to see if there is a U-shaped relationship between age cohorts and the demand for economic equality.

This study examines the demand for economic equality as both a dependent and explanatory variable, which is measured by a question that asks respondents whether the state ensuring people’s incomes are equal is an essential characteristic of democracy, with one meaning “not an essential characteristic of democracy” and ten as “an essential characteristic of democracy.” It is similar to the question used by the European Social Survey (ESS, 6th round) to explore ordinary people's socialist conceptions of democracy (Hernández, 2016). The social-democratic conceptions form the normative views ordinary people have about an ideal democracy and they tend to use these beliefs as the benchmark to evaluate the performance of political regimes (i.e. output) based on their experiences with economic and social policies (Torcal & Trechsel, 2016). Analyzing ESS data, Torcal and Trechsel (2016) found that strong beliefs in socialist conceptions led to negative evaluations of social democracy, indicating that such beliefs (expectations) were frustrated by social and economic realities. Accordingly, demand for equal distribution (socialist conception) implies that income distribution may be “perceived” as not being equal in the country, thus the expectation. Consequently, there is likely to be a negative relationship between demand for economic equality and support for democracy in its current form.

Besides the life cycle effect, we also include several theoretically related predictors of demand for economic equality. Ideology is measured by a question item that asks respondents whether large income differences are justifiable. Respondents having this tendency should be less likely to think of economic equality as an essential characteristic of democracy. Factors related to self-interest (food shortage and subjective class) and risk perception (job insecurity) are also included in the model, with the anticipation that they should have impacts on the demand for economic equality (Moene & Wallerstein, 2001; Shepelak & Alwin, 1986).

In terms of support for democracy, it is measured by two questions: “What do you think about having a democratic political system? Would you say it is a very good, fairly good, fairly bad or very bad way of governing this country?” and “How important (from one to ten) is it for you to live in a country that is governed democratically?” Both questions aim to test support for democracy (see for instance Foa and Mounk (2016, 2019) and Claassen (2019)), and the correlation coefficient between them is 0.3234. Looking more closely, however, the first question is more like a regime preference question, while the second a governance question. A person might not prefer a democratic political system but would like her government to act democratically (e.g., responsive, transparent, etc.). The correlation coefficient between the two indicators in China is 0.1141, but it is 0.5018 in the United States. The comparison indicates that the two questions entail slightly different meanings to respondents. Thus, we use both as our dependent variables and name the first support for democracy as a principle and the latter support for democracy in practice. We run analyses separately to see whether we can generate the same findings.

Despite recent challenges, ordinary people generally associate democracy with political freedom and economic prosperity. As a result, support for democracy is usually high since people tend to give “lip service” to democracy when all other alternatives have been discredited (Inglehart, 2003, 52). Based on our data, support for democracy as a principle received an average of 3.31 out of 4 and support for democracy in practice received an average of 8.3 out of 10. The high score is partly due to regime types; in authoritarian regimes, support for democracy is a manifest rejection of the present regime (Claassen, 2019). In addition, the ambiguous understanding of democracy among the masses also blurs the line between liberal democracy and authoritarianism (Kirsch & Welzel, 2018; Lu & Chu, 2021). Accordingly, high scores could also mean support for the current authoritarian regime, which the masses mistakenly believed to be democratic. Political culture and regime experiences were also found to affect conceptions of democracy and the evaluation of political regimes (Ariely, 2015; Ariely & Davidov, 2011; Jacobsen & Fuchs, 2020). As a result, comparative survey studies of the “D” word have to consider cultural and political contexts. We control for the degree of democracy based on the Freedom House index by taking into account the regime type as a source of political socialization. Based on the combined scores of political rights and civil liberties, Freedom House assigns each country the status of Free, Party Free, or Not Free (Freedom House, 2022). As a result, democracy is classified into three levels: non-democracies (not free), semi-democracies (partly free), and democracies (free). We expect to see stronger support for democracy in less free countries based on previous studies discussed above. For control variables, we include theoretically related predictors—demand for gender equality and life satisfaction (as post-materialist orientations) (Graham & Pettinato, 2001; Inglehart, Norris, and Welzel 2003; M.-H. Huang et al., 2008), confidence in the government and evaluation of democracy at work (output performance) (Bratton & Mattes, 2001; Chang et al., 2013), political interest (psychological involvement) (Mattes & Bratton, 2007) as well as standard controls — education, gender, and income level in the model.

At the national level, one structural factor is of interest: The old-age dependency ratio is measured as the percentage of the population older than 65, compared to the percentage of the working-age population (15–64). Societies then are divided into two groups based on World Health Organization categorization: aged societies with the old-age dependency ratio exceeding 14 percent, and young societies with ratios of less than 14 percent. We run analyses by different specifications and expect to see that only in aged societies does cohort form a U-shaped relationship with demand for economic equality. Gini coefficients are compared with the demand to see whether they produce similar results on support for democracy. Since the argument presented in this article depends on population aging as a condition–if this structural factor is, in fact, measuring the Gini coefficient, it would undermine our argument. Previous research has used the relative size of mature-aged cohorts as an instrument of the Gini coefficient, arguing that a “fat” mature-aged cohort (i.e. working-age population) indicates better economic equality due to more labor market competition (Krieger and Meierrieks 2016, 16). If this is the case, we would expect to see the Gini coefficient and old-age dependency ratio to be positively correlated (i.e. a higher old-age dependency ratio indicates a smaller proportion of working-age population, and thus, a higher unequal distribution). The Pearson coefficient of the two is actually negative (− 0.2996), defying the expectation. In addition, in less free countries, political inequality might lead to economic inequality and vice versa, so the level of democracy and the Gini coefficient might be negatively correlated; free countries are associated with equal distribution. The coefficient of the two is merely 0.0042, so collinearity is not a concern. All models also control for GDP growth rate and GNI per capita. We assume that good economic conditions should increase the level of support for democracy at the individual level (Cho, 2014). Question wordings and summary statistics are reported in the “Appendix”.

Since we have variables at both the individual and national levels, the standard application should be a multilevel linear model. However, as we are interested in combining information from three equations, a seemingly unrelated regression (SUR) serves the purpose. A SUR consists of several relationships that have correlated disturbances. The correlation between the equation disturbances of support for democracy as a principle and in practice is likely to be observed in our example because they share similar explanatory variables. It is plausible that demand for economic equality decreases support for democracy, but it is also likely that decreasing democratic support leads to inaccurate estimates of income distribution. Accordingly, demand for economic equality and support for democracy are mutually causal in theory, so even if estimated with different equations, it cannot be assumed that the error terms of the two equations are uncorrelated, as OLS assumes (Zellner, 1962). As a result, we use SUR to estimate three equations simultaneously: demand for economic equality, support for democracy as a principle, and support for democracy in practice.

Results and Discussion

Column 1 in Table 1 presents the results that in aged societies, youth and elders have a higher level of demand for economic equality than adults compared to other societies. Without taking population aging into account, the relationship between cohorts and demand looks inverted U-shaped, meaning that demand for economic equality increases first, followed by a decrease along the age cohort spectrum. The interaction terms between cohort and cohort squared and old-age dependency ratio are negative and positive respectively, turning the relationship between cohorts and demand for economic equality to be U-shaped. Such a relationship is more comprehensible when we run analyses separately in aged and young societies. In aged societies, the young and the old have a higher demand for economic equality than adults, but in young societies, the adult cohorts have the highest level of demand instead (see Tables 5 and 6 in the “Appendix”). Aside from this difference, the signs of other predictors (ideology, food shortage, job insecurity, and subjective class) are consistent across societies, indicating that population aging only has the effect on age cohorts’ demand for economic equality. Since the majority of aged societies are democracies (21 out of 30 country cases), the findings explain why the young and old voters in those societies are likely to be attracted by extreme political appeals, such as socialism for the young (for instance, Bernie Sanders’s popularity among young Democrats), and right-wing populism for the elders who increasingly vote for preserving national culture and anti-immigration platforms in the context of declining economic security (Inglehart & Norris, 2017).

Table 1.

The determinants of demand for economic equality and support for democracy

Variable	Demand Eco. Equality	Support D. as principle	Support D. in practice
	(1)	(2)	(3)
	Coef. (p)	Coef. (p)	Coef. (p)
Cohort	0.287 (0.00)	− 0.004 (0.49)	0.103 (0.00)
Cohort2	− 0.027 (0.00)	0.001 (0.08)	− 0.007 (0.001)
Cohort*old age dep	− 0.020 (0.00)
Cohort2*old age dep	0.002 (0.00)
Demand eco. equ		0.003 (0.01)	0.009 (0.01)
Demand*old age dep		− 0.0007 (0.00)	− 0.001 (0.00)
Ideology	− 0.146 (0.00)
Food shortage	− 0.124 (0.00)
Job insecurity	0.062 (0.00)
Subjective class	− 0.149 (0.00)
Gender equality		0.054 (0.00)	0.216 (0.00)
Political interest		0.061 (0.00)	0.110 (0.00)
Evaluation of democracy		0.016 (0.00)	0.122 (0.00)
Life satisfaction			0.083 (0.00)
Confidence in the government			− 0.086 (0.00)
Income level		− 0.002 (0.07)	− 0.025 (0.00)
Male	− 0.098 (0.00)	0.033 (0.00)	0.059 (0.00)
Education	− 0.275 (0.00)	0.013 (0.00)	0.174 (0.00)
Gini coefficient	− 0.022 (0.00)	− 0.002 (0.00)	− 0.010 (0.00)
GNI per capita	− 0.029 (0.00)	− 0.001 (0.00)	− 0.003 (0.00)
GDP growth rate	− 0.006 (0.00)	− 0.004 (0.00)	− 0.010 (0.00)
Level of democracy	− 0.416 (0.00)	− 0.012 (0.00)	0.033 (0.00)
Old age dependency ratio	0.043 (0.00)	0.005 (0.00)	0.006 (0.00)
Wave 7	0.264 (0.00)	− 0.030 (0.00)	− 0.0007 (0.95)
Constant	9.398 (0.00)	2.725 (0.00)	5.118 (0.00)
No. of observations	118,540
No. of countries	67
R2	0.0844	0.0441	0.1305
Adjusted R2	0.0842	0.0440	0.1303

Model: Seemingly unrelated regression

At the national level, high GNI per capita lowers the level of demand for economic equality at the individual level, indicating that people tend to emphasize the importance of inequality less when the average income is high. GDP growth rate also has a negative effect on the demand. The surprising finding is that the Gini coefficient decreases the level of the demand for economic equality, defying our common sense. This finding echoes the puzzle that constantly emerges in research on economic inequality: the gap between objective measures of economic inequality and subjective assessments of income distribution. The United States, for instance, has the most unequal wealth distribution among the industrialized countries, and yet Americans are found to underestimate the real level of inequality (Chambers et al., 2014; Norton & Ariely, 2011). The possible reason may be that people living in countries with high inequality tend to justify the circumstances in which they find themselves (Trump, 2017). Despite the fact that rising inequality is a reality, people may not demand an equal share when they believe inequality is a reflection of meritocracy, and thus, underestimate its seriousness (Mijs, 2019).

Democracies, on average, have a lower level of demand for economic equality at the national level compared to semi- and non-democracies. This indicates that political inequality may go hand-in-hand with economic inequality, leading to equal share demand. Lastly, demand for economic equality further increased in Wave 7 (2017–2020) of the World Values Survey, ten years after the financial crisis, indicating that policy measures at the time failed to restore existential security but instead exacerbated inequality that further fuels demand for economic equality.

Model 2 in Table 1 shows whether demanding economic equality negatively impacts support for democracy as a principle across societies. The sign of the demand is positive, but it becomes negative when the old-age dependency ratio increases. As seen in Tables 5 and 6 in the “Appendix”, the coefficient of demanding economic equality is negative in aged societies, but positive (though not statistically significant) in young societies. There is evidence that population aging exacerbates the negative effects of the demand for economic equality. When the proportion of old-age dependents is low, the demand may even have a positive effect on people's support for democracy. There is also a consistent relationship between demanding economic equality and the Gini coefficient in aged societies, with both indicators relating to lower support for democracy as a principle at the national and individual levels. Given the surveys were taken after the 2008 financial crisis, the finding echoes previous research that finds income shocks and inequality respectively lowers democratic quality (Kotschy & Sunde, 2021).

Based on the results of Model 3 in Table 1, democracy in practice appears to be supported in a similar way to Model 2; the effect of demanding economic equality becomes negative as the old-age dependency ratio increases. However, in young societies, the effect is positive and statistically significant. It is likely that the divergent signs of demanding economic equality in aged and young societies are due to the endogeneity of economic inequality in support for democracy. Even when democracy level and Gini coefficient are controlled for in the model, people living in less free and likely unequal countries tend to support democracy (no matter what the conception of democracy is). According to Model 3 of Table 6 in the “Appendix”, democracies on average have a lower level of support for democracy than other types of regimes, suggesting that demanding economic equality and political constraints are the main factors determining support for democracy in young societies. Nevertheless, in aged societies the effect of demand for economic equality is negatively correlated with support for democracy as a principle and in practice.

Cohorts alone still have an impact on support for democracy, though. Young and elderly people tend to support democracy in practice less in aged societies, a similar finding as Lechler and Sunde’s (2019). Both cohorts’ support for democracy further declines due to their high levels of demand for economic equality. Other predictors follow our expectations, except confidence in government; it actually lowers democratic support. Further investigation is needed, but the likely explanation is that people with traditional or authoritarian values tend to trust the government more, associating democracy with conflicts and inefficiency (Ma & Yang, 2014; Shi, 2001). Confidence in government, in this regard, is likely to negatively affect support for democracy.

In this article, it is argued that the negative effects of demanding economic equality on democratic support are most pronounced among the young and old in democracies with aged populations. However, previous analyses have not directly addressed the issue, despite the fact that the majority of aged societies are democracies. Table 2 outlines the analysis in the sub-sample of aged democracies, and the results are similar to those in Table 1 and Table 5 in the “Appendix”. Unlike the results reported in Table 7 in the “Appendix” for young democracies, only in aged democracies do the youth and elders have similar attitudes toward economic equality, and such attitudes negatively affect democratic support (Models 2 and 3). When the United States is excluded from the sample, the results remain unchanged (see Table 8 in the “Appendix”). Findings indicate that declining support for democracy among the youth and elderly is not unique to the United States; other advanced democracies also show a similar trend. Alternatively, cohorts do not affect the demand for economic equality in democracies that are demographically young (Table 7), although the demand still contributes to lower support for democracy. Thus, our hypothesis is supported by survey data.

Table 2.

The determinants of demand for economic equality and support for democracy in aged democracies

Variable	Demand Eco. Equality	Support D. as principle	Support D. in practice
	(1)	(2)	(3)
	Coef. (p)	Coef. (p)	Coef. (p)
Demand eco. equ		− 0.009 (0.00)	− 0.024 (0.00)
Cohort	− 0.228 (0.00)	0.046 (0.00)	0.265 (0.00)
Cohort2	0.021 (0.00)	− 0.002 (0.08)	− 0.021 (0.00)
Ideology	− 0.222 (0.00)
Food shortage	0.161 (0.00)
Job insecurity	0.174 (0.00)
Subjective class	− 0.166 (0.00)
Gender equality		0.079 (0.00)	0.252 (0.00)
Political interest		0.082 (0.00)	0.209 (0.00)
Evaluation of democracy		0.041 (0.00)	0.165 (0.00)
Life satisfaction			0.056 (0.00)
Confidence in the government			− 0.073 (0.00)
Income level		0.010 (0.00)	− 0.001 (0.75)
Male	− 0.170 (0.00)	0.026 (0.00)	− 0.030 (0.09)
Education	− 0.526 (0.00)	0.006 (0.34)	0.158 (0.00)
Gini coefficient	− 0.069 (0.00)	0.001 (0.01)	− 0.004 (0.01)
GNI per capita	− 0.055 (0.00)	− 0.0007 (0.01)	− 0.006 (0.00)
GDP growth rate	− 0.085 (0.00)	− 0.024 (0.00)	− 0.019 (0.00)
Wave 7	0.455 (0.00)	0.047 (0.00)	− 0.131 (0.00)
Constant	12.293 (0.00)	2.057 (0.00)	4.412 (0.00)
No. of observations	30,722
No. of countries	21
R2	0.1643	0.1194	0.2064
Adjusted R2	0.1640	0.1190	0.2061

Model: Seemingly unrelated regression

Robustness Checks

For robustness checks, we test whether our findings will differ if we chose to code our key variable–cohorts differently. Using multiverse analysis, we aim to examine the stability of our findings, ensuring that they are not the result of specific data processing (Steegen et al., 2016). The following processing options are used: (1) age and age squared; (2) three cohort categories (youth, adults, elderly); (3) generation and generation squared; and (4) five generation categories (silent, baby boomers, Generation X, Millennials, and Generation Z). Furthermore, we examine whether the coefficient of demand for economic equality will be different if we exclude the squared factors of cohort, age, and generation. The coefficients and p-values are presented in Table 3.

Table 3.

Results of multiverse analysis

	Demand Eco. Equality	Support D. as principle	Support D. in practice
	alternative * old age dep	Demand Eco. Equality * old age dep
	(1)	(2)	(3)
Cohort	− 0.020 (0.00)	− 0.0007 (0.00)	− 0.001 (0.00)
Cohort2	0.002 (0.00)	–	–
Age	− 0.002 (0.00)	− 0.0007 (0.00)	− 0.001 (0.00)
Age2	0.00002 (0.00)	–	–
Generation	− 0.041 (0.00)	− 0.0007 (0.00)	− 0.001 (0.00)
Generation2	0.004 (0.00)	–	–
Youth	0.021 (0.00)	− 0.0007 (0.00)	− 0.001 (0.00)
Elder	− 0.009 (0.003)
Millennials	0.014 (0.77)	− 0.0007 (0.00)	− 0.001 (0.00)
Generation X	− 0.002 (0.95)
Baby Boomer	− 0.010 (0.83)
Silent	− 0.007 (0.88)

p values are in the parentheses

Table 3 shows two results. The U-shaped relationship between cohort-alternative variables and the demand for economic equality in aged societies is only evident when this variable is coded as a continuous variable. The first column in Table 3 shows that when we use age, cohort and generation, as well as their squared factors, the signs of the interaction terms between them and old-age dependency ratio are all the same. When we use categorical variables for generations, the coefficients of the interaction terms are not statistically significant. When respondents are divided into three cohorts, the interaction term between youth and old-age dependency ratio is positive, while it is negative for elders. The results seem not to match previous findings. In the equation, the coefficient of youth is − 0.293, but if the old-age dependency ratio is 21, this coefficient will increase to 0.148. At the same time, the coefficient of demanding economic equality will be 0.016 for the elderly. As a result, youth and elders still have a higher demand for economic equality than adults in aged societies. Multiverse analysis confirms three of the four alternatives to cohort and cohort squared, indicating that our findings remain robust regardless of coding choice. Accordingly, the higher demands for economic equality among the young and old are due to the life cycle effect, not generation effect. Regardless of the specifications, the coefficients between demand for economic equality and old-age dependency ratio are all the same when squared factors are excluded from the equations for support for democracy. This further supports the robustness of our findings.

An alternative measure of socialist democracy is also examined to understand whether findings above are due to our choice of measurement. The variable is measured by asking: Are governments taxing the rich and subsidizing the poor essential to democracy? The results are presented in Table 9 in the “Appendix”. A positive correlation exists between economic equality and redistribution demand, as shown in Models 1 and 2. The age effect, however, does not have a U-shaped relationship with the demand for redistribution. The demand for redistribution increases as one ages, meaning that young people are less likely to believe that redistribution is essential to democracy, while the elderly are more likely to believe that it is. In spite of similar attitudes toward economic equality, youth and elders hold differing views about redistribution policy. There is also a divergence between the two variables' effects on democratic support, where a higher demand for redistribution leads to increased support for democracy, not the opposite as with the demand for economic equality (Models 3 and 4). Although the results seem contradictory, given that most advanced democracies have some form of welfare policy, its positive effect on democratic support indicates that the output has met the expectations of redistribution. There are, however, other ways to reduce economic inequality. Specifically, youth are more concerned about policies that will help them land better jobs or subsidies that will reduce education costs, housing problems, and child-rearing expenses. Although both young and old are concerned about economic inequality, conflicts over policy priorities are likely to turn them against one another. The results support the argument that population aging and rising inequality put pressure on government finances, which raise concerns about worsening economic inequality and in turn, bring down youth and the elderly's faith in democracy in its current form, even when redistribution exists.

Conclusion

In this article we analyze survey data from 67 country cases and find that both youth and elders in aged societies (democracies) hold similar views on economic inequality, which undermine support for democracy both as a principle and in practice. The findings explain why growing inequality is particularly harmful for advanced democracies since these countries are facing an additional problem of population aging. However, If both young and older people think inequality is a pressing problem in advanced democracies, the conflict between the two generations needs some explanation. Previous studies have viewed the confrontation as a clash of values between materialists and post-materialists who grew up in different economic environments (Inglehart, 2016a). At the same time, parties have adopted a survival strategy of shifting competition from economic to social issues when both left and right parties embrace neoliberalism (Berman & Snegovaya, 2019; Inglehart & Norris, 2017). The findings of this article imply that youth and the elders actually have something in common but this commonality is overshadowed by their diverging priorities over welfare policies and polarizing political opinions, which were fueled by the rise of populism that threatens liberal democracy. Shifting competition back to economic issues and identifying compromise solutions, although certainly not an easy task, might be the first step to find common ground and regain the sense of community that has been undermined by identity-based political polarization.

As the global population projected to get older in the next three decades, in which the share of the population aged 65 years or over is expected to reach 16 percent in 2050 (United Nations, 2020, 1), such structural change is likely to exacerbate the inequality problem by activating the youth’s political discontent. While the older generation who grew up in the post-war order typically prefer the redistribution approach, the young seek stable jobs and salaries (reward) compatible with their (educational) investment. Thus, besides the wealth tax advocated by Piketty (2014), institutional arrangements that strengthen workers’ bargaining power against employers should be restored and re-calibrated to suit the new environment. The reason for the failure of left-wing parties is because not only they abandoned their core value of protecting the working and lower classes but also due to their obsolete methods of organizing labor (it is impossible to organize people working in the service sector in traditional unions). The left needs to innovate to catch up with the ever-evolving economic environment shaped by new technology. Considering that the world is heading towards stagflation due to the effects of the Covid-19 pandemic and the Russia-Ukraine war, it is likely that global inequality will rise again, this time affecting even adults. In order to adjust the course of neoliberalism and develop creative solutions to meet the technological revolution, bold reforms are needed in the West. By doing so, liberal democracy can regain the trust of its citizens and continue to inspire millions of people living in less free countries.

Appendix

See Tables 4, 5, 6, 7, 8, and 9.

Table 4.

Question wording and summary statistics

Variable	Question Wording	Scale and Summary Statistics
Individual Level
Demand for economic equality	Essential characteristic of democracy: The state makes people’s incomes equal	10-point scale, 10 = an essential characteristic; mean = 6.007, s.d. = 3.034
Support for democracy as principle	Types of political system governing the country: having a democratic political system	4-point scale: 4 = very good; mean = 3.313; s.d. = 0.791
Support for democracy in practice	How important is it for you to live in a country that is governed democratically?	10-point scale, 10 = absolutely important; mean = 8.334, s.d. = 2.091
Demand for gender equality	Essential characteristic of democracy: women have the same rights as men	10-point scale, 10 = an essential characteristic; mean = 7.940, s.d. = 2.576
Demand for redistribution	Essential characteristic of democracy: Governments tax the rich and subsidize the poor	10-point scale, 10 = an essential characteristic; mean = 6.329, s.d. = 3.002
Life satisfaction	How satisfied are you with your life as a whole these days?	10-point scale, 10 = completely satisfied; mean = 6.979, s.d. = 2.260
Confidence in government	How much confidence you have in the government?	4-point scale: 4 = a great deal; mean = 2.428; s.d. = 0.969
Political interest	How interested would you say you are in politics?	4-point scale: 4 = very interested; mean = 2.369; s.d. = 0.959
Evaluation of democracy	How democratically is this country being governed today?	10-point scale, 10 = completely democratic; mean = 6.139, s.d. = 2.573
Income level	We would like to know in what (income) group your household is	10-point scale, 10 = highest group; mean = 4.829, s.d. = 2.082
Male	Respondent’s sex	2-point scale: 1 = male; mean = 0.494; s.d. = 0.500
Education	What is the highest educational level that you have attained?	3-point scale, 3 = tertiary education, with degree; mean = 2.182; s.d. = 0.652
Age cohort	You are___ years old; 1 = 16–29; 2 = 30–39; 3 = 40–49; 4 = 50–59; 5 = 60–69; 6 = 70–79; 7 = above 80	Mean = 2.76; s.d. = 1.55
Wave	Survey wave	Wave 6 = 54;
		Wave 7 = 44
Gini coefficient	Figures in the survey year, from World Bank, http://api.worldbank.org/v2/en/indicator/SI.POV.GINI?downloadformat=excel, accessed on 2020/9/17	mean = 39.520; s.d. = 7.643
GNI per capita	World Development Indicators, from http://api.worldbank.org/v2/en/indicator/NY.GNP.PCAP.KD?downloadformat=excel, accessed on 2020/9/17	In 1,000 unit, mean = 14.312; s.d. = 16.964
GDP growth rate	World Development Indicators, from https://databank.worldbank.org/source/world-development-indicators, accessed on 2020/1/2	mean = 3.344; s.d. = 4.761
Old-age dependency ratio	Age dependency ratio, old (% of working-age population), from http://api.worldbank.org/v2/en/indicator/SP.POP.DPND.OL?downloadformat=excel, accessed on 2020/9/17	mean = 14.202; s.d. = 8.422
Level of democracy	Freedom House Index, from https://freedomhouse.org/countries/freedom-world/scores, accessed on 2020/9/17	3 Democracy = 25; 2 semi-democracy = 26; 1 non-democracy = 21
Aged societies	Australia, Estonia, Georgia, Germany, Taiwan, Greece, Hong Kong, Japan, Netherlands, New Zealand, Puerto Rico, Romania, Serbia, Slovenia, Spain, Sweden, Ukraine, United States, Uruguay, Argentina, Armenia, Belarus, China (2018), Chile (2018), Cyprus, South Korea, Macau, Poland, Russia, Thailand (2018)
Young societies	Algeria, Azerbaijan, Bangladesh, Bolivia, Brazil, Myanmar, Chile (2011), China (2012), Colombia, Ecuador, India, Indonesia, Kazakhstan, Kyrgyzstan, Lebanon, Malaysia, Mexico, Morocco, Nicaragua, Pakistan, Peru, Philippines, Singapore, Vietnam, South Africa, Thailand (2013), Tunisia, Turkey, Egypt, Iran, Ethiopia, Ghana, Iraq, Jordan, Libya, Nigeria, Rwanda, Zimbabwe, Tajikistan, Yemen

Table 5.

The determinants of demand for economic equality and support for democracy in aged societies

Variable	Demand Eco. Equality	Support D. as principle	Support D. in practice
	(1)	(2)	(3)
	Coef. (p)	Coef. (p)	Coef. (p)
Demand eco. equ		− 0.012 (0.00)	− 0.018 (0.00)
Cohort	− 0.151 (0.00)	0.033 (0.00)	0.211 (0.00)
Cohort2	0.014 (0.001)	− 0.001 (0.11)	− 0.020 (0.00)
Ideology	− 0.200 (0.00)
Food shortage	− 0.072 (0.00)
Job insecurity	0.116 (0.00)
Subjective class	− 0.225 (0.00)
Gender equality		0.077 (0.00)	0.275 (0.00)
Political interest		0.085 (0.00)	0.207 (0.00)
Evaluation of democracy		0.029 (0.00)	0.152 (0.00)
Life satisfaction			0.061 (0.00)
Confidence in the government			− 0.072 (0.00)
Income level		0.003 (0.08)	− 0.014 (0.001)
Male	− 0.169 (0.00)	0.018 (0.007)	− 0.005 (0.75)
Education	− 0.545 (0.00)	0.003 (0.584)	0.128 (0.00)
Gini coefficient	− 0.010 (0.00)	− 0.001 (0.00)	0.002 (0.09)
GNI per capita	− 0.037 (0.00)	− 0.002 (0.00)	− 0.006 (0.00)
GDP growth rate	− 0.068 (0.00)	− 0.007 (0.00)	0.012 (0.00)
Level of democracy	− 0.414 (0.00)	0.050 (0.00)	0.285 (0.00)
Wave 7	0.378 (0.00)	0.061 (0.00)	− 0.078 (0.00)
Constant	11.265 (0.00)	2.254 (0.00)	3.344 (0.00)
No. of observations	45,703
No. of countries	30
R2	0.1540	0.0953	0.2058
Adjusted R2	0.1537	0.0950	0.2055

Model: Seemingly unrelated regression

Table 6.

The determinants of demand for economic equality and support for democracy in young societies

Variable	Demand Eco. Equality	Support D. as principle	Support D. in practice
	(1)	(2)	(3)
	Coef. (p)	Coef. (p)	Coef. (p)
Demand eco. equ		0.001 (0.21)	0.011 (0.00)
Cohort	0.120 (0.00)	− 0.005 (0.49)	0.051 (0.01)
Cohort2	− 0.014 (0.002)	0.001 (0.91)	− 0.002 (0.45)
Ideology	− 0.118 (0.00)
Food shortage	− 0.110 (0.00)
Job insecurity	0.031 (0.002)
Subjective class	− 0.106 (0.00)
Gender equality		0.042 (0.00)	0.186 (0.00)
Political interest		0.042 (0.00)	0.055 (0.00)
Evaluation of democracy		0.006 (0.00)	0.101 (0.00)
Life satisfaction			0.086 (0.00)
Confidence in the government			− 0.087 (0.00)
Income level		− 0.007 (0.00)	− 0.031 (0.00)
Male	− 0.031 (0.14)	0.039 (0.00)	0.089 (0.00)
Education	− 0.152 (0.00)	0.026 (0.00)	0.180 (0.00)
Gini coefficient	− 0.042 (0.00)	0.0006 (0.21)	− 0.001 (0.33)
GNI per capita	− 0.009 (0.00)	− 0.004 (0.00)	− 0.006 (0.00)
GDP growth rate	− 0.003 (0.05)	− 0.003 (0.00)	− 0.009 (0.00)
Level of democracy	− 0.274 (0.00)	− 0.069 (0.00)	− 0.182 (0.00)
Wave 7	0.201 (0.00)	− 0.090 (0.00)	0.032 (0.04)
Constant	9.738 (0.00)	2.983 (0.00)	5.666 (0.00)
No. of observations	72,837
No. of countries	40
R2	0.0390	0.0328	0.1021
Adjusted R2	0.0388	0.0326	0.1019

Model: Seemingly unrelated regression

Table 7.

The determinants of demand for economic equality and support for democracy in young democracies

Variable	Demand Eco. Equality	Support D. as principle	Support D. in practice
	(1)	(2)	(3)
	Coef. (p)	Coef. (p)	Coef. (p)
Demand eco. equ		− 0.010 (0.00)	− 0.020 (0.01)
Cohort	0.110 (0.22)	− 0.020 (0.41)	0.050 (0.42)
Cohort2	− 0.013 (0.35)	0.004 (0.30)	− 0.007 (0.49)
Ideology	− 0.115 (0.00)
Food shortage	0.203 (0.00)
Job insecurity	− 0.001 (0.96)
Subjective class	− 0.040 (0.22)
Gender equality		0.052 (0.00)	0.220 (0.00)
Political interest		0.050 (0.00)	0.054 (0.01)
Evaluation of democracy		0.010 (0.006)	0.163 (0.00)
Life satisfaction			0.109 (0.00)
Confidence in the government			− 0.048 (0.05)
Income level		− 0.039 (0.00)	− 0.055 (0.00)
Male	0.009 (0.87)	0.015 (0.38)	0.116 (0.01)
Education	0.147 (0.01)	0.112 (0.00)	0.238 (0.00)
Gini coefficient	− 0.046 (0.00)	− 0.002 (0.23)	− 0.017 (0.00)
GNI per capita	0.032 (0.01)	0.009 (0.009)	0.049 (0.00)
GDP growth rate	− 0.180 (0.00)	0.092 (0.00)	0.092 (0.00)
Wave 7	− 1.021 (0.00)	0.134 (0.00)	0.657 (0.00)
Constant	8.798 (0.00)	2.385 (0.00)	4.404 (0.00)
No. of observations	8454
No. of countries	5
R2	0.0502	0.1336	0.1481
Adjusted R2	0.0488	0.1322	0.1466

Model: Seemingly unrelated regression

Table 8.

The determinants of demand for economic equality and support for democracy in aged democracies, excluding the U.S

Variable	Demand Eco. Equality	Support D. as principle	Support D. in practice
	(1)	(2)	(3)
	Coef. (p)	Coef. (p)	Coef. (p)
Demand eco. equ		− 0.007 (0.00)	− 0.015 (0.00)
Cohort	− 0.201 (0.00)	0.036 (0.00)	0.242 (0.00)
Cohort2	0.021 (0.00)	− 0.001 (0.27)	− 0.022 (0.00)
Ideology	− 0.222 (0.00)
Food shortage	0.074 (0.008)
Job insecurity	0.169 (0.00)
Subjective class	− 0.154 (0.00)
Gender equality		0.077 (0.00)	0.239 (0.00)
Political interest		0.071 (0.00)	0.177 (0.00)
Evaluation of democracy		0.034 (0.00)	0.153 (0.00)
Life satisfaction			0.056 (0.00)
Confidence in the government			− 0.054 (0.00)
Income level		0.009 (0.00)	− 0.007 (0.12)
Male	− 0.133 (0.00)	0.021 (0.01)	− 0.056 (0.003)
Education	− 0.496 (0.00)	0.002 (0.70)	0.150 (0.00)
Gini coefficient	− 0.074 (0.00)	0.014 (0.00)	0.009 (0.00)
GNI per capita	− 0.058 (0.00)	0.004 (0.00)	0.0003 (0.69)
GDP growth rate	− 0.091 (0.00)	− 0.016 (0.00)	− 0.010 (0.02)
Wave 7	0.428 (0.00)	0.020 (0.02)	− 0.152 (0.00)
Constant	12.491 (0.00)	1.621 (0.00)	4.056 (0.00)
No. of observations	26,863
No. of countries	20
R2	0.1379	0.1208	0.1871
Adjusted R2	0.1375	0.1203	0.1866

Model: Seemingly unrelated regression

Table 9.

The determinants of demand for economic equality, demand for redistribution, and support for democracy in aged democracies

Variable	Demand for economic equality	Demand for redistribution	Support D. as principle	Support D. in practice
	(1)	(2)	(3)	(4)
	Coef. (p)	Coef. (p)	Coef. (p)	Coef. (p)
Demand for redistribution	0.548 (0.00)		0.005 (0.00)	0.033 (0.00)
Demand for economic equality		0.530 (0.00)	− 0.012 (0.00)	− 0.034 (0.00)
Cohort	− 0.232 (0.00)	0.127 (0.00)	0.048 (0.00)	0.261 (0.00)
Cohort2	0.019 (0.00)	− 0.006 (0.25)	− 0.002 (0.07)	− 0.021 (0.00)
Ideology	− 0.165 (0.00)	0.013 (0.02)
Food shortage	0.151 (0.00)	− 0.071 (0.003)
Job insecurity	0.102 (0.00)	0.045 (0.002)
Subjective class	− 0.089 (0.00)	− 0.052 (0.005)
Gender equality			0.078 (0.00)	0.249 (0.00)
Political interest			0.082 (0.00)	0.207 (0.00)
Evaluation of democracy			0.040 (0.00)	0.162 (0.00)
Life satisfaction				0.056 (0.00)
Confidence in gov				− 0.074 (0.00)
Income level			0.011 (0.00)	0.001 (0.836)
Male	− 0.154 (0.00)	0.059 (0.04)	0.026 (0.002)	− 0.031 (0.08)
Education	− 0.492 (0.00)	0.224 (0.00)	0.006 (0.33)	0.158 (0.00)
Gini coefficient	− 0.030 (0.00)	− 0.036 (0.00)	0.002 (0.01)	− 0.003 (0.06)
GNI per capita	− 0.050 (0.00)	0.019 (0.00)	− 0.001 (0.01)	− 0.006 (0.00)
GDP growth rate	− 0.089 (0.00)	0.050 (0.00)	− 0.025 (0.00)	− 0.020 (0.00)
Wave 7	0.219 (0.00)	0.188 (0.00)	0.045 (0.00)	− 0.143 (0.00)
Constant	7.051 (0.00)	3.088 (0.00)	2.039 (0.00)	4.263 (0.00)
No. of obs	30,439
No. of countries	21
R2	0.1856	0.0624	0.1202	0.2100
Adjusted R2	0.1852	0.0620	0.1198	0.2096

Demand for redistribution (question wording): Governments tax the rich and subsidize the poor

Model: Seemingly Unrelated Regression

Declarations

Competing interests

The author(s) declare none.