1 Introduction
Economic inequality has been a global challenge for decades. Since the 1990s, global inequality (inequality across all individuals in the world) has been declining.1 However, income inequality has increased within countries, including most developed countries and some middle-income countries, such as China and India.2 From the 1980s to the 2010s, the share of the top 1 percentage earners in total income increased in most OECD countries, which indicates that the top 1 percentage earners are able to capture a disproportionate share of overall income growth.3 Economic inequality is hence a pressing issue for most countries.
Economic inequality could have significant political-social impacts. First, the literature suggests that a large income gap could lead voters to demand higher income redistribution (Meltzer & Richard, 1981). Cross-national surveys suggest that perceived economic inequality leads to stronger support for government redistribution (Gimpelson & Treisman, 2018). Second, economic inequality can be a source of relative deprivation, as economic inequality has the potential to evoke feelings of disappointment and anguish among low-income citizens as they compare their lives with those of the affluent (Goldstone, 1982, 192). The relative deprivation and social discontent caused by economic inequality can thus push citizens to join protests and even social revolutions, as in the Egyptian revolution after the Second World War (see Chen et al., 2023; Gurr, 1968; Alesina & Perotti, 1996).4 In countries where economic inequality is salient across ethnic lines, the economic inequality can further be a source of ethnic conflict, as in the case of Rwanda in the 1990s (Cramer, 2003). Moreover, Acemoglu and Robinson (2006) argue that income inequality could lead to democratization as democracy allocates more resources to ordinary citizens than autocracy; thus, ordinary citizens will prefer democracy and call for democratization.
Advertisement
Hong Kong, as one of the most unequal developed economies, offers an intriguing case of examining the political effects of economic inequality. Hong Kong’s GDP per capita reached US$ 46,324 in 2020, which ranked 25th in the world. However, Hong Kong maintained high income inequality during its rapid economic development between 1960 and 1990 (Wang & Xia, 2012). Since the handover in 1997, economic inequality in Hong Kong has been rising (Wong, 2017). Hong Kong’s Gini coefficient before the handover was 0.518 in 1996.5 After the handover, the Gini coefficient increased slightly to 0.533 in 2006 and 0.539 in 2016.6 After considering tax and in-kind social benefits, the Gini coefficient in Hong Kong was still as high as 0.473 in 2016,7 which was higher than most developed economies in 2016, such as Singapore (0.356), the United States (0.391), the UK (0.351), Australia (0.337) and Canada (0.318).8 The gap between the rich and the poor is astonishing. A government report states that 1,098,000 Hong Kong citizens (15.8% of the total population) lived in poverty in 2019.9 In contrast, it is estimated that 515,000 Hong Kong citizens had a net asset of more than 10 million Hong Kong dollars in 2020, which means that one of every twelve Hong Kong citizens is rich, with net assets of 10 million Hong Kong dollars.10 Various factors, including the residual welfare regime, inequality in the education sector, and inflow of high-skilled global talent and capital, all contribute to the high level of economic inequality in Hong Kong (Chan et al., 2022; Mok, 2015; Nagy, 2015; Yang, Miao, and Wu, 2022).
Hong Kong has experienced dramatic socio-political changes in recent years, including massive social protests and the rise of localism. Most notably, the Umbrella Movement in 2014 and the Anti-Extradition Law Amendment Bill (Anti-ELAB) Movement in 2019 attracted mass participation (Cheng et al., 2022; Yang, 2019). The Umbrella Movement was a democratic movement that targeted universal suffrage, but the student leaders also made an argument that Hong Kong’s political-economic systems were biased toward big business and that this had led to economic inequality; therefore, Hong Kong needed to achieve democratization to address inequality. The class boycott campaign by the Hong Kong Federation of Students claimed that without democratic reform, Hong Kong would continue to be an ‘absurd city” with the highest level of inequality in the world.11 In addition to social movements, Hong Kong has been an epicenter of identity conflicts in China since the mid-2000s (Lee, 2020, 371; Steinhardt, Li, and Jiang 2018). After the Umbrella Movement, localism rose with a strong claim of self-determination and even independence (Veg, 2017; Yang, 2023). There are two major types of localism (Chen & Szeto, 2015; So & Ip, 2020), one is civic localism based on inclusive democratic values (Veg, 2017), and the other is anti-mainland localism that led to anti-immigration protests starting in the 2010s (Yuen & Chung, 2018). Given the severe economic inequality in Hong Kong, one may wonder about the impact of economic inequality on the rise of social movements and localism in recent years.
This study aims to investigate the political impact of economic inequality in Hong Kong, with a specific focus on its effects on social protests and demands for political autonomy. With an original random sample survey done in 2017, we adopt statistical analyses to examine the political effects of economic inequality from three perspectives, namely, its impacts on citizens’ preference for redistribution, their preference for political autonomy, and their propensity to join social protests. We find that the perceived economic inequality is positively correlated with citizens’ preference for redistribution and their preference for political autonomy. However, economic inequality does not exhibit a significant correlation with the propensity of citizens to engage in protest activities. We find that political values and citizens’ identities influence their propensity to join protests. Our study thus sheds light on the political effect of inequality in a developed but contentious society.
2 The Political Consequences of Economic Inequality in Hong Kong
Some literature has examined economic inequality in Hong Kong and its impacts. Regarding the study of economic inequality in Hong Kong, the literature suggests that Hong Kong citizens have a high tolerance for economic inequality. Based on surveys from 1988 to 2006, research showed that perceived income disparities were persistently high in Hong Kong; many attributed income inequality to individual factors and blamed the poor (Wong et al., 2009). Based on a survey conducted in 2007, Wu argued that Hong Kong citizens had a higher degree of tolerance for income inequality than citizens of mainland China, which could be attributed to the laissez-faire capitalism adopted since the British colonial era (Wu, 2009).
Advertisement
The most direct effect of economic inequality may be its impact on citizens’ support for redistribution. The literature suggests a weak link between economic inequality and the demand for redistribution. In a survey study conducted in 2013, Wu and Chou (2017) found that the perceived magnitude of income inequality was positively correlated with support for redistribution in a bivariate analysis, but this correlation was not significant when other variables were added in multiple and logistic regressions.
Economic inequality could further contribute to the rise of localism that calls for political autonomy. Localism focuses on preserving Hong Kong’s identity and autonomy, with goals ranging from greater autonomy to independence (Kwong, 2016). The perceived threat from mainland China contributes to the rise of a Hong Kong-based identity (Ho, 2022; Fong, 2017) and leads citizens with the Hong Kong identity to participate in protests (Chan, Nachman, and Mok, 2021). Based on qualitative interviews, Nagy (2015) argued that some citizens believed the inflow of mainland professionals and investors had negative impacts, such as causing the property boom, which led to economic and social inequality. As a result, the perceived social inequality gave rise to localism as a response to tackle the problem. However, no quantitative study has yet been carried out to examine the impact of economic inequality on citizens’ support for political autonomy.
Economic inequality may influence citizens’ participation in politics through institutionalized channels. Some scholars have tried to examine the impact of economic inequality on citizens’ voting participation. Based on a novel district-level income inequality dataset, a recent study found that individuals who lived in districts with higher income inequality in Hong Kong were less likely to vote in elections (Wong & Wong, 2022). Economic inequality could hence have a negative impact on citizens’ participation in institutionalized channels.
The literature has presented a mixed picture of the impact of economic inequality on social protests in Hong Kong. Based on a survey conducted in 2014, Wong (2017) suggested that citizens’ perception of the deteriorating income equality was positively correlated with their support for the Umbrella Movement in 2014. Sing (2020) found that citizens’ perceived economic performance was not correlated with their participation in the Anti-ELAB movement in 2019. More specifically, citizens’ perception that “economic problems are the main reasons for the youths’ dissatisfaction” was not correlated with their participation in the Anti-ELAB movement, and citizens’ perception that “housing is the main cause for the youths’ dissatisfaction” was not correlated with their participation in the movement. The literature provided valuable insights into how economic factors influence citizens’ support for or participation in a particular movement in Hong Kong. However, an examination of the impact of economic inequality on citizens’ general experiences and propensity to participate in social protests remains to be conducted.
In summary, regarding the political effects of perceived economic inequality on income redistribution, the picture presented in the literature of Hong Kong studies is blurred. In terms of the political effects of income inequality on the rise of localism, a quantitative study remains to be done. Regarding the influence of economic inequality on social protests, the literature discusses how this inequality affects the levels of support for or participation in a specific movement, but a general inquiry into economic inequality on citizens’ propensity to participate in social protests has yet to be done.
3 The Perception of Economic Inequality and its Political Effects
In the existing empirical literature on the impacts of economic inequality, as Lei Yawen highlights (Lei, 2020, 1), some scholars argue that it is perceived economic inequality rather than the actual level of inequality that has important political consequences (Lei, 2020; Gimpelson & Treisman, 2018; Whyte, 2016). Perceived economic inequality matters: many citizens are shaped by their misperception of income inequality since they have no idea of the actual income distribution (Gimpelson & Treisman, 2018). Empirical studies also find that perceived economic inequality significantly correlates with the demand for redistribution and class conflict, while the impact of actual inequality is tenuous (Alesina et al., 2018). However, it should be noted that in some recent studies, scholars find that the actual level of inequality can influence citizens’ political participation and political trust; inequality may influence citizens through many channels even though the inequality is not perceived (Wong & Wong, 2022; Zhou & Jin, 2018).
Based on the literature on the political effects of inequality, this study examines the effects of perceived economic inequality on three dependent variables: preference for redistribution, support for political autonomy, and propensity for joining social protests. It should be noted that perceived economic inequality could influence these dependent variables through many channels. For instance, perceived income inequality could lead to relative deprivation, or perceived income inequality could make some citizens attribute other social problems, such as the high crime rate, to inequality; some citizens might view economic inequality as unjust. Based on the practice in the literature, we do not specify all intermediate variables to examine the political effects of economic inequality (Zhou & Jin, 2018, 1039). In the following sections, we will first review the factors that are likely to influence the abovementioned dependent variables and then raise our hypotheses.
3.1 Explaining the Preference for Redistribution
The literature has identified two main approaches that explain people’s preferences for redistribution. The first approach argues that individuals’ values and beliefs matter in shaping their support for government redistribution (Fong, 2001; Svallfors, 2012). First, perceived inequality is positively correlated with support for government redistribution. However, actual inequality, measured by the actual income Gini coefficient, is not significantly correlated with the preference for government redistribution (Gimpelson & Treisman, 2018). Second, individuals’ values that define the proper relationship between the state and individuals influence their support for government redistribution (Blekesaune & Quadagno, 2003). More specifically, if individuals attribute inequality and poverty to structural factors such as social injustice, they are more likely to demand government redistribution. In contrast, if people attribute inequality and poverty to individuals, such as their ability and hard work, they are less likely to demand government redistribution. Moreover, the perception of social mobility also matters. People are more likely to tolerate economic inequality when they are opportunistic about upward chances (Larsen, 2016; Whyte, 2016). Fourth, the literature argues that identity matters (Keely & Tan, 2008; Klor & Shayo, 2010). It is argued that national identification could reduce support for redistribution, as poor people’s strong national identification may undermine their class identification (Shayo, 2009).
The second approach to examining individuals’ preference for redistribution treats individuals as rational self-interested people. Citizens’ calls for redistributive policies grow in proportion to the risks they face (Rehm, 2011). For example, individuals who perceive themselves as the lower class are more likely to support income redistribution (Svallfors, 2004). Individuals’ age also influences their preference for government redistribution; the elderly tend to need more public health and social assistance resources and are thus likely to support more generous government redistribution (Busemeyer et al., 2009).
Based on the abovementioned literature and our major research interest, we propose the first main hypothesis:
Hypothesis 1
People who are concerned about economic inequality are more likely to support government redistribution.
3.2 Explaining the Preference for Political Autonomy
In the study of people’s preference for political autonomy, scholars have found that subjective economic evaluations, identity, and material interests matter (Hierro & Queralt, 2021; Howe, 1998; Rodon & Guinjoan, 2018). First, subjective economic evaluations are expected to influence public opinion on political autonomy. Individuals who are dissatisfied with the current economic situation are more likely to support the change in the status quo (Hooghe & Marks, 2004). Second, identity plays a major role in driving people’s preference for political autonomy. When national identity and regional identities coexist, regional identity contributes to individuals’ preference for decentralization or secession (Rodon & Guinjoan, 2018). Third, people’s material interests influence their support for regional autonomy or secession. For instance, in the study of the preference for secession in Catalonia, individuals employed in industries and companies focused on the Spanish market were predominantly against secession, as it would introduce significant disruptions to their business operations (Hierro & Queralt, 2021).
In the case of Hong Kong, the rise of localism suggests that some citizens tend to blame the intervention from mainland China for causing various socioeconomic problems (Kwong, 2016; Nagy, 2015). Based on the above literature and the context of Hong Kong, we propose the second main hypothesis:
Hypothesis 2
People who are concerned about economic inequality in Hong Kong are more likely to demand political autonomy.
3.3 Explaining Protest Propensity
Protest propensity is an individual’s willingness to join a protest (Su & Feng, 2013). In the literature explaining why people join protests, scholars have acknowledged that economic inequality can be a source of grievance and relative deprivation that drives people to join protests (Goldstone, 1982; Gurr, 1968). Economic inequality can create structural barriers that prevent low-income citizens from accessing key opportunities and resources, such as schooling and healthcare (Suárez Álvarez and López Menéndez, 2020; Pop et al., 2013). The economic inequality could make low-income citizens feel frustrated and even anguished when comparing their own lives with those of the rich, leading to a sense of relative deprivation that pushes them to join social protests. Moreover, economic inequality can contribute to other social problems, such as a high crime rate, causing citizens’ discontent. Some citizens could take economic inequality as social injustice, which motivates them to join street protests. Therefore, the “social volcano thesis” argues that economic inequality could push citizens to join protests and social revolutions (Zhou & Jin, 2018). For instance, the huge gap in individual wealth in Egypt after the Second World War led to a call for an egalitarian society and became a cause of the Egyptian Revolution in 1952 (Davies, 1962, 13–14).
Scholars have also argued that individuals’ identities and political values matter. Identity is the way that individuals categorize themselves and others. When identity is politicalized, a large number of people can mobilize to protect the interests of their identity group. For instance, nationalism led to waves of anti-Japan protests in China in 2012 (Wallace & Weiss, 2015). In Northern Ireland in the 1970s and Rwanda in the 1990s (Bhavnani & Backer, 2000), politicalized identity further contributed to ethnic violence. In addition to individuals’ identities, their ideologies and value commitments are crucial for movement mobilization. In the study of Western industrialized societies, scholars have argued that the rise of post-materialist values due to modernization has contributed to social activism (Inglehart & Catterberg, 2002; Norris, 2002; Dalton et al., 2010). In the field of Hong Kong studies, scholars find that post-materialism correlates with citizens’ participation in protests, boycotts, and strikes (Charm & Lin, 2023; Cheng, Chung, and Cheng, 2023). Recent research further suggests that localism is positively correlated with support of the Anti-ELAB Movement, but not correlated with actual participation in the movement (Wong, Zheng, and Wan, 2023). Therefore, this study will incorporate the variables of identities and political values as control variables.
Individuals’ socio-economic and class statuses could also influence their protest propensity. Research indicates that certain demonstrations are associated with individuals who have lower educational levels, earn less income, and face higher unemployment rates compared to the general population (Downes, 1970), while some protests are associated with people with higher socioeconomic status (Mason & Murtagh, 1985). Although the impact of socioeconomic status on protest propensity is inconclusive in the literature, this study will incorporate variables of socioeconomic status as control variables.
Based on the literature and the focus of our research on economic inequality, we propose the third main hypothesis:
Hypothesis 3
People who are concerned about economic inequality are more likely to join protests.
4 Empirical Analysis
4.1 Data and Measures
The data for this project were collected through a telephone survey using random sampling by the Hong Kong University Public Opinion Programme in February 2017. The interviewees were adults in Hong Kong who were 18 years old or above. The interviewees were able to speak the main local language, Cantonese. For random sampling, first, we randomly selected telephone numbers as seed numbers based on the latest residential telephone directories. In the next step, we obtained the telephone numbers for interviews by randomly adding 1 or 2 to the seed numbers or subtracting 1 or 2 from the seed numbers. After picking the telephone numbers, our telephone interviewers called the selected numbers. If there was more than one respondent in a household who fulfilled our criteria, our interviewers selected the respondent whose birthday was closest to the interview date. The sample size of our study was 1002, and the response rate was 72.3%.
4.2 Dependent Variables
Our analysis included three dependent variables (The descriptive statistics for all variables are shown in Table 1). The first dependent variable was the preference for income redistribution. In our questionnaire, one question asked “Whether the government should reduce the differences in income between people with high incomes and those with low incomes”. Respondents could choose among the categories from 1 to 7, with 1 indicating that the government should absolutely reduce the income difference and 7 indicating that the government should absolutely not reduce the income difference. International surveys, such as the International Social Survey Program (ISSP), use the same question to measure the preference for redistribution. This measurement of citizens’ preference for redistribution has been commonly used in the literature (Gimpelson & Treisman, 2018). As the proportional odds assumption (i.e., parallel regression assumption) for ordinal logistic regression did not hold in our dataset (p < 0.001 in the Brant test of parallel regression assumption), one option was to use a nonordinal alternative (Williams, 2006, 62), and this study opted for binary logistic regression. Hence, we recoded this variable into a binary variable in which support for government redistribution was coded as 1, while indifference or no support for government redistribution was coded as 0 (1–3 were recoded as 1, while 4–7 were recoded as 0). In our robustness checks, we further tried to recode indifference or support as 1 and no support as 0 (1–4 were recoded as 1, while 5–7 were recoded as 0). Another ordinal alternative when the proportional odds assumption did not hold was to use a partial proportional odds model; we used this model in the robustness checks (Williams, 2006).
Table 1
Descriptive statistics
Variables | Percentage/Mean (SD) |
|---|---|
Dependent variables | |
Preference for redistribution | 0.49 (0.50) |
1 = Support | 48.7% |
0 = Indifference or no support | 51.3% |
Preference for autonomy | 0.77 (0.42) |
1 = Support | 77.4% |
0 = Indifference or no support | 22.6% |
Propensity to join protests | 0.55 (0.50) |
1 = Yes | 54.6% |
2 = Will never join | 45.5% |
Independent variables | |
Income differences in Hong Kong are too high | 4.39 (0.86) |
Control variables | |
Obey the government | 2.37 (1.2) |
Equal say for all | 4.30 (0.99) |
Identity | 4.44 (1.83) |
Optimism about the future | 3.28 (0.69) |
Injustice as primary cause of poverty | 0.31 (0.46) |
Perceived unfairness of elites’ privilege | 3.70 (1.26) |
Meritocratic explanation | 2.70 (1.28) |
Optimism about mobility | 3.63 (1.09) |
Class | 3.90 (0.89) |
Age group | 4.60 (1.78) |
Education | 4.26 (2.29) |
Male | 46.4% |
Married | 60.7% |
The second dependent variable was the preference for political autonomy. We used one question in the questionnaire to measure the preference for political autonomy: “Hong Kong’s political affairs should be completely free from intervention from mainland China”. The responses fell into one of five categories: strongly agree/agree/neither agree nor disagree/disagree/strongly disagree. As the proportional odds assumption (i.e., parallel regression assumption) for ordinal logistic regression was violated (p < 0.05 in the Brant test of the parallel regression assumption), we recoded this variable into a binary variable in which agreement on this statement was recoded as 1, and indifference or disagreement was coded as 0. In our robustness checks, we further recoded indifference or support as 1 and disagreement as 0. Moreover, we further ran the partial proportional odds model in the robustness checks.
The third dependent variable measured citizens’ propensity to join protests. In our questionnaire, one question asked whether the interviewees had joined protests or demonstrations. The interviewees could choose among the following options: very often/sometimes/maybe join such activities/will never join such activities. As this variable measured respondents’ experiences and willingness to join protests, we took this variable to measure citizens’ propensity to join protests. As the proportional odds assumption (i.e., parallel regression assumption) for ordinal logistic regression was violated (p < 0.001 in the Brant test of the parallel regression assumption), we also recoded this variable to be a binary variable in which very often/sometimes/maybe join such activities were recoded as 1 to indicate a propensity to join protests, while the answer “will never join such activities” was recoded as 0 to indicate the propensity to not join social protests. In the robustness checks, we further incorporated the partial proportional odds model.
4.3 Independent Variables
Our main interest in the independent variable was citizens’ perceptions of inequality. We measured citizens’ perceptions and concerns about inequality through the item “Income differences in Hong Kong are too high”. The responses fell into one of five categories: strongly agree/agree/neither agree nor disagree/disagree/strongly disagree (we recoded this variable so that 1-strongly disagree to 5-strongly agree). The same question has been used in international surveys such as the ISSP. Many existing studies have used this single variable to measure citizens’ perceptions of economic inequality (Mijs, 2019; Larsen, 2016; Lei, 2020).
To study the political consequences of economic inequality, we used the perception of inequality as the main covariate in the analysis. We examined the three dependent variables with several covariates. First, to study the extent to which citizens think inequality was caused by structural factors, we incorporated a dummy variable for whether the respondents thought social injustice was the primary cause of poverty (1-agree, 0-not agree). We also incorporated a question to ask whether the respondents agreed with the statement that social inequality was mainly caused by control and manipulation carried out by a small group of people in power (5-strongly agree, 1-strongly disagree). Second, to study the extent to which citizens attributed economic inequality to meritocracy, we included a covariate based on responses to the following statement: “Social injustice was mainly caused by individuals’ gift and abilities” (5-strongly agree, 1-strongly disagree). Third, we also included a covariate about perceived mobility based on the following item: “With hard work, people can overcome all difficulties and succeed” (5-strongly agree, 1-strongly disagree). Fourth, to measure citizens’ optimism about the future, we added a covariate based on the response to the following question: “For the next 12 months, you think you and your family members will have a better life, a worse life, and no difference?” (5-strongly agree, 1-strongly disagree). Fifth, to measure citizens’ attitudes toward identity, we asked the respondents to reveal their identity through categories 1 to 7, in which 1 represented “I’m a Hong Konger”, 4 represented “I’m both a Hong Konger and a Chinese”, and 7 represented “I’m a Chinese” (we recoded this variable so that 7- “I’m a Hong Konger”, 1- “I’m a Chinese”). Moreover, we added two indicators to measure citizens’ political values. The first indicator asked whether the respondents agreed that people with different levels of education should have the same right to speak. Responses fell into one of the five categories: strongly agree/agree/neither agree nor disagree/disagree/strongly disagree (we recoded this variable so that 1 = strongly disagree to 5 = strongly agree). The second question asked whether the respondents agreed that the decision-makers in government were similar to the masters of the house and that people should obey their decisions. The responses were put into one of five categories: strongly agree/agree/neither agree nor disagree/disagree/strongly disagree (we recoded this variable from 1 = strongly disagree to 5 = strongly agree). Finally, we further included a set of control variables, including self-perceived class, age, education level, gender, and marital status.
Our analytical model used the logistic regression in the form of the following equation, in which the dependent variables were binary variables of preference for redistribution, preference for political autonomy, and propensity to join protests:
$$\begin{gathered} \log it \, (P(DV = 1)) = \beta_{0} + \beta_{1} Inequality + \beta_{2} ObeyGovernment \hfill \\ \quad \quad \quad \quad \quad \quad \quad \quad \;\quad \; + \beta_{3} EqualSay + \beta_{4} Identity + \beta_{5} Optimism + \beta_{6} Injustice \hfill \\ \quad \quad \quad \quad \quad \quad \quad \quad \;\quad \; + \beta_{7} Unfairness + \beta_{8} Meritocracy + \beta_{9} Mobility + \beta_{10} Class \hfill \\ \quad \quad \quad \quad \quad \quad \quad \quad \;\quad \; + \beta_{11} Age + \beta_{12} Education + \beta_{13} Male + \beta_{14} Married + e \hfill \\ \end{gathered}$$
5 Results
5.1 Preference for Redistribution
The binary logistic regressions from Model 1 to Model 3 examined covariates that were associated with citizens’ preference for redistribution. In Model 1, we incorporated only the main independent variable and demographic variables. In Model 2, we added all control variables. In Model 3, we further controlled the other two dependent variables (Table 2). In the robustness checks, we further adopted a partial proportional odds model by using the ordinal dependent variable (Model 12) and recoded the dependent variable (Model 10), and our findings remained robust.
Table 2
Binary logistic regressions of citizens’ preference for redistribution
Model 1 | Model 2 | Model 3 | |
|---|---|---|---|
Perception of inequality | 0.551*** | 0.343*** | 0.349*** |
Preference for autonomy | − 0.110 | ||
Propensity to join protests | 0.062 | ||
Obey the government | − 0.083 | − 0.081 | |
Equal say for all | 0.076 | 0.077 | |
Identity | 0.083+ | 0.086+ | |
Optimism about future | 0.063 | 0.056 | |
Injustice as primary cause of poverty | 0.407* | 0.412* | |
Perceived unfairness of elites’ privilege | 0.159* | 0.160* | |
Meritocratic explanation | − 0.180** | − 0.180** | |
Optimism about mobility | − 0.126+ | − 0.126+ | |
Class | 0.089 | 0.087 | |
Age | − 0.180*** | − 0.109+ | − 0.108+ |
Education | 0.118** | 0.112** | 0.106* |
Male | 0.445** | 0.399* | 0.400* |
Married | − 0.007 | 0.073 | 0.069 |
Adjusted R2 | 0.09 | 0.13 | 0.13 |
Our regression models show that citizens’ perceptions of economic inequality are positively correlated with support for the government to address the income difference between citizens with high income and citizens with low income. In Model 2, all else being equal, a one-unit increase in the perception of inequality is expected to multiply the odds of preferring a redistributive policy by 1.409 (p < 0.001). This finding confirms our first hypothesis that people who are concerned about economic inequality are more likely to support government redistribution.
We find that citizens who strongly identify as Hong Kongers are more likely to demand a redistributive policy, which suggests an interesting correlation between identity and support for a redistributive policy (p < 0.1). This may illustrate that citizens who have strong Hong Kong identities are dissatisfied with the existing structure of the political economy and would like the government to address the income gap. Moreover, the higher the respondents’ education levels are, the stronger their preference for a redistributive policy (p < 0.01). This could suggest that citizens with a higher level of education are more likely to think it is the government’s responsibility to address economic inequality.
Our findings support the literature that social-economic values matter. More specifically, citizens who think injustice is the primary cause of poverty demonstrate a stronger preference for redistribution (p < 0.05). We also find a positive correlation between citizens who perceive the unfairness of privileges granted to those with political power and the preference for redistribution (p < 0.05). Citizens who perceive political elites as having unfair privilege are more likely to support the government in addressing the income difference. In contrast, citizens who attribute social injustice to individuals’ gifts and abilities are less likely to support redistributive policies (p < 0.01). Citizens’ perceptions of mobility are also negatively correlated with their support for redistributive policies (p < 0.1). Citizens who are optimistic about mobility are less likely to support redistributive policies, as they believe they can achieve social mobility through their own efforts.
We further find that age is negatively correlated with support for a redistributive policy. It is interesting to note that in the case of Hong Kong, younger citizens are more supportive of the view that the government should address the income difference between the rich and the poor (p < 0.1). Our findings confirm existing research that older adults in Hong Kong tend to hold a negative view of welfare redistribution for the poor population (Wu & Chou, 2021). We also find that male citizens are more likely to support redistributive policies (p < 0.05).
5.2 Preference for Political Autonomy
Regarding the preference for political autonomy, in Model 4, we incorporated only the independent variable and demographic variables. In Model 5, we added all control variables. In Model 6, we further controlled the other two dependent variables (Table 3). In the robustness checks, we further performed a partial proportional odds model by keeping the ordinal dependent variable (Model 13) and recoding the dependent variable (Model 11), and our findings remained robust.
Table 3
Binary logistic regressions of citizens’ preferences for political autonomy
Model 4 | Model 5 | Model 6 | |
|---|---|---|---|
Perception of inequality | 0.425*** | 0.258* | 0.269* |
Preference for redistribution | − 0.111 | ||
Propensity to join protests | 0.260 | ||
Obey the government | − 0.123 | − 0.112 | |
Equal say for all | 0.123 | 0.123 | |
Identity | 0.315*** | 0.310*** | |
Optimism about future | − 0.246+ | − 0.265+ | |
Injustice as primary cause of poverty | 0.266 | 0.286 | |
Perceived unfairness of elites’ privilege | 0.112 | 0.112 | |
Meritocratic explanation | − 0.139+ | − 0.133+ | |
Optimism about mobility | − 0.053 | − 0.052 | |
Class | − 166 | − 0.167 | |
Age | − 0.161* | − 0.090 | − 0.085 |
Education | − 0.084+ | − 0.146** | − 0.155** |
Male | 0.307+ | 0.293 | 0.291 |
Married | − 0.231 | − 0.077 | − 0.082 |
Adjusted R2 | 0.04 | 0.13 | 0.14 |
We find that citizens’ concerns about economic inequality are positively correlated with support for the statement that Hong Kong’s political affairs should be completely free from intervention from the mainland (Model 5) (p < 0.05). In Model 5, a one-unit increase in the perception of inequality is expected to multiply the odds of preferring political autonomy by 1.294 with other variables controlled in the model. The finding supports our second hypothesis that citizens who are concerned about economic inequality are more likely to demand political autonomy. Our findings fit the literature suggesting that some Hong Kong citizens may attribute social and economic inequality to the inflow of mainland professionals and investors, which then leads to the rise of localism that calls for political autonomy (Nagy, 2015).
We find that citizens’ identities and education levels are correlated with their preference for political autonomy. First, citizens who identify as Hong Kongers are more likely to have a stronger preference for political autonomy (p < 0.001). In Model 5, all else being equal, a one-unit increase in the identity of being a Hong Konger is expected to multiply the odds of preferring political autonomy by 1.370. This finding fits our intuition that individuals’ identities will have an impact on their preference for political autonomy. Second, we find that people with a lower level of education are more likely to think that Hong Kong’s political affairs should be completely free from intervention from the mainland. Moreover, we find that citizens who are optimistic about the future and trust meritocracy are less likely to support the statement that Hong Kong’s political affairs should be completely free from intervention from the mainland.
We further examine the potential mediation effect that could influence the preference for autonomy. The potential mediation effect might involve how the preference for economic redistribution influences the preference for autonomy, as the impacts of the independent variable (perceived inequality) are statistically significant for both variables. Based on the existing literature, the book entitled “Hong Kong as a City-State” contributes to the growth of localism in Hong Kong (Kwong, 2016, 65). In the book, the author argues that the neo-liberalism supported by the Hong Kong government has led to predatory capitalism in Hong Kong, which only benefits tycoons while undermining the interests of small businesses and ordinary people. His solution to the economic inequality in Hong Kong is that Hong Kong should act as a “city-state” with much greater political autonomy.
Hence, for citizens who follow localism, the perceived economic inequality may push them to call for a higher level of redistribution, which may result in an increased preference for political autonomy as a means to alter redistributive policies. This study thus tested whether the preference for redistribution is a mediator between perceived inequality and preference for political autonomy. In the mediation analysis, the Monte Carlo test following Zhao, Lynch and Chen’s approach is not significant, showing that there is only a direct effect from perceived economic inequality on preference for political autonomy. The Baron and Kenny approach to testing mediation also shows that there is no mediation.
5.3 Propensity for Joining Protests
We further examined whether the perception of economic inequality was associated with citizens’ experiences and propensity to join protests. In Model 7, we incorporated only the independent variable and demographic variables. In Model 8, we added all control variables. In Model 9, we further controlled the other two dependent variables (Table 4). In the robustness checks, we further performed a partial proportional odds model by keeping the ordinal dependent variable (Model 14), and our findings remained robust.
Table 4
Binary logistic regressions of citizens’ propensity for joining protests
Model 7 | Model 8 | Model 9 | |
|---|---|---|---|
Perception of inequality | 0.102 | − 0.048 | − 0.062 |
Preference for redistribution | 0.067 | ||
Preference for political autonomy | 0.209 | ||
Obey the government | − 0.286*** | − 0.282*** | |
Equal say for all | 0.084 | 0.080 | |
Identity | 0.165** | 0.154** | |
Optimism about future | 0.298* | 0.305* | |
Injustice as primary cause of poverty | 0.031 | 0.012 | |
Perceived unfairness of elites’ privilege | 0.087 | 0.081 | |
Meritocratic explanation | − 0.217*** | − 0.210** | |
Optimism about mobility | − 0.069 | − 0.066 | |
Class | 0.040 | 0.045 | |
Age | − 0.203*** | − 0.116* | − 0.111+ |
Education | 0.234*** | 0.238*** | 0.242*** |
Male | 0.256 | 0.233 | 0.221 |
Married | 0.027 | 0.135 | 0.135 |
Adjusted R2 | 0.11 | 0.17 | 0.17 |
Our logistic regression models (Model 8) indicate that citizens’ concerns about income inequality in Hong Kong are not correlated with their propensity to join protests. Our finding thus rejects the third hypothesis. Perceived economic inequality is not correlated with the propensity to participate in social protests in our model. One possible explanation is that most social movements in Hong Kong have focused their efforts on advocating for political and civil rights, and citizens are not driven by economic grievances to join social protests. Thus, citizens typically do not connect economic inequality to participation in social protests.
Our regression models suggest that citizens’ political values and their identities have significant impacts on their propensity to join protests. Based on the binary logistic regression in Model 8, citizens’ support for the statement that citizens should always obey the government is negatively correlated with their experience and propensity to join protests (p < 0.001). In contrast, citizens who identify as Hong Kongers are more likely to join protests (p < 0.01). A one-unit increase in perceived identity as a Hong Konger is expected to multiply the odds of joining protests by 1.179. Moreover, it is interesting to find that citizens who support the meritocratic explanation that social injustice is caused by an individual’s gifts and abilities are less likely to join protests (p < 0.001). Their support for a meritocratic explanation of social injustice may reduce their grievances and attribute injustice to individual factors, which makes them less likely to join protests. The findings suggest that protests in Hong Kong are primarily driven by citizens’ political values and identities (Tang & Cheng, 2021), while perceived economic inequality does not play a significant role in protest participation.
The models also illustrate that citizens who are optimistic about the future are more likely to join social protests (p < 0.05). We further find that citizens with higher education levels are more likely to join protests (p < 0.001). In contrast, the age of citizens is inversely correlated with their propensity to participate in protests (p < 0.05). Younger people are more likely to join protests, which fits our intuition.
6 Conclusion
This study examines the political effects of economic inequality in Hong Kong. Based on a random sampling survey, we first find that perceived economic inequality is positively correlated with citizens’ support for government redistribution. Our finding fits the existing cross-national literature on the impacts of perceived economic inequality on redistribution support (Gimpelson & Treisman, 2018), and the finding makes new contributions to Hong Kong studies as the impact of perceived economic inequality on support for the redistributive policy in Hong Kong is blurred in the literature (Wu & Chou, 2017; Chen, Wu, and Lin, 2023). Our finding further argues that perceived economic inequality is positively correlated with citizens’ support for political autonomy. The literature argues that some citizens attribute social and economic inequality to the inflow of mainland professionals and capital, which then leads to the rise of localism (Nagy, 2015). Our research confirms the linkage between economic inequality and support for political autonomy based on survey data. However, this study finds that perceived economic inequality has no statistically significant relationship with citizens’ propensity to join social protests. Our finding suggests that economic inequality is not associated with social protests in Hong Kong; rather, identity politics and political values are positively correlated with citizens’ propensity to join social protests.
Particularly, the impact of identity politics is thought-provoking. Our models illustrate that identity as a Hong Konger has a positive correlation with all three dependent variables, namely, citizens’ support for redistribution, their support for complete political autonomy, and their propensity to join social protests. The findings confirm that identity politics plays an important role in shaping Hong Kong politics. With a rising Hong Kong identity in recent years, such an identity change could further increase citizens’ demands for policy change and political change in the coming years.
Our research also presents an interesting picture of generational politics in Hong Kong. Our models show that the younger generation in Hong Kong tends to support more welfare redistribution and complete political autonomy and is more likely to join protests. Regarding welfare redistribution, our finding presents an intriguing phenomenon in the case of Hong Kong, it is not the elderly, traditionally perceived as more economically vulnerable, who exhibit a preference for heightened welfare redistribution (see also Wu & Chou, 2021).
Another finding deserving attention is the impact of education. Our models illustrate that the higher the level of education a person receives, the more likely he or she will be to demand more welfare redistribution and participate in social protests in Hong Kong. This indicates that individuals with a higher level of education are more concerned with the income gap and social injustice. However, individuals with higher education levels are less likely to demand complete political autonomy from mainland China.
This research has significant policy implications. The Hong Kong SAR government should pay good attention to addressing the problem of economic inequality, as perceived economic inequality could lead citizens to call for not only more welfare redistribution but also political change. Moreover, our findings suggest that increasing welfare redistribution and addressing the income difference between the rich and the poor can help to address the concerns of citizens who are younger and more well-educated and who play a significant role in Hong Kong society. As citizens’ political values and identities as Hong Kongers have a significant positive correlation with their participation in social protests, this study further highlights that the government needs to understand citizens’ democratic values and identities to preempt serious discontent and even social unrest in the future.
It is worth noting that many scholars and opinion leaders in mainland China tend to attribute the social protests in Hong Kong to economic factors. The argument is that economic inequality, the huge gap between the rich and the poor, and the lack of social mobility of young citizens cause the social movements in Hong Kong. The economic explanation for the social protests in Hong Kong may have some impacts on the Chinese central government’s policy toward Hong Kong. The finding of this article instead challenges the economic explanation of social protests through solid empirical data.
Hong Kong has changed on politico-socio-economic fronts dramatically in recent years. Our survey data were collected in 2017, but we believe our findings still hold. Recent research on the Anti-ELAB Movement in 2019 confirms that economic factors had no significant impact on citizens’ participation in the movement, which fits our findings (see Sing, 2020). Although economic inequality is not correlated with protest participation, our research suggests that economic inequality has significant political-social implications, which need to be addressed by the government and Hong Kong people.
Acknowledgement
For helpful advice on the paper, we would like to thank Chen Chuanmin, Hu Peng, Hu Yue, Ji Chengyuan, Zeng Yu, and particularly Su Zheng. We also appreciate the feedback received from the 3rd Shanghai Forum on Chinese Political Economy Research at Shanghai Jiao Tong University and the 2023 Lien International Conference at Nanyang Technological University in Singapore.
Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by/4.0/.
Publisher's Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.