Abstract
Using data from the Gallup World Poll, we examined whether progressive taxation is associated with increased levels of subjective well-being. Consistent with Rawls’s theory of justice, our results showed that progressive taxation was positively associated with the subjective well-being of nations. However, the overall tax rate and government spending were not associated with the subjective well-being of nations. Furthermore, controlling for the wealth of nations and income inequality, we found that respondents living in a nation with more-progressive taxation evaluated their lives as closer to the best possible life and reported having more positive and less negative daily experiences than did respondents living in a nation with less-progressive taxation. Finally, we found that the association between more-progressive taxation and higher levels of subjective well-being was mediated by citizens’ satisfaction with public goods, such as education and public transportation.
What is a good society? From the time of Plato on, many philosophers have answered that a good society is “a happy society,” or a society in which citizens feel satisfied with their lives (e.g., Bentham, 1789/2008; Plato, trans. 1993). Over the last two decades, social scientists have empirically investigated the characteristics of happy societies. They have found that happy societies are wealthy societies (e.g., Diener, Diener, & Diener, 1995; Stevenson & Wolfers, 2008; Veenhoven, 1995), in which citizens have a subjective sense of freedom (Inglehart, Foa, Peterson, & Welzel, 2008), trust (Helliwell, 2007), and social support (Oishi & Schimmack, 2010; see Diener, Kahneman, & Helliwell, 2010, for a review).
In terms of the role of government in shaping a happy society, government efficiency and the lack of corruption are associated with increased levels of citizen’s subjective well-being (e.g., Bjørnskov, Dreher, & Fischer, 2007; Helliwell & Huang, 2008). In contrast, the size of the government (e.g., as indexed by government spending as a percentage of gross domestic product, or GDP, and by welfare spending) is associated with decreased levels of subjective well-being (Bjørnskov et al., 2007; Veenhoven, 2000; but see Flavin, Pacek, & Radcliff, 2011, for an exception). However, to our knowledge, no previous research has examined the role of progressive taxation in the subjective well-being of nations.
The influential philosopher John Rawls (1971/1999) argued that a good society redistributes its wealth so that all its citizens have the same opportunity for future success in the form of equal access to public goods, such as quality education and health care. According to Rawls, then, progressive taxation (a higher tax rate for the rich than for the poor) is a more just policy than flat taxation (a constant tax rate across income categories). Using data from the Gallup World Poll, we examined whether progressive taxation is associated with the subjective well-being of nations, and if so, whether this link is explained by citizens’ satisfaction with public and common goods, such as public transportation, the school system, health care, roads and highways, good affordable housing, and air and water quality. Because progressive taxation might be associated with national wealth, income equality, the average tax rate, and government spending, we also tested whether these factors were associated with the subjective well-being of nations, and, if so, whether progressive taxation would predict the subjective well-being of nations above and beyond these confounding factors.
Method
For its 2007 world poll, the Gallup organization conducted representative surveys of 132 countries. On the basis of available tax information, we included data from 54 of these nations in our study; in addition, individual-level data were available from 59,634 respondents.
In the Gallup World Poll, subjective well-being was assessed for each nation using three types of items. First, a global-life-evaluation index was obtained using Cantril’s (1965) Self-Anchoring Striving Scale, which asks respondents to evaluate their current life on a scale ranging from 0 (worst possible life) to 10 (best possible life). Second, positivity of daily experiences was assessed by taking the mean of all answers (1 = yes, 0 = no) to the following 10 items: “Did you smile or laugh a lot yesterday?” “Were you proud of something you did yesterday?” “Did you learn or do something interesting yesterday?” “Did you feel enjoyment during a lot of the day yesterday?” “Did you feel love during a lot of the day yesterday?” “Would you like to have more days just like yesterday?” “Did you feel well-rested yesterday?” “Were you treated with respect all day yesterday?” “Were you able to choose how you spent your time all day yesterday?” and “Did you have good tasting food to eat yesterday?”
Then, negativity of daily experiences was assessed by asking respondents whether they felt worry, sadness, depression, boredom, anger, and shame “during a lot of the day yesterday.” We took the mean of all responses to create a score for each nation. Finally, we took the mean of respondents’ ratings of satisfaction (1 = satisfied, 0 = dissatisfied) with the following seven public and common goods: “public transportation system,” “the availability of quality health care,” “the availability of good affordable housing,” “educational system or the schools,” “roads and highways,” “quality of air,” and “quality of water.” The reliability estimates were .996 for global life evaluation, .986 for positivity of daily experiences, .980 for negativity of daily experiences, and .996 for satisfaction with public and common goods. 1
We obtained the highest and lowest tax rates for personal income from 54 of the nations surveyed in the Gallup World Poll from Worldwide-Tax.com (http://www.worldwide-tax.com/index.asp#partthree). We then calculated each nation’s degree of progressive taxation, which is defined as the difference between the highest and the lowest tax rates. Because of tax benefits and social security taxes, official tax rates for income could be misleading. To address this issue, we also obtained two types of “effective” tax rates from Taxing Wages 2007/2008 (Organisation for Economic Co-operation and Development, OECD, 2008): (a) personal income tax plus social security tax paid by earners and (b) personal income tax plus social security tax paid by earners minus tax benefits. This information was available for 27 of the 54 nations.
In addition, we calculated “effective” progressive taxation indices by taking the difference between the effective tax rates for individuals whose earning was 67% of the national average and individuals whose earning was 167% of the national average (two tax rates that the OECD reports in addition to the rate for the average earner). Because tax rates differ depending on family types, we first computed separately the effective progressive taxation rate for singles without children and for families with two children, and then we took the mean of these two rates. Because overall tax rate and the degree of progressive taxation could be confounded with overall government spending, we obtained the share of GDP that represents general government consumption spending for 69 nations from Bjørnskov et al. (2007). The 2007 per capita GDP 2 data for each nation were taken from the World Development Indicators online database (World Bank, 2009). We obtained each nation’s income inequality index (Gini coefficient) from the 2007 World Factbook (Central Intelligence Agency, 2007).
Results
Nation-level mediation analyses
For the sake of clarity, we first present results at the national level (see Table 1 for descriptive statistics and correlations among the key variables of this analysis). In this model, global life evaluation, positivity of daily experiences, and negativity of daily experiences were included as indicators of subjective well-being, whereas satisfaction with public goods was included as a mediator.
Descriptive Statistics and Correlations Among Key Variables at the Nation Level
Note: N = 54. Standard deviations are given in parentheses.
p < .05. **p < .01.
Supporting Rawls’s (1971/1999) theory of justice, our results showed that nations with more-progressive taxation were higher in global life evaluation and positivity of daily experiences and lower in negativity of daily experiences than nations with less-progressive taxation (see Fig. 1). This was true even when we used the two effective progressive taxation indices. The degree of progressive taxation calculated by using personal income tax and social security tax was positively associated with life evaluation, r(25) = .41, p < .05, and positivity of daily experiences, r(25) = .51, p < .01, but unrelated to negativity of daily experiences (r = −.23, n.s.). The degree of progressive taxation calculated by using personal income tax, social security tax, and tax benefits was also marginally positively associated with global life evaluation, r(25) = .36, p = .07, and significantly positively associated with positivity of daily experiences, r(25) = .46, p < .05, but unrelated to negativity of daily experiences (r = −.21, n.s.). Because the patterns of correlations between our three indices of progressive taxation were similar, we report the following results utilizing the first progressive tax index, which had the most data (N = 54 nations).
Scatter plot (with best-fitting regression line) showing mean global-life-evaluation rating as a function of progressive taxation (calculated as the difference between the highest and lowest tax rate; N = 54 nations).
One potential alternative explanation for the link between progressive taxation and the subjective well-being of nations is that nations with more-progressive taxation have higher overall rates of taxation, and their governments spend more money on various services than governments of nations with less-progressive taxation. This interpretation, however, did not explain our findings. Whereas more-progressive taxation was associated with higher subjective well-being of nations, the two types of effective tax rates for average earners (personal income tax + social security tax; personal income tax + social security tax – tax benefit) were both unrelated to global life evaluation, r(25) = −.17 and .03, respectively, ps > .41, and to negativity of daily experiences, r(25) = −.05, and −.19, respectively, ps > .31.
Furthermore, the two effective tax rates for average earners were marginally negatively associated with positivity of daily experiences, r(25) = −.45, p < .05, and r(25) = −.34, p = .08, respectively. Namely, residents of nations with higher effective tax rates for average earners reported marginally lower positivity of daily experiences than did residents of nations with lower effective tax rates for average earners. Likewise, replicating the findings of Bjørnskov et al. (2007), our results showed that government spending as a percentage of GDP was negatively associated with life evaluation, r(67) = −.46, p < .01, and positivity of daily experiences, r(67) = −.45, p < .01, but positively associated with negativity of daily experiences, r(67) = .26, p < .05. It is also noteworthy that government spending was negatively associated with satisfaction with public and common goods, r(67) = −.44, p < .01, and the two effective tax rates for average earners were unrelated to satisfaction with public and common goods, rs(25) = −.06 and .19, respectively, ps > .36. However, progressive taxation was positively associated with satisfaction with public and common goods (see Table 1). That is, it was not the overall tax rate or government spending but the degree of progressive taxation that was positively associated with the subjective well-being of nations.
Nations with progressive taxation could be richer nations with less income inequality. As Table 1 shows, however, the degree of progressive taxation was not associated with the GDP per capita or Gini coefficients. In contrast, GDP per capita and income equality were positively associated with global life evaluation. Thus, we next tested whether more-progressive taxation was associated with higher levels of subjective well-being, above and beyond the wealth of the nations and income inequality. Multiple regression analyses showed that nations with more-progressive taxation were indeed higher in global life evaluation, b = 0.02, 95% confidence interval (CI) = [0.007, 0.031], β = 0.27, t(51) = 3.09, p < .01, and positivity of daily experiences, b = 0.002, 95% CI = [0.001, 0.003], β = 0.44, t(51) = 3.46, p < .01, but lower in negativity of daily experiences, b = −.005, 95% CI = [−0.010, −0.000], β = −0.29, t(51) = −2.08, p < .05, above and beyond GDP per capita and income inequality (Gini coefficient) simultaneously. That is, if two nations were equally wealthy and income distribution was the same, people living in the nation with a more-progressive taxation policy were more satisfied with their lives in general and had more positive daily experience and fewer negative daily experiences than people living in the nation with the less-progressive taxation policy.
After eliminating the main alternative explanations regarding why residents of nations with a progressive taxation policy might be happy, we tested our central hypothesis: namely, that the link between progressive taxation and each of the indicators of subjective well-being was mediated by citizens’ satisfaction with public and common goods. To test this, we used the bias-corrected bootstrap method in Mplus (Version 4.21; Muthén & Muthén, 2007), with the resampling number set to 5,000. As Figure 2 shows, citizen’s satisfaction with public and common goods fully mediated the association between progressive taxation and the subjective well-being of nations, indirect effect = 0.03, 95% CI = [0.02, 0.04], z = 4.45, p < .01, for global life evaluation; indirect effect = 0.001, 95% CI = [0.001, 0.002], z = 3.26, p < .01, for positivity of daily experiences; and indirect effect = −0.005, 95% CI = [−0.009, −0.002], z = −2.79, p < .01, for negativity of daily experiences. That is, citizens of nations with more-progressive taxation were on average more satisfied with their lives, had more positive daily experiences, and had fewer negative daily experiences than citizens of nations with less-progressive taxation. This is true, at least in part, because the latter were more satisfied with public and common goods, such as the quality of education and the availability of health care.
Results of nation-level mediation analyses showing the influence of progressive taxation on three measures of the subjective well-being of nations: global life evaluation (top panel), positivity of daily experiences (middle panel), and negativity of daily experiences (bottom panel). The mediator in the models was citizens’ satisfaction with public and common goods (e.g., education, public transportation, health care, roads and highways, affordable housing, quality of air, and quality of water). The figure shows standardized regression coefficients. Asterisks and solid lines denote significant results (**p < .01).
Multilevel mediation analysis
The nation-level mediation analyses yielded a clear picture of the associations between progressive taxation and the well-being of nations. However, the aggregation of individual responses ignores within-nations variation. It also makes it impossible to control for individual-level variables, such as income and marital status, that are known to affect people’s subjective well-being. We thus conducted a multilevel mediation analysis (Preacher, Zyphur, & Zhang, 2010) using Mplus (Version 4.21; Muthén & Muthén, 2007) to test our central multilevel mediation hypothesis. As Figure 3 shows, each of the three indicators of subjective well-being was predicted by respondents’ income (28 income brackets, in the U.S. dollar, grand-mean-centered), sex (1 = male, 0 = female), age (grand-mean-centered), marital status (1 = married, 0 = not married), and satisfaction with public and common goods at the individual level of analysis. In addition, the path from income to satisfaction with public and common goods was added at the individual level (Level 1). At the nation level (Level 2), mean-adjusted subjective well-being and Level 1 slopes were predicted by log-linear GDP per capita, the degree of progressive taxation, national mean income, and national mean satisfaction with public and common goods. Furthermore, income and satisfaction with public and common goods were predicted from GDP per capita and progressive tax.
Results of the multilevel mediation analyses showing the effects of progressive taxation, gross domestic product (GDP) per capita, satisfaction with public and common goods, income, sex, age, and marital status on subjective well-being. Variables at the nation level are in gray boxes, and variables at the individual level are in white boxes. Significant paths (p < .05) are represented by solid lines (thick lines for variables of theoretical interest and thin lines for control variables), and nonsignificant paths are represented by dotted lines.
As predicted, we found a significant multilevel mediation effect (progressive taxation yields greater satisfaction with public and common goods, which in turn results in higher subjective well-being). This effect held for all three indicators of subjective well-being: indirect effect = 0.003, 95% CI = [0.002, 0.004], SE = 0.001, z = 4.78, p < .01, for global life evaluation; indirect effect = 0.001, 95% CI = [0.000, 0.001], SE = 0.000, z = 4.67, p < .01, for positivity of daily experiences; and indirect effect = −0.003, 95% CI = [−0.004, −0.002], SE = 0.001, z = −5.14, p < .01, for negativity of daily experiences. That is, residents of nations with more-progressive taxation were more satisfied with public and common goods, and therefore also evaluated their lives to be closer to the best possible life, had more positive daily experiences, and had fewer negative daily experiences than people living in nations with less-progressive taxation. There was no evidence that the effect of progressive taxation was mediated by income (|z|s < 0.67, ps > .50).
It is interesting that the well-known association between GDP per capita and subjective well-being (e.g., Diener et al., 1995) was mediated by two factors. The first was personal income, indirect effect = 0.306, 95% CI = [0.214, 0.397], SE = 0.047, z = 6.55, p < .01, for global life evaluation; indirect effect = 0.017, 95% CI = [0.011, 0.022], SE = 0.003, z = 5.74, p < .01, for positivity of daily experiences; and indirect effect = −0.142, 95% CI = [−0.191, −0.093], SE = 0.025, z = −5.67, p < .01, for negativity of daily experiences. The second mediating factor was satisfaction with public and common goods, indirect effect = 0.025, 95% CI = [0.001, 0.048], SE = 0.012, z = 2.04, p < .05, for global life evaluation; indirect effect = 0.005, 95% CI = [0.001, 0.009], SE = 0.002, z = 2.19, p < .05, for positivity of daily experiences; and indirect effect = −0.027, 95% CI = [−0.053, −0.001], SE = 0.013, z = −2.07, p < .05, for negativity of daily experiences. That is, residents of rich nations evaluate their lives to be closer to the best possible life and have more positive daily experiences and less negative daily experiences than people living in poor nations. This is in part because residents of rich nations earn more money and are in general more satisfied with public and common goods than residents of poor nations are.
Once these mediating processes and individual-level variables were taken into account, the wealth of nations was no longer positively associated with global life evaluation (z = 0.89, n.s.) or positivity of daily experiences (z = −0.55, n.s.). Indeed, everything else being equal, residents of rich nations reported higher levels of negativity of daily experiences than did residents of poorer nations, b = 0.385, 95% CI = [0.227, 0.542], z = 4.79, p < .01 (see Daly, Oswald, Wilson, & Wu, 2011). In contrast, even after controlling for various mediating processes and individual-level variables, we found that progressive taxation remained marginally positively associated with global life evaluation, b = 0.016, 95% CI = [−0.003, 0.036], z = 1.67, p < .10, and significantly positively associated with positivity of daily experiences, b = 0.003, 95% CI = [0.001, 0.004], z = 3.18, p < .01, but it was not associated with negativity of daily experiences, b = −0.010, 95% CI = [−0.024, 0.003], z = −1.50, n.s.
Another important question that the multilevel analyses address is potential variation in Level 1 (within-nations) associations. The relationship between income and global life evaluation was marginally stronger in nations with less-progressive taxation than in nations with more-progressive taxation (mean slope = 0.096, moderation effect = −0.128, z = −1.74, p < .10). Similarly, the inverse relation between income and negativity of daily experiences (mean slope = −0.044) was stronger in nations with less-progressive taxation than in nations with more-progressive taxation (moderation effect = 0.125, z = 2.15, p < .05). In general, then, personal income appears to matter more to the subjective well-being of residents in nations with less-progressive taxation than to the subjective well-being of residents of nations with more-progressive taxation. The link between satisfaction with public and common goods and positivity of daily experiences was stronger in nations with less-progressive taxation than in nations with more-progressive taxation (mean slope = 0.148, moderation effect = −0.276, z = −2.17, p < .05). Finally, the inverse relation between satisfaction with public and common goods and negativity of daily experiences (mean slope = −0.795) was stronger in richer nations than in poorer nations (moderation effect = −0.203, z = −2.92, p < .01).
Discussion
In the study reported here, we found that residents of nations with a more-progressive taxation policy reported higher levels of subjective well-being on average than residents of nations with a less-progressive taxation policy, controlling for various individual-level variables and the wealth of nations. Whereas more-progressive taxation was consistently associated with higher levels of national subjective well-being, the tax rate for average earners and government spending were not. Our findings therefore do not provide support for the simple “big government” idea that the larger a role the government plays, the better the quality of life becomes. Indeed, government spending as a percentage of GDP was associated with lower levels of subjective well-being. 3 Thus, it is the tax policy that aims at a fair redistribution of wealth, not large government spending or a high average tax rate per se, that is associated with higher levels of national subjective well-being.
More important, we found that the relation between progressive taxation and subjective well-being was explained by citizens’ satisfaction with public and common goods, such as public transportation, the education system, and health care. A major contribution of the current research is its demonstration that satisfaction with public and common goods is one reason why residents of nations with more-progressive taxation are in general happier than residents of nations with less-progressive taxation. The findings from our mediation analysis also suggest that a key to a happy society is quality public and common goods. That is, a progressive taxation policy or a nation’s wealth in and of itself might not create a happy society. Rather, it is the use of the nation’s wealth to provide citizens with better public goods that results in increased well-being. In this regard, it is interesting to note that government spending as a percentage of GDP per capita was negatively associated with satisfaction with public and common goods. That is, large government spending per se does not guarantee high-quality public and common goods. This might be a reason why government spending as a percentage of GDP or of the overall tax rate was not associated with increased levels of subjective well-being. The flip side of this finding is that even if a society does not adopt a progressive tax, as long as it can afford good public transportation, a quality educational system, health care, and so forth, citizens are likely to be happy. It is important in future research to identify other social indicators that affect citizens’ satisfaction with public and common goods.
Our interpretation of the present findings is that a fair redistribution of wealth via progressive taxation increases the mean happiness of a nation’s citizens. We have arrived at this interpretation in part because the other causal direction (unhappiness giving rise to the adoption of a flat tax) is possible but unlikely: Conservatives who favor a flat tax tend to be slightly happier than liberals who favor a progressive tax (Napier & Jost, 2008). However, it is important to keep in mind that our interpretation is still vulnerable to a third-variable account. Namely, there might be any number of third variables (e.g., social cohesion) that are associated with both higher levels of subjective well-being and more-progressive taxation. For instance, trust, reciprocity, and social support among citizens are known to be associated with increased levels of subjective well-being (Dunn, Aknin, & Norton, 2008; Konow & Earley, 2008; Oishi & Schimmack, 2010), on the one hand, and could lead to a more-progressive tax, on the other hand. To this end, it is critical in the future to examine whether a change toward more-progressive taxation in nations will result in a higher average level of subjective well-being; however, researchers should also assess other societal, interpersonal, and psychological changes (potential third variables). It is also important to identify the optimal level of taxation for the happiness of a nation’s citizens, as after a certain point, more-progressive taxation might not be associated with higher levels of subjective well-being. 4
Finally, it should be noted that a taxation policy that works in one nation might not work in another nation. In the future discussion of “what is a happy society?” it will be important to identify the moderating role of local history and culture in this regard (cf. Alesina, Di Tella, & MacCulloch, 2004; Berg & Veenhoven, 2010). Taxation is one of the major policies that affect citizens’ everyday lives. We hope that taxation will be examined not only in terms of economic measures (e.g., GDP, unemployment), but also in terms of psychological measures (e.g., happiness) in the future.
Footnotes
Acknowledgements
We thank Xi Wang, Sharon Kim, Felicity Miao, Jamie Schiller, and Casey Eggleston for their assistance, and Carol Graham and John Helliwell for their valuable comments. We also thank the Gallup organization for making the data available to us.
The authors declared that they had no conflicts of interest with respect to their authorship or the publication of this article.