Can Money Buy Happiness? Income and Multidimensional Life Satisfaction in Brunei Darussalam

Abstract

This paper examines the nexus of income and multidimensional life satisfaction (LS) in the oil- and gas-rich Belait district of Brunei Darussalam. Using a random sample of 1,000 Belait residents and principal component factor analysis to sort 11 domains-of-life satisfaction into three uncorrelated LS spheres-LS with materialist life (job, stress, and income); LS with postmaterialist life (family, spirituality, neighbors, and community activities); and LS with public life (health, neighborhood facility, feeling safe at home, and quality of living environment)-we discover the following. First, positive income effects on LS with materialist life sphere are evidenced from lowermiddleto high-income bands. Second, a positive income effect on LS with post-materialist life sphere is only found in the high-income band. Income does not relate to LS with public life sphere. These findings are robust to using each domains-of-life satisfaction and treating scores on domain satisfaction as ordinal and cardinal measures.

Headnote

This paper examines the nexus of income and multidimensional life satisfaction (LS) in the oil- and gas-rich Belait district of Brunei Darussalam. Using a random sample of 1,000 Belait residents and principal component factor analysis to sort 11 domains-of-life satisfaction into three uncorrelated LS spheres-LS with materialist life (job, stress, and income); LS with postmaterialist life (family, spirituality, neighbors, and community activities); and LS with public life (health, neighborhood facility, feeling safe at home, and quality of living environment)-we discover the following. First, positive income effects on LS with materialist life sphere are evidenced from lowermiddleto high-income bands. Second, a positive income effect on LS with post-materialist life sphere is only found in the high-income band. Income does not relate to LS with public life sphere. These findings are robust to using each domains-of-life satisfaction and treating scores on domain satisfaction as ordinal and cardinal measures.

Keywords: Brunei Darussalam, domains-of-life satisfaction, happiness, income, life satisfaction, multidimensional, oil- and gas-dependent economy, Southeast Asian country

JEL codes: D60, I31, O53

(ProQuest: ... denotes formulae omitted.)

I. Introduction

Conventionally, having a high income allows an individual to have higher consumption of material goods and services, yielding higher individual utility or satisfaction (higher indifference curve). The pursuit of happiness, therefore, has been intimately associated with income and material wealth. Moreover, higher income also conveys higher social status (Frey and Stutzer 2002) and may enable other nonmaterial life fulfillment (e.g., the ability to meet religious or spiritual and family obligations). There is a growing corpus of research that is devoted to the study of the nexus between income and subjective well-being both on the macro and micro levels (Frey and Stutzer 2002; Clark, Frijters, and Shields 2008; Frey 2020). A catalyst of this literature is the Easterlin paradox-the puzzle that despite growing real national income over time, especially among developed countries, their average subjective life satisfaction (LS) remains largely unchanged (Easterlin 1974)-suggesting no linkages between economic development and average happiness. This paradox received both empirical support (Easterlin 2001, 2013, 2015) and empirical contradiction (Stevenson and Wolfers 2008, Di Tella and MacCulloch 2008, Veenhoven and Vergunst 2014). Furthermore, country-level studies using microdata usually find a positive impact of individual income on subjective LS (Di Tella, MacCulloch, and Oswald 2003; Clark, Frijters, and Shields 2008; Killingsworth 2021). However, the income-happiness literature has mostly focused on the aggregate measure of LS that summarizes overall self-perceived happiness or satisfaction with life as a whole.1 There is a scarcity of empirical studies that look at this nexus focusing on multidimensional, self-perceived LS or domains-of-life satisfaction (e.g., satisfaction with income, stress, health, family, spirituality or religiosity, neighbors, and community, among other life domains) (Frey 2020). Such inadequate attention to the impacts of income on different domains-of-life satisfaction-capturing aspects of both Hedonic and Eudaimonic LS- deprives us of further detailed insights.2 In this study, we fill this gap by examining the nexus between income and multidimensional domains-of-life satisfaction using a case study of Brunei Darussalam, a small but oil- and gas-rich welfare state.

Brunei Darussalam is part of the dynamic Southeast Asian region. It is the least explored in terms of the aforementioned nexus even though it is a wealthy country with a population of less than half a million (Inoguchi and Fujii 2013; Tsui 2014; Ngoo, Tey, and Tan 2015; Lim et al. 2020; Cuong 2021). Its economy depends heavily on oil and gas, which account for over 91% of the country's total commodity exports, 76% of government revenue, and about 66% of gross domestic product. Most of its oil and gas reserves and production are concentrated in the Belait district, located in the western part of the country, which is the largest district in land size. With wealth from oil and gas, Brunei Darussalam provides its citizens with generous welfare support (e.g., highly subsidized necessity goods and services, free health care, free education, low-cost housing, old-age pension, and no income tax) (Koh 2020). Brunei Darussalam may provide an interesting case study within the context of a resourcedependent Southeast Asian state with an extensive welfare system.3 This is because it may add further insight-at least along with the income-LS nexus-to a recent line of cross-economy studies investigating the influence of natural resource dependence on overall happiness and subjective well-being (Ali, Murshed, and Papyrakis 2020; Vaskovskyi 2020; Mignamissi and Kuete 2021; Fenton Villar 2022). These studies attempt to look beyond the traditional focus on economic growth and to link resource rents with happiness at the cross-economy level.4 Thus, we add to this line of literature a case study investigating the income and multidimensional domains-of-life satisfaction nexus within the context of a small, resource-dependent economy and Southeast Asian welfare state. Using primary data randomly collected at the district level (i.e., the oil- and gas-rich Belait district), we seek to shed nuanced insights on 11 dimensions of domains-of-life satisfaction (and the corresponding newly constructed composite LS that spans across materialist and post-materialist aspects of LS)-the self-perceived satisfaction with income, health, job, stress, partner and family, religiosity or spirituality, neighbors, community engagements, feeling safe at home, neighborhood facilities, and quality of living environment-to provide deeper insights on whether income plays any role in the multidimensional domains-of-life satisfaction in Brunei Darussalam.

Our empirical analysis presents several important results. First, in most cases under consideration, income does not matter for domains-of-life satisfaction at the lower end of income distribution. This may underscore the important role of welfare support for the poor. Second, income is neutral to domains-of-life satisfaction with stress and community activities. Third, income promotes domains-of-life satisfaction with regard to job, income, and quality of the living environment for residents in the lower-middle- to higher-income bands; while positive effects of income on domains-of-life satisfaction with family, spirituality, health, home security, neighbors, and neighborhood facilities are found only at higher-income bands. Fourth, when principal component factor analysis was used on the 11-dimensional domains-of-life satisfaction, they can be sorted into three uncorrelated factors (in order of importance): LS with (i) public life sphere, (ii) post-materialist life sphere, and (iii) materialist life sphere. Further analysis revealed that, in line with findings using individual domains-of-life satisfaction, a positive income effect on LS with post-materialist life sphere is evidenced only at higher-income bands; while a positive income effect on LS with materialist life sphere is evidenced from the lower-middle through higherincome bands. Finally, conforming with our prior expectation for the natural-resourcedependent welfare state of Brunei Darussalam, income does not seem to matter in terms of LS with the public life sphere.

The rest of this paper proceeds as follows. Section II provides a concise review of the literature. Section III explains the data, variables, and methodology. Section IV discusses the empirical findings. Section V provides the concluding remarks.

II. Background Literature

LS, or happiness, is usually measured in a survey by an individual self-reporting his or her feelings of happiness, life satisfaction, or well-being, which are commonly treated as interchangeable (see Veenhoven 2012, Easterlin 2015) to reflect more closely the experience utility-the pleasure derived from consumption (Kahneman and Thaler 1991; Di Tella, MacCulloch, and Oswald 2003).5 LS is commonly taken to mean "an individual's introspective hedonic evaluation of life" (Hirschauer, Lehberger, and Musshoff 2015, 649). Two approaches are commonly used to study subjective wellbeing through LS or happiness: (i) satisfaction with life as a whole and (ii) satisfaction with domains of life (Rojas 2007, 2008; Van Praag and Ferrer-i-Carbonell 2008). The former assesses subjective well-being via overall LS, while the latter views LS as the outcome of satisfaction across various domains of life (e.g., satisfaction with income, job, family, and spirituality, among others). Using data from Mexico, Rojas (2007) showed that socioeconomic and demographic factors have differential influences on satisfaction with health, economic, job, family, friendship, personal, and community domains of life. He found that income only affects LS with respect to the economic and job domains. Similarly, Van Praag and Ferrer-i-Carbonell (2008) used microdata from Germany and the United Kingdom to provide socioeconomic evidence, including income impacts on the multiple domains of satisfaction-e.g., satisfaction with job, finances, health, housing, leisure, social life, and environment. They showed that income does not have a uniform positive or negative influence across these domains.

Furthermore, macro empirical evidence derived from cross-economy data reported a positive correlation between income and happiness (overall LS), but over time such a relationship disappears (Easterlin 2001). Other studies show that the strength of this correlation is not robust across different income groups. For example, Kahneman et al. (2006) argued that income has a transitory (or adaptation) effect on happiness; what matters is relative income. However, Paul and Guilbert (2013) found no empirical support for the adaptation effect in the case of Australia. Other scholars show that moving from lower to higher income can monotonically increase relative happiness. For example, Caporale et al. (2009) used European Social Survey data to empirically show that an increase in income is monotonically associated with an increase in happiness in European countries.6 Luo (2021) also found similar conclusions for Germany.

Although macro literature is less conclusive on the income-LS nexus, studies using cross-sectional survey data of one country or a panel of countries over some period usually conclude with finding a positive income effect on happiness (Clark, Frijters, and Shields 2008). For instance, using a sample of 250,000 randomly sampled respondents in Europe and the United States from 1975 to 1992, Di Tella, MacCulloch, and Oswald (2003) showed that income positively correlates with happiness. In a more recent study that uses a much larger random sample of 1.7 million Americans, Killingsworth (2021) also found similar results and underlines that income is linearly and positively related throughout the entire range of income. Using a German socioeconomic panel, Pouwels, Siegers, and Vlasblom (2008) showed that the positive effects found by past studies were underestimated as they ignored "working hours"- the disutility-associated with earnings.

Furthermore, in the context of welfare states, studies have shown that welfare benefits-measured in different forms such as government welfare expenditure, in-kind and in-cash benefits, and the extensiveness of the benefits-are associated with increased happiness (Pacek and Radcliff 2008; Di Tella and MacCulloch 2008; Easterlin 2013; Flavin, Pacek, and Radcliff 2014; Martela et al. 2020). Pacek and Radcliff (2008) provided evidence that welfare state generosity-the extent of emancipation from market dependency along with pensions, income maintenance for the ill or disabled, and unemployment benefits-can increase LS. This finding is supported by Flavin, Pacek, and Radcliff (2014), who further showed that the positive effects of welfare policy on LS are felt among both low-income and high-income earners. Easterlin (2013) argued that welfare policies-measured by the generosity index reflecting the income replacement rate and the scope and duration of benefits for unemployment, sickness, and pensions-in ultra-welfare European Union states (e.g., Denmark, Finland, and Sweden) are uniformly more generous than semi-welfare European Union states (e.g., Austria, France, Germany, and the United Kingdom). He shows that the respondents in the ultra-welfare states are more satisfied with their work, health, family life, and overall LS than in the semi-welfare states. Thus, the study argues that full employment and a generous and comprehensive social safety net increase happiness. The conclusions from these relevant studies suggest that welfare state policies are conducive to human happiness.

While there is a relatively large collection of literature on the income-happiness nexus in advanced economies, such research on Asian and Southeast Asian societies is only emerging (Oshio, Nozaki, and Kobayashi 2011; Inoguchi and Fujii 2013; Tsui 2014; Ngoo, Tey, and Tan 2015; Lim et al. 2020; Cuong 2021). The balance of evidence on the nexus in Asia so far is that money buys different levels of happiness. In the context of the People's Republic of China, Japan, and the Republic of Korea, Oshio, Nozaki, and Kobayashi (2011) found that relative income and individual income matter for perceived happiness in these East Asian economies. Likewise, for Taipei,China, Tsui (2014) found positive income effects on happiness across all measures of income: absolute, relative, and expected income. Using data from the Asian Barometer survey, Ngoo, Tey, and Tan (2015) found similar results across 28 economies from different Asian subregions even when they additionally control for the measure of the standard of living. A more recent study by Lim et al. (2020) used World Value Survey data on East Asian and Southeast Asian economies to show that (perceived relative) income promotes LS. More importantly, they show that societal values (secular values along with defiance, relativism, and skepticism) play an important moderating role on the nexus via substitution effects between negative societal value-LS and positive incomeLS nexuses. Furthermore, looking at the income-happiness nexus among older people in Viet Nam, Cuong (2021) also showed that money does buy happiness.

So far, Brunei Darussalam has been absent in all these studies. The sole exception is Inoguchi and Fujii (2013), who used Asia Barometer survey data, which reflect the daily life views of Central, East, South, and Southeast Asians (including Bruneians) to study socioeconomic and demographic correlates of subjective well-being-measured along with happiness, enjoyment, and achievement. With respect to the income-LS nexus, their findings are mixed across many Asian economies. For example, the nexus with happiness is absent (i.e., not statistically significant) for Hong Kong, China; the People's Republic of China; the Republic of Korea; and Taipei,China; as well as for many other Asian economies. The nexus for Japan was negative and statistically significant. This is in contradiction with the study by Ngoo, Tey, and Tan (2015) that used the same source of data. Of particular interest is the case of Brunei Darussalam, where Inoguchi and Fujii's (2013) findings showed a negative and significant effect of income on happiness. Thus, even within this scant empirical literature on the incomehappiness nexus in Asia, inconclusive evidence exists. In the context of Brunei Darussalam, further study is warranted at the district and national levels. This study contributes to the emerging literature by providing a first look at Brunei Darussalam's data at the district level.

We approach this nexus from the perspective of a natural-resource-abundant (or an oil- and gas-exporting) small state with a generous welfare support system. This is in the interest of the recent largely separated line of empirical literature that investigates the connection between resource rents and LS or happiness. Countries rich with natural resources have access to the resources needed to satisfy their people's aspirations and material consumption, among other factors. This suggests that rents from these resources should be associated with higher levels of LS. However, a recent debate emerges, with a few studies (see, for example, Ali, Murshed, and Papyrakis 2020; Vaskovskyi 2020; Mignamissi and Kuete 2021) showing that natural resource rents reduce LS and arguing that the so-called resource curse-the tendency of resourcerich countries to have dismal economic performances-also extends to perceived LS. However, Fenton Villar (2022) found no such evidence for Latin American countries. The question then is, would Brunei Darussalam be subject to this pattern of finding? In this study, we do not look directly at the resource rents but take Brunei Darussalam's background as an oil- and gas-dependent economy with generous welfare support as given. We then investigate the nexus of income and multidimensional LS in the Belait district where oil and gas reserves and production are concentrated. We argue that this provides one way to link the two lines of literatures in our attempt to contribute to it with respect to the income-happiness nexus in a resource-abundant country.

III.Data and Methodology

A. Data and Variables

This study utilizes primary data from a demographic and socioeconomic survey on the oil- and gas-rich Belait district of Brunei Darussalam. The survey was conducted by the Centre for Advanced Research (CARe) of the Universiti Brunei Darussalam in October-November 2019. The survey collected baseline data of 1,000 respondents from the district's population aged 18 years and above on their demographic and socioeconomic characteristics. The sample size was randomly drawn with a 95% confidence level and a ±3% margin of error.7

The 11 dimensions of perceived domains-of-life satisfaction involve questions about satisfaction with income, job, health, partner and family, level of stress, interaction with neighbors, religion and spirituality, community activity and engagement, safety at home, neighborhood facilities, and quality of the living environment. The answer to these questions is on a five-point scale as follows: (1) unhappy, (2) less happy, (3) average, (4) happy, and (5) very happy.8 The Cronbach's Alpha-the reliability test that measures the consistency between the questions to capture the intercorrelation of the same concept or construct (Tavakol and Dennick 2011)-for satisfaction with these 11 domains of life is 0.816. This is well within the acceptable intervals of between 0.75 and 0.95 (Ngoo, Tey, and Tan 2015), indicating a good degree of reliability in these measures of perceived LS. We also conduct an exploratory factor analysis to check whether these measures can be further grouped into different LS spheres in our next analysis. The income is measured by the monthly earning levels of the respondents, which is coded from 1 to 13 as follows:9

1. No monthly earnings

2. Monthly earnings below B$250

3. Monthly earnings B$250-B$499

4. Monthly earnings B$500-B$999

5. Monthly earnings B$1,000-B$1,999

6. Monthly earnings B$2,000-B$2,999

7. Monthly earnings B$3,000-B$3,999

8. Monthly earnings B$4,000-B$4,999

9. Monthly earnings B$5,000-B$5,999

10. Monthly earnings B$6,000-B$6,999

11. Monthly earnings B$7,000-B$7,999

12. Monthly earnings B$8,000-B$8,999

13. Monthly earnings B$9,000 and above

Each monthly earnings band is captured by an indicator (or dummy) variable that takes a value of 1 if a respondent's income falls within it and zero otherwise. There are 12 indicator variables included in the following baseline regression models with "no monthly earnings" being a reference group. In the LS equation (1) below, the coefficients on these included indicator variables capture the effects of income on multidimensional domains-of-life satisfaction in the district. Besides these focal income variables, in the empirical analysis we control for the following standard demographic and socioeconomic variables based on (Oshio, Nozaki, and Kobayashi 2011; Inoguchi and Fujii 2013; Tsui 2014; Ngoo, Tey, and Tan 2015; Lim et al. 2020) gender, marital status, educational attainment, age, and occupations. We also consider additional socioeconomic characteristics such as homeownership; self-rating of the adequacy of monthly earnings to support oneself and the family; and monthly consumption expenditure on food, clothing, and personal care items. In including these controls, we follow the strategy recently suggested by Bartram (2021) and others discussed in the following section. Since there are missing observations for some of the included variables, the final sample sizes are between 679 and 736 observations that are used across different models.

Table 1 shows the summary statistics. The largest proportion of respondents in any income band is 22.5% for B$1,000-B$1,999, while 11.6% have no monthly earnings. All in all, the means (or medians) of different aspects of domains-of-life satisfaction are above 3, with satisfaction with income having the lowest mean and satisfaction with spirituality and religion having the highest mean. Furthermore, these 11 dimensions may potentially measure some overlapping LS (e.g., domains-of-life satisfaction with job and income); we, therefore, conduct a further check on whether these aspects of satisfaction can be grouped into different spheres or domains. We do so by using exploratory factor analysis to sort them into uncorrelated common factors (or domains). This allows for further analysis of the effects of income bands on satisfaction across life spheres (Inoguchi and Fujii 2013).

The principal component factor analysis is conducted on the 11 aspects of satisfaction to summarize their correlated structure into fewer common factors. These factors are extracted using principal component analysis (PCA), based on Kaiser criteria to retain the factors (i.e., factor is retained when its eigenvalue is more than 1), and then employing varimax orthogonal rotation to distribute factor loading evenly among the retaining factors (see Slesman, Baharumshah, and Ra'ees 2015). Table 2 reports the results. It shows that these 11 aspects of satisfaction can be sorted into three different uncorrelated domains, with the first factor explaining 36.3% of the total variance, while the second and third factors account for 11.2% and 9.6% of the variance, respectively. In total, the three domains explain about 57% of the total variance underlying these LS data. The first factor (F1) loads predominantly on satisfaction with health, neighborhood facilities, quality of the living environment, and feeling safe at home. The second factor (F2) loads significantly on religion and spirituality, family, neighbors, and involvement with community activities. The third factor (F3) loads dominantly on LS with income, job, and the level of stress.

F1 reflects the "LS with public life sphere" domain that captures public infrastructure provisions. It is in line with Inoguchi and Fujii's (2013) so-called "public life sphere" or "quality of life (QOL) enabling factor." F2 captures most of the underlying variables that relate more closely to the so-called "LS with postmaterialist life sphere" that associates with self-expression concerning nonmaterialist life fulfillments. F3 relates closely to "LS with materialist life sphere" that centers on self-expression concerning materialist life fulfillment (Inglehart 1971, Inoguchi and Fujii 2013). We conjecture that these findings mimic quite closely the welfare state nature and close-family culture of Brunei Darussalam's society. This finding is in the spirit of Inoguchi and Fujii (2013, 67), who found three life spheres using the Brunei Darussalam sample-the "public life sphere," which ranked first, followed by the "post-materialist" and "materialist" spheres of life-from 16 aspects of LS.10 In our empirical analysis, we also employ each of these LS spheres along with its constituent components to gauge the comprehensive (summary) effects of income on LS with materialist life sphere, post-materialist life sphere, and public life sphere.

B. Methodology

The following standard multivariate regression model is considered where multidimensional satisfaction across domains of life is expressed as a function of income and the relevant set of personal characteristics (i.e., the control variables that are identified based on the nonparametric approach of causal diagrams-also known as Direct Acyclic Graphs [DAGs]) (Elwert and Winship 2014, Bartram 2021):

...

(1) Subscript i indexes individual respondents. The dependent variable is subjective life satisfaction LSf for individual i in domain k, where k indexes the 11 domains-oflife satisfaction and the three newly constructed LS: F1, F2, and F3. The focal variable is INCy, which is the indicator variable for each level of income, where j indexes the included 12 levels of income as coded above, with j = 1 representing the omitted reference income level (no earnings group), j = 2 gives INCi2, which is the indicator variable for monthly earnings below B$250, and so forth. Any effect of INC^· for j e[2,13] is compared with the reference income band.11 The following regression function spells out the effects for each income band in relation to the reference income at any domain-of-life satisfaction

k:

... 11 In our robustness check, we merge those having monthly earnings of B$6,000-B$6,999 and above together, as their respective shares in the sample are too small to see their impacts, if any, on different domains of satisfaction (see Table

Xi is the vector of control variables to be identified-both indicator and continuous variables-as detailed in Table 1, and e? is the normally distributed error term. The coefficients ß capture the effect of each included level of income, while 0'k is a vector of parameters measuring the effect of each identified and included control variable on LSf. To include the set of control variables, we follow a recent DAG strategy advocated by Bartram (2021) for the study of happiness to minimize the estimation biases in our attempt to gauge the effects of income on LS in Brunei Darussalam. This DAG approach suggests we include only confounding variables and exclude intervening variables. Figure 1 details the DAG paths that identify the status of each potential control variable. The confounding variable is a variable influencing both income and LS (i.e., arrows pointing toward both), while the intervening variable is a variable acting as a channel through which income affects LS (i.e., the arrow from income passes through it before reaching LS). Thus, we only include covariates in X listed in Table 1 that are confounding variables.

Figure 1 shows, first and foremost, that age and gender are clearly confounding variables-as income does not determine age and gender. Second, occupation status- employed, unemployed, retired, and other classifications-and education level determine whether a person earns a higher or lower income and most likely influence his or her level of LS. Third, marital status is, however, less clear in this context of whether it is a confounding or intervening variable. Intuitively, the direction of the arrows between income and marital status can go both ways: Sufficient income would induce an individual to get married, while it is also plausible that marital responsibilities may induce individuals to work harder to earn more income to support their family. Empirically, marital status is usually found to be statistically significant in the earning equations (Schoeni 1995), while income was also found to influence marital status (Watson and McLanahan 2011). Furthermore, marital status was also found to influence happiness (Stack and Eshleman 1998). With this uncertainty, we treat it as a confounding variable and include it in X. In our sensitivity analysis, we exclude it (thus treating marital status as an intervening variable) to see whether the main findings on the effects of income on LS remain robust. Finally, consumption, owning a home, and (self-rated) having sufficient monthly earnings to support oneself and family may determine LS. However, these variables are certainly determined by income, making them intervening variables. We, therefore, exclude these variables from the Xt controls. Our final set of controls consists of gender, marital status, education, occupation, age, and age squared.

It should be noted that, within the context of DAG, the coefficient interpretations for income variables differ from that of the control variables. While coefficients (ßj) on each income level represents the "total effect" of that income level (in relation to reference income) on domain-of-life satisfaction at any level of control variables, the coefficient on each of the included control variables captures the "direct effect" of that control relative to income-it is the portion of the effect of that control (e.g., education) on LSf that is not mediated through its influence on income (Westreich and Greenland 2013).12 Thus, the interpretations of the (statistically significant) coefficients on the included controls below reflect this subtle distinction.

Equation (1) is estimated using both the ordinary least squares (OLS) and the ordered probit regression (OPR) estimators. When OLS is used, the assumption is that the LS numerical response scores are cardinal, while the OPR assumes the LS numerical scores are ordinal.13 Past studies have shown that assuming ordinal LS scores (hence fitting the model with OPR) or cardinal LS scores (hence fitting the model with OLS estimation) make little difference on the estimation results-the sign, statistical significance, and magnitudes of the coefficients are qualitatively the same (Dunn 1993, Ferrer-i-Carbonell and Frijters 2004). There is, however, a trade-off for going from the ordinal numerical response (with less measurement error but also fewer information contents) to cardinal (or continuous) numerical response (with increased measurement error but more information contents). Dunn (1993) used both Monte Carlo and actual survey data to show that treating survey response as a continuous response (i.e., cardinality) outperformed the one that treated it as categorical (i.e., ordinality) in the estimation. This is also supported by a more recent study conducted by Ferrer-i-Carbonell and Frijters (2004). Taking this trade-off into account and these recent findings, we use the OLS estimation of equation (1) as our baseline estimation while complementing it, as part of our robustness analysis, with the OPR analysis. All models are estimated with robust standard errors-i.e., using the corrected variancecovariance matrix-to correct for any autocorrelation and heteroscedasticity in the error residuals that may be prevalent in cross-sectional data.14

IV. Empirical Results and Discussions

Tables 3A-3C summarize the OLS and OPR main results on the effects of income on multidimensional domains-of-life satisfaction in Belait district. Table 3A summarizes the findings on domain-of-life satisfaction with job, stress, and income, and the PCA-generated LS with materialist life sphere (or F3) from these three dimensions. Table 3B shows the results for domain-of-life satisfaction with life partner and family, religion and spirituality, neighbors, and involvement with community activities, and the PCA-generated LS with post-materialist life sphere (or F2). Table 3C reports the findings on domain-of-life satisfaction with health, feeling safe at home, neighborhood facilities, and quality of the living environment, and the PCA-generated LS with public life sphere (or F1). All the models in Tables 3A-3C are well specified, as they conform with necessary regression assumptions (e.g., correct model specification) for the inference.15

First and foremost, the findings on the control variables capturing demographic and socioeconomic variables are generally robust across OLS and OPR models and are largely in line with past studies on Asian societies (Oshio, Nozaki, and Kobayashi 2011; Inoguchi and Fujii 2013; Ngoo, Tey, and Tan 2015; Lim et al. 2020). The findings reported in Tables 3A-3C show that the controls that characterize demographic factors-gender, marital status, occupation status, age, and age squared-do not have the same direct effects (and not all of them have statistically significant effects) on each domain-of-life satisfaction of the Belait residents.

Specifically, the results show, with respect to the statistically significant coefficients, that female respondents tend to report themselves as more satisfied with their income (Models 3 a and 3b, Table 3A), but less happy with their health (Models 8a and 8b, Table 3C), than their male counterparts. Unmarried respondents are more likely to state that they not only feel more stressed (Models 2a and 2b, Table 3A)- but also are less satisfied with family (Models 4a and 4b, Table 3B), religion and spirituality (Models 5a and 5b, Table 3B), their interaction with neighbors (Models 6a and 6b, Table 3B), community activities (Models 7a and 7b, Table 3B), and their health (Models 8a and 8b, Table 3C)-compared to married Belait residents. Similarly, those who are divorced, widowed, or separated are also more likely to report themselves being less happy with family (Models 4a and 4b, Table 3B). However, those with a college or higher education are more likely to state that they are satisfied with their family (Models 4a and 4b, Table 3B), but less so with neighbors (Models 10a and 10b, Table 3C), compared to those without formal education.

Furthermore, occupation status is only statistically significantly correlated to satisfaction with job, family, involvement in community activities, health, feeling safe at home, and quality of the living environment. The findings suggest that unemployed respondents are relatively less satisfied with their current jobless status (Models 1a and 1b, Table 3A) and family (Models 4a and 4b, Table 3B); while self-employed respondents are relatively happy with their job (Models 1a and 1b, Table 3A), respondents having nonregular employment (e.g., employed part-time, freelance workers, and those under the i-Ready program-a government-sponsored apprenticeship program for fresh graduates) and retirees are happier with community activities (Models 7a and 7b, Table 3B) than those having a regular job. It is interesting to observe that out of all occupation statuses, self-employed residents are the only group that reported being relatively happier with their job than the regular job holders, which is in line with the findings of Benz and Frey (2008). This suggests that promoting entrepreneurship enhances satisfaction across these domains of life in the Belait district. Finally, the statistical significance of the positive-signed coefficient on age and the negative-signed coefficient on age squared suggest that although respondents report increasing domains-of-life satisfaction as they become older- especially with job, income, neighbors, community activities, facilities, and quality of living environment-this feeling is reversed for older residents beyond a certain age, which is 73 years old based on Model 1a.

Turning to our focal indicator variables capturing different income bands, we observe from Tables 3A-3C that there are heterogeneous effects of income on different domains-of-life satisfaction in the Belait district. First, Table 3A shows that those earning higher incomes starting from the B$1,000-B$2,999 band are happier with their job (Models 1a and 1b), but only those in the income band of B$3,000B$3,999 or above are likely to report more satisfaction with their income (Models 3a and 3b). Furthermore, residents in the lower-income band of B$250-B$499 are less likely to state that they are satisfied with their income. The finding that belonging to higher-income bands is positively correlated to satisfaction with income and job implies that satisfaction does not adapt to income in the Belait district. This is in line with the recent studies that investigate other contexts (see, for example, Paul and Guilbert 2013 and Luo 2021 for studies on Australia and Germany, respectively). Stress seems to have no statistically significant effect, suggesting that a stressful life is not tied to income among Belait residents. Furthermore, using an aggregate measure of LS with materialist life sphere, Model 3 a seems to summarize and confirm the nonadaptation effects of income. In short, low-income earners are either dissatisfied with or are less likely to view their income as making any difference to domain-oflife satisfaction with income, and similarly with job and materialist life sphere, while high-income earners are more likely to view their income as improving these domainsof-life satisfaction. Thus, only high-income earners are more likely to report a relative increase in satisfaction with their income, job, and aggregate materialist life sphere. One possible postulation may be that, in the welfare state of Brunei Darussalam, the state's guarantee of basic needs for survival (i.e., welfare support) for low-income earners may relieve some of the pressures that income and job would otherwise have to provide for their well-being.16 This may play into the observed absence of the connection between income and domain satisfaction with job and materialist sphere at lower-income bands. Thus, in this context, belonging to the lower end of the income distribution may not necessarily be associated with unhappiness.

Table 3B shows that Belait residents in the B$6,000-B$7,999 income band tend to report themselves as having more satisfaction with their life partner and family (Models 4a and 4b), while those earning B$6,000-B$6,999 are happier with their spirituality (Models 5a and 5b). Though higher income seems to contribute to improved LS along these two post-materialist life-fulfilling dimensions, it does not increase satisfaction to interactions with neighbors, especially for those in the B$8,000-B$8,999 band (Models 6a and 6b), compared to residents with no formal earnings. Income is not linked to satisfaction with community activities (Models 7a and 7b), as they are not statistically significant at any conventional level. Thus, income (and its conveying social status) does not relate to enjoying community activities and engagement. Further, Model 3b reports the overall effects of income bands on the LS within the post-materialist life sphere. It confirms that the overall effects seem positive at the higher-income band of B$6,000-B$6,999. This finding implies that income at higher levels does contribute to nonmaterialist life fulfillment with family and spirituality among Belait residents.17 Thus, the statistically significant connections at the higherincome bands for materialist and post-materialist life spheres may be in accordance with earlier evidence found across Asia (Ngoo, Tey, and Tan 2015). Nevertheless, all models for domains-of-life satisfaction reported in Table 3B have relatively lower ^-squared values, which is in line with past studies using microdata (see, for example, Di Tella, MacCulloch, and Oswald 2003; Rojas 2007, 2008; Van Praag and Ferreri-Carbonell 2008; Inoguchi and Fujii 2013; Killingsworth 2021), suggesting a weak partial explanation of income on the variations in domain satisfaction.

Lastly, Table 3C reports that residents within income bands of B$7,000-B$7,999 and B$6,000-B$6,999 are more likely to state satisfaction with their health (Models 8a and 8b) and feeling safe at home (Models 9a and 9b), respectively, compared to residents in the base band. This may reflect the fact that, especially concerning satisfaction with health domain, Brunei Darussalam provides free health-care services making the impact of income (especially at lower bands) insignificant in influencing LS.18 We conjecture that this may be the case as the state-guaranteed health-care fulfillment for all citizens may allow Belait residents to allocate their income toward other unfulfilled priorities. However, higher income (e.g., the B$7,000-B$7,999 income band) provides flexibility in meeting health-care needs such as opting for relatively more customized private or overseas health care, thus improving satisfaction with health. Concerning satisfaction with neighborhood facilities, Models 10a and 10b show that residents with the lowest income band (less than B$250) report more satisfaction, while those earning B$8,000-B$8,999 are associated with less satisfaction. Furthermore, residents in the B$500-B$999 income band are the only residents to report less satisfaction with their quality of living environment compared to the base income band. Although these findings may reflect the nuanced difference that life fulfillment conveys on lowincome earners versus high-income earners among Belait residents, Model 3c reports that income seems to not matter in influencing LS with public life sphere as none of the results for any income band are statistically significant. This again, at the aggregate level, may reflect the welfare-state nature of Brunei Darussalam's society, whereby life fulfillment with health, security, facilities, and living environment is heavily subsidized by the government.19 As noted earlier in Table 3B that it is common to obtain low Ä-squared and pseudo Ä-squared, the models reported in Table 3C also reported similar cases. Thus, income may weakly and partially explain these domainsof-life satisfaction.

Nevertheless, our findings are largely in line with past research looking at the whole of Brunei Darussalam and Asia (Inoguchi and Fujii 2013; Ngoo, Tey, and Tan 2015) and the European context (e.g., Caporale et al. 2009; Drichoutis, Nayga Jr., and Lazaridis 2010). Ngoo, Tey, and Tan (2015), for example, showed that moving up the income ladder (using a composite measure of living standards) is associated with higher LS in Central and West Asia, East Asia, South Asia, and Southeast Asia. Beyond confirming past evidence, our study sheds further nuanced insight, in the context of the oil- and gas-rich Belait district of a small welfare state, that income bands have differential effects on multidimensional LS. We show that positive income effects on LS with materialist life sphere (job and income) appear to take place only starting from lower-middle-income bands upward, while the positive income effect on LS with post-materialist life sphere (family and spirituality) is only evident at higherincome bands. Conforming with our prior expectation about the resource-dependent, welfare-state context, money seems neutral to LS with the public life sphere among Belait residents. Overall, although most microstudies in the literature found a positive relationship between income and overall subjective well-being (Clark, Frijters, and Shields 2008), our findings demonstrate that, in the context of the oil- and gas-rich Belait district, this happens only at higher-income bands and with specific domainsof-life satisfaction. Thus, not all income bands influence different domains-of-life satisfaction the same way.

Finally, as part of our robustness analysis on the control variable of marital status, we exclude it from equation (1)-thus treating marital status as an intervening variable (Figure 1)-to see whether the finding holds up. The results show that the main findings as reported in Tables 3A-3C remain intact. Thus, the findings are robust to the exclusion of marital status from the DAG's identified set of controls. Furthermore, as the shares of high-income earners in the sample are small (Table 1), we merge the monthly earning bands of B$6,000-B$6,999 and B$7,000-B$7,999 into one category and the remaining high-income bands of B$8,000-B$8,999 and B$9,000 and above into another category (and alternatively merge monthly earnings of B$6,000-B$6,999 and all higher categories into a single category) to further conduct robustness checks on our main findings.20 The results are largely in line with our main findings reported in Tables 3A-3C. Further, the findings are robust across OLS and OPR estimations; thus, they are robust to treating LS numerical scores as both ordinal and cardinal 21 measures.

V. Conclusion

The question of whether more income leads to more happiness has received great interest from scholars across disciplines in the social sciences. Although there is a large corpus of literature on happiness focusing on advanced economies in North America and Europe, there is a dearth of such research on Southeast Asian societies. Brunei Darussalam is not an exception to this observation. Its oil- and gas-rich Belait district presents an interesting case to the nexus between income and multidimensional happiness, especially within the context of a resource-dependent economy. Furthermore, the bulk of income-LS literature has focused on overall LS and paid less attention to the multidimensional domains-of-life satisfaction. Using the recent demographic and socioeconomic survey data from the Belait district, this study examines the impact of income on self-rated multidimensional domains-of-life satisfaction-with job, stress, income, life partner and family, neighbors, spirituality and religiosity, community engagement and activities, health, feeling safe at home, neighborhood facilities, and quality of living environment.

Based on the DAG empirical strategy and the OLS and OPR estimations, the findings generally show that income does not have the same effect across different domains-of-life satisfaction among the Belait residents in Brunei Darussalam. Specifically, income does not relate with most domains-of-life satisfaction at the lower ends of the income distribution. Second, income is neutral to domains-of-life satisfaction with stress levels and community activities. Third, income positively influences satisfaction with job, income, and quality of living environment only starting from the lower-middle-income bands; while the positive effects of income on satisfaction with family, religiosity or spirituality, health, feeling safe at home, neighbors, and neighborhood facilities are only evidenced at the higher-income bands.

Using PCA-generated LS, we further confirm the findings using individual domains-of-life satisfaction. The findings show that the positive income effects on LS with post-materialist life sphere appear to be evidenced only at higher-income bands, while those of LS with materialist life sphere are present only starting from middle-income bands. income may not only mean having access to a more comfortable material life but also higher social status, as well as allowing individuals to satisfy their post-materialist life such as spirituality and religious obligations. Furthermore, conforming with our prior expectation for the natural-resource-rich welfare state of Brunei Darussalam, we find that income does not seem to matter with LS in the public life sphere. Our findings provide a nuanced glimpse of the role income plays in influencing different dimensions of LS in Brunei Darussalam, which can be relevant to other welfare states and resource-abundant countries.

Footnote

1 This overall happiness or LS is usually measured with survey questions such as the following: "All things considered, how satisfied are you with your life as a whole these days?" (Used in World Value Survey); "Taking all things together, how would you say things are these days-would you say you're very happy, fairly happy, or not too happy these days?" or "On the whole, are you very satisfied, fairly satisfied, not very satisfied, or not at all satisfied with the life you lead?" (see Di Tella, MacCulloch, and Oswald 2003, 810-11).

2 According to Ryan and Deci (2001, 144), Hedonic well-being focuses on LS or happiness in the sense of maximizing pleasure and minimizing pain (i.e., "the experience of pleasure versus displeasure broadly construed to include all judgments about the good/bad elements of life"), while Eudaimonic wellbeing or LS focuses on meaning, purpose, self-realization, or self-actualization. It is defined as the extent to which a person is fully functioning. In other words, Hedonic LS captures "short-term happiness," while Eudaimonic LS reflects "long-term happiness" or deeply felt happiness (Frey 2020).

3 Extensive welfare policies (e.g., as is the case among European Union countries) have been empirically shown to be conducive to higher LS (Pacek and Radcliff 2008; Flavin, Pacek, and Radcliff 2014).

4 The evidence so far is mixed.

5 We use the terms happiness, LS, and well-being interchangeably in this study.

6 However, Drichoutis, Nayga Jr., and Lazaridis (2010) showed that the finding on reference income in Caporale et al. (2009) was not robust.

7 A detailed explanation of the random sampling design is available upon request.

8 With respect to domains of stress, interaction with neighbors, feeling safe at home, and spirituality, respondents in the survey were asked to rank on a scale of 1 (worse), 2 (bad), 3 (average), 4 (good), and 5 (best). For example, for satisfaction with the domain of stress, the respondents were asked to assign a rank of 1 (very stressful), 2 (stressful), 3 (average), 4 (less stressful), or 5 (not stressful); they were asked the same for interacting with neighbors, feeling safe at home, and spirituality.

9 Ideally, income derived from other sources, including government transfers, would also be included. However, these data are unavailable in the survey.

10 When using the Asian sample, the study found a "materialist sphere," which ranked first, followed by a "post-materialist sphere" and a "public sphere of life" from 16 aspects of domains-of-life satisfaction. Based on the factor loading among the items, Inoguchi and Fujii (2013) also termed the public life sphere as a "QOL-enabling factor," the post-materialist life sphere as a "QOL-enriching factor," and the materialist life sphere as a "QOL-sustaining factor." It is important to note that, unlike our study, Inoguchi and Fujii (2013) did not consider these life spheres as the outcome variables but examined their effects on the overall level of happiness in Asian societies.

12 In short, this "direct effect" is the portion of the effect of each included controls (e.g., education) embodied in the arrows pointing from that control variable toward LS in Figure 1.

13 Ordinal LS scores assume the ranked orders of different degrees of LS (e.g., from unhappy to happier, or from 1 to 5), but not the magnitude of differences between any two-point score in the scale. While the cardinal LS scores further allow for that magnitude differences to be meaningful and that the scale of that difference is the same (OECD 2013). For example, for ordinal data, a score of 5 is higher than 4, so is 3 compared to 2. However, ordinal scores say nothing about whether such magnitude differences (e.g., between 5 and 4 as well as between 3 and 2) are meaningful and represent the same-scale step. This represents the loss of information content. Cardinal LS score assumes both-the rank order and magnitude differences are of the same-scale step. Thus, moving from ordinal to cardinal assumption of the survey responses represents an increase in the information content of the LS data. However, it is at the same time known that such an increase in information content comes with an increase in measurement error (Dunn 1993).

14 As our sample size is reasonably large enough, we invoke the central limit theorem and make use of asymptotic normal distribution for the residuals. Furthermore, the robust standard error used also indirectly normalizes the residual distributions.

15 In our robust analysis, we also relax the assumed asymptotic normality assumption and estimate all models in Tables 3A-3C using bootstrapped standard errors (and bootstrapped p-values) and our findings remain intact; hence, we confirm our assumed asymptotic normal distribution for the residuals. Due to space constraints, the results from these exercises are not reported here but are available upon request.

16 In the context of Europe's high-income tax regime and welfare setting, Di Tella, MacCulloch, and Oswald (2003) showed that more unemployment benefits are associated with higher national happiness.

17 Brunei Darussalam is predominantly a Muslim country. As part of Islamic religious obligation, Muslims must perform a Hajj-an Islamic pilgrimage to the city of Mecca in the Kingdom of Saudi Arabia-at least once in their lifetime if they can manage to do so, especially financially. Having performed a Hajj, a person is called Haji (for male) or Hajah (for female), which carries relatively higher social status in Brunei Darussalam society.

18 It may also be that satisfaction with health is perceived from health conditions the respondents are currently facing and, hence, is uncorrelated with income. This is a conjecture that requires empirical verification.

19 Utility is linked to consumption, which constitutes one portion of income. Thus, satisfaction with the provision of many public goods-e.g., security, facilities, living environment, and (to a lesser extent) health that partially reflects free health-care provision-can directly be linked to the share of income spent on them. In the welfare setting of Brunei Darussalam, where there is no income tax, such fulfillment is provided by the state. Though this requires empirical verification, we postulate that such a setting may free the share of an individual's income needed to spend on these things, thereby resulting in income playing an insignificant role in residents' perceived satisfaction with health, security, facilities, and living environment.

20 We thank the two anonymous reviewers of this journal for directing us to this concern.

21 The above results are not reported here but are available upon request.

References

References

Ali, Sabna, Syed Mansoob Murshed, and Elissaios Papyrakis. 2020. "Happiness and the Resource Curse." Journal of Happiness Studies 21 (2): 437-64. Bartram,

David. 2021. "Cross-Sectional Model-Building for Research on Subjective WellBeing: Gaining Clarity on Control Variables." Social Indicators Research 155 (2): 725-43.

Benz, Matthias, and Bruno S. Frey. 2008. "Being Independent Is a Great Thing: Subjective Evaluations of Self-Employment and Hierarchy." Economica 75 (298): 362-83.

Caporale, Guglielmo Maria, Yannis Georgellis, Nicholas Tsitsianis, and Ya Ping Yin. 2009. "Income and Happiness across Europe: Do Reference Values Matter?" Journal of Economic Psychology 30 (1): 42-51.

Clark, Andrew E., Paul Frijters, and Michael A. Shields. 2008. "Relative Income, Happiness, and Utility: An Explanation for the Easterlin Paradox and Other Puzzles." Journal of Economic Literature 46 (1): 95-144.

Cuong, Nguyen Viet. 2021. "Does Money Bring Happiness? Evidence from an Income Shock for Older People." Finance Research Letters 39: 01605.

Di Telia, Rafael, and Robert J. MacCulloch. 2008. "Gross National Happiness as an Answer to the Easterlin Paradox?" Journal of Development Economics 86 (1): 22-42.

Di Tella, Rafael, Robert J. MacCulloch, and Andrew J. Oswald. 2003. "The Macroeconomics of Happiness." Review of Economics and Statistics 85 (4): 809-27.

Drichoutis, Andreas C., Rodolfo M. Nayga Jr., and Panagiotis Lazaridis. 2010. "Do Reference Values Matter? Some Notes and Extensions on 'Income and Happiness across Europe'." Journal of Economic Psychology 31 (4): 479-86.

Dunn, Lucia F. 1993. "Category versus Continuous Survey Responses in Economic Modelling: Monte Carlo and Empirical Evidence." Review of Economics and Statistics 75 (1): 188-93.

Easterlin, Richard A. 1974. "Does Economic Growth Improve the Human Lot? Some Empirical Evidence." In Nations and Households in Economic Growth, 89-125. Cambridge, US: Academic Press.

_. 2001. "Income and Happiness: Towards a Unified Theory." Economic Journal 111 (473): 465-84.

_. 2013. "Happiness, Growth, and Public Policy." Economic Inquiry 51 (1): 1-15.

_. 2015. "Happiness and Economic Growth: The Evidence." In Global Handbook of Quality of Life, 283-99. Dordrecht: Springer.

Elwert, Felix, and Christopher Winship. 2014. "Endogenous Selection Bias: The Problem of Conditioning on a Collider Variable." Annual Review of Sociology 40: 31-53.

Fenton Villar, Paul. 2022. "Is There a Mineral-Induced 'Economic Euphoria'? Evidence from Latin America." Journal of Happiness Studies 23 (4): 1403-30.

Ferrer-i-Carbonell, Ada, and Paul Frijters. 2004. "How Important Is Methodology for the Estimates of the Determinants of Happiness?" Economic Journal 114 (497): 641-59.

Flavin, Patrick, Alexander C. Pacek, and Benjamin Radcliff. 2014. "Assessing the Impact of the Size and Scope of Government on Human Well-Being." Social Forces 92 (4): 1241-58.

Frey, Bruno S. 2020. "What Are the Opportunities for Future Happiness Research?" International Review of Economics 67 (1): 5-12.

Frey, Bruno S., and Alois Stutzer. 2002. "What Can Economists Learn from Happiness Research?" Journal of Economic Literature 40 (2): 402-35.

Hirschauer, Norbert, Mira Lehberger, and Oliver Musshoff. 2015. "Happiness and Utility in Economic Thought-Or: What Can We Learn from Happiness Research for Public Policy Analysis and Public Policy Making?" Social Indicators Research 121 (3): 647-74.

Inglehart, Ronald. 1971. "The Silent Revolution in Europe: Intergenerational Change in PostIndustrial Societies." American Political Science Review 65 (4): 991-1017.

Inoguchi, Takashi, and Seiji Fujii. 2013. The Quality of Life in Asia: A Comparison of Quality of Life in Asia, Vol. 1. Berlin: Springer Science & Business Media.

Kahneman, Daniel, Alan B. Krueger, David Schkade, Norbert Schwarz, and Arthur A. Stone. 2006. "Would You Be Happier If You Were Richer? A Focusing Illusion." Science 312 (5782): 1908-10.

Kahneman, Daniel, and Richard Thaler. 1991. "Economic Analysis and the Psychology of Utility: Applications to Compensation Policy." American Economic Review 81 (2): 341- 46.

Killingsworth, Matthew A. 2021. "Experienced Well-Being Rises with Income, Even Above $75,000 Per Year." Proceedings of the National Academy of Sciences 118 (4): e2016976118.

Koh, Wee Chian. 2020. "Brunei: Economy." The Far East and Australasia 2020, 51st edition. Oxford: Routledge.

Lim, Hock-Eam, Daigee Shaw, Pei-Shan Liao, and Hongbo Duan. 2020. "The Effects of Income on Happiness in East and South Asia: Societal Values Matter?" Journal of Happiness Studies 21 (2): 391-415.

Luo, Jianbo. 2021. "Happiness Adaptation to High Income: Evidence from German Panel Data." Economics Letters 206: 109995.

Martela, Frank, Bent Greve, Bo Rothstein, and Juho Saari. 2020. "The Nordic Exceptionalism: What Explains Why the Nordic Countries Are Constantly among the Happiest in the World." World Happiness Report 7: 129-46.

Mignamissi, Dieudonné, and Yselle Flora Malah Kuete. 2021. "Resource Rents and Happiness on a Global Perspective: The Resource Curse Revisited." Resources Policy 71: 101994.

Ngoo, Yee Ting, Nai Peng Tey, and Eu Chye Tan. 2015. "Determinants of Life Satisfaction in Asia." Social Indicators Research 124 (1): 141-56.

Organisation for Economic Co-operation and Development (OECD). 2013. "OECD Guidelines on Measuring Subjective Well-Being." Paris: OECD Publishing. http://dx.doi. org/10.1787/9789264191655-en.

Oshio, Takashi, Kayo Nozaki, and Miki Kobayashi. 2011. "Relative Income and Happiness in Asia: Evidence from Nationwide Surveys in China, Japan, and Korea." Social Indicators Research 104 (3): 351-67.

Pacek, Alexander C., and Benjamin Radcliff. 2008. "Welfare Policy and Subjective Well-Being across Nations: An Individual-Level Assessment." Social Indicators Research 89 (1): 179-91.

Paul, Satya, and Daniel Guilbert. 2013. "Income-Happiness Paradox in Australia: Testing the Theories of Adaptation and Social Comparison." Economic Modelling 30: 900-10.

Pouwels, Babette, Jacques Siegers, and Jan Dirk Vlasblom. 2008. "Income, Working Hours, and Happiness." Economics Letters 99 (1): 72-74.

Rojas, Mariano. 2007. "The Complexity of Well-Being: A Life-Satisfaction Conception and a Domains-of-Life Approach." In Researching Well-Being in Developing Countries: From Theory to Research, 259-80. Cambridge: Cambridge University Press.

_. 2008. "Experienced Poverty and Income Poverty in Mexico: A Subjective Well-Being Approach." World Development 36 (6): 1078-93.

Ryan, Richard M., and Edward L. Deci. 2001. "On Happiness and Human Potentials: A Review of Research on Hedonic and Eudaimonic Well-Being." Annual Review of Psychology 52 (1): 141-66.

Schoeni, Robert F. 1995. "Marital Status and Earnings in Developed Countries." Journal of Population Economics 8 (4): 351-59.

Slesman, Ly, Ahmad Zubaidi Baharumshah, and Wahabuddin Ra'ees. 2015. "Institutional Infrastructure and Economic Growth in Member Countries of the Organization of Islamic Cooperation (OIC)." Economic Modelling 51: 214-26.

Stack, Steven, and J. Ross Eshleman. 1998. "Marital Status and Happiness: A 17-Nation Study." Journal of Marriage and the Family 60 (2): 527-36.

Stevenson, Betsey, and Justin Wolfers. 2008. "Economic Growth and Subjective Well-Being: Reassessing the Easterlin Paradox." Brookings Papers on Economic Activity Spring.

Tavakol, Mohsen, and Reg Dennick. 2011. "Making Sense of Cronbach's Alpha." International Journal of Medical Education 2: 53.

Tsui, Hsiao-Chien. 2014. "What Affects Happiness: Absolute Income, Relative Income or Expected Income?" Journal of Policy Modeling 36 (6): 994-1007.

Van Praag, Bernard, and Ada Ferrer-i-Carbonell. 2008. Happiness Quantified: A Satisfaction Calculus Approach. Oxford: Oxford University Press.

Vaskovskyi, Anton. 2020. "Natural Resources and Quality of Life: International Evidence." Singapore Economic Review. https://doi.org/10.1142/S0217590820500605.

Veenhoven, Ruut. 2012. "Happiness: Also Known as 'Life Satisfaction' and 'Subjective Well-Being'." In Handbook of Social Indicators and Quality of Life Research, 63-77. Dordrecht: Springer.

Veenhoven, Ruut, and Floris Vergunst. 2014. "The Easterlin Illusion: Economic Growth Does Go with Greater Happiness." International Journal of Happiness and Development 1 (4): 311-43.

Watson, Tara, and Sara McLanahan. 2011. "Marriage Meets the Joneses Relative Income, Identity, and Marital Status." Journal of Human Resources 46 (3): 482-517.

Westreich, Daniel, and Sander Greenland. 2013. "The Table 2 Fallacy: Presenting and Interpreting Confounder and Modifier Coefficients." American Journal of Epidemiology 177 (4): 292-98.