Absolute and relative restriction and consumer behavior: implications for understanding global consumption.

Our premise is that researchers have much to gain from an understanding of the global marketplace experiences of impoverished consumers. We argue that influences of absolute and relative restriction, across peoples and societies, are particularly critical. Therefore, this research makes progress by evaluating consumer data from diverse cultures and nations using hierarchical linear models, revealing ways restriction through poverty and consumption impacts well-being. We find that understanding both absolute and relative poverty is necessary for a more complete picture. Specifically, interactions show that absolute restriction moderates relationships between relative restriction and consumption and well-being by muting or exacerbating the effects.

RESEARCH FRAMEWORK

The role of various restrictions on consumer decision making has received some focus, guided by the premise that all people may face roadblocks to meeting wants and needs despite relative affluence and availability of goods and services (Botti et al. 2008). However, extreme and dire circumstances characterize the consumptive lives of poor citizens around the world, inspiring research on poverty, deprivation and restriction (Hill, Felice, and Ainscough 2007; Viswanathan, Rosa, and Ruth 2010). Recent statistics paint a grim picture and reveal that about three billion people survive on less than $2.50 per day, 640 million lack adequate shelters, and one in seven has no access to health care (http://www.povertycrisis.com). Of course, statistics fail to provide a complete understanding of the impact of poverty restrictions on consumer decision making worldwide. Previous research on subsets of impoverished people gives an indication that nuanced differences resulting from deprivation exist (Hill 1991; Hill and Stamey 1990), but most of this stream of work characterizes poverty in either relative or absolute terms and with smaller subsets of consumers. We advance such perspectives by considering both absolute and relative poverty with a diverse multinational sample.

Accordingly, this study contributes to research by providing a novel examination of the impact of poverty-as-restriction on consumer well-being. Specifically, we evaluate poverty in absolute and relative terms, individually and in combination. Drawing from diverse theories, we expand on previous research by using a multifaceted perspective of impoverishment, juxtaposing absolute restriction (i.e., human development) with relative restriction (i.e., inequality). We are particularly interested in providing as broad a representation as possible to offer insights about the role of impoverishment on a subset of humanity. We continue this discussion by explaining the proposed framework and resulting hypotheses, followed by a description of the study sample and data. Next, we detail the multilevel methodology and analysis used to test our question. We then reveal our findings and highlight how results explain the nuanced role played by poverty-as-restriction on consumption, with implications for the relationship between consumption and well-being to guide consumer theory and research.

Restriction and Consumer Behavior

The general framework that underpins our hypotheses involving restriction reveals that the nature of consumption opportunities impacts the relationship between relative or absolute poverty and subjective evaluations of well-being (Zhong and Mitchell 2010). However, our position is that their interaction also impacts consumer well-being. As a result, perspectives of access to or restriction from goods and services used in this study are based on assessments of absolute material circumstances as well as relative evaluations of what others have (Easterlin 2004). Zagorski, Kelley, and Evans (2007) suggest that these goods and services may be viewed as serving primary needs, which are similar across persons and associated with basic survival, and status desires, which denote social standing and orient us relative to other people. These comparisons have a lengthy history across times and places as consumer culture has developed and matured (Hill and Gaines 2007; Veblen 1934), culminating in feelings of satisfaction or dissatisfaction depending upon the final result (Steward 2006).

Impact of Relative or Absolute Restriction on Well-being

Previous research on relative restriction and well-being calls upon two research streams. The materialism literature in consumer behavior suggests that there is widespread agreement that people and societies characterized by materialistic values tend to experience lower levels of well-being, regardless of country of origin. Thus, consumers are caught in a catch-22. Those who strive to achieve wealth and possessions find they lack real meaning in their lives and experience lower levels of well-being (Arndt et al. 2004). Yet relative deprivation theory suggests that impoverished consumers, whose primary needs and desires are largely unmet, face even greater discontent that manifests when they see unfair discrepancies between their material situations and that of more affluent others (see Davis 1959). Runciman (1966, p. 10) provides this explanation: "If A does not have something but wants it, compares himself to B, who does have it, then A is 'relatively deprived' with reference to B." Importantly, this approach to social comparisons shows that the wider the discrepancies within a society, the greater the negative impact on the very poor (Hill 2001).

There also has been much debate in the consumer literature on the influence of absolute restriction on various measures of well-being (e.g., Wood and Bettman 2007). The question often asked is: "Are wealthier members of society usually happier than the poor?" (Easterlin 1974, p. 90). Researchers tasked with addressing this issue typically arrive at similar conclusions. First, they find that citizens across nations share an interest in and concern for their material lives on par with or above most other important life topics (Easterlin 2000; Hill, Peterson, and Dhanda 2001). Second, their scholarship reveals that subjective evaluations best capture the aspects of human well-being that include life satisfaction (Inoguchi and Shin 2009). Third, well-being is tied to the absolute nature of material deprivation, especially in nations characterized by severe impoverishment (Hill, Felice, and Ainscough 2007). In the final analysis, while the poor have coping mechanisms to buoy themselves and their material lives, the absolute restriction in their material environments significantly reduces their well-being relative to affluent counterparts.

Yet these simple relationships between restriction and well-being fail to take into account consumption, even though it is at the center of most consumer theory (Rindfleisch, Burroughs, and Wong 2009). Consider Zagorski, Kelley, and Evans (2007), who posit goods and services have "absolute aspects" common across consumers that do not distinguish one person from the next, while other products have "positional aspects" that are reflective of status and determine our comparative standing. Hsee et al. (2009) present a similar approach to possessions, but categorize goods and services rather than their characteristics as Type A/basic commodities and Type B/status products. In a look at middle class vs. impoverished consumers' expenditures in thirteen countries, Banerjee and Duflo (2007) found that wealthier consumers invest more in goods and services that advance personal status relative to basic commodities when compared to poorer consumers, despite their joint preference for status-enhancing products. The consumer literature opines that these less-affluent consumers are aware of and desire many of the same consumption options, even if they lack the resources to acquire them (Ozanne, Hill, and Wright 1998; Rucker and Galinsky 2008). Taken together, we hypothesize that the consumption of status-enhancing products moderates the restriction-well-being relationship as follows:

H1: Access to the preferred set of higher-order goods and services moderates/reduces negative relationships between relative restriction and well-being.

H2: Access to the preferred set of higher-order goods and services moderates/reduces negative relationships between absolute restriction and well-being.

Impact of Relative and Absolute Restriction

Regarding our expectations for interaction effects, we turn to the small body of collective literature that signals dissension in the ranks and crosses the absolute/relative divide. While D'Ambrosio and Frick (2006) believe that relative restriction has more influence over well-being than absolute restriction, Hsee and colleagues (2009) examine both constructs and indicate that the importance of one vs. the other remains unresolved. Early research by Easterlin (1974) is in agreement that something is amiss given variance in well-being across rich and poor countries cannot be adequately accounted for by looking individually at either of these constructs (also see Easterlin 2009 for additional discussion). His body of work intimates that consumption norms arise within societies based on the availability of goods and services, and if some consumers are denied them due to impoverishment, their value rises and well-being for those who must go without may fall (Inglehart 2000 comes to a similar conclusion).

Given the focus of our research, we concentrate this portion of the investigation on some possible differences in well-being as absolute restriction increases under conditions of low vs. high relative restriction. The higher absolute restriction nations are often characterized by a lack of goods and services as well as severely limited access to marketplaces by most consumers (Hill and Gaines 2007), creating even starker differences between the haves and have-nots if they exist. The consumer and quality-of-life literatures tell us that high absolute restriction conditions may cause great hardship, depressing measures of life quality or well-being (see Chakravarti 2006). Yet the materialism and relative deprivation literatures suggest the possibility that some consumers are able to make positive social comparisons under conditions of high relative restriction, while low relative restriction counterparts in their entirety lack access and subsequent status-differentiating possessions (see Runciman 1966). This situation leads to the perverse outcome of lower well-being among the worst off when greater equality exists, and we hypothesize:

H3: Rising absolute marketplace restrictions increases negative effects on well-being for consumers living in low relative restriction nations compared to consumers living in high relative restriction nations.

METHOD

Variables

Individual, Subjective Well-being

Our framework examines how consumption and restriction influence individual well-being. To craft appropriate measures, we consider respondents' subjective assessments of overall happiness and life satisfaction. These items are measured in the World Values Survey (WVS), and they are commonly accepted variables for the study of well-being (Diener, Oishi, and Lucas 2003; Easterlin 2001), where life satisfaction represents the cognitive component of subjective well-being and happiness represents the affective component (e.g., Chirkov et al. 2003; Oishi et al. 2007; Suh 2002; Sub et al. 1998). As suggested by this literature and because these items are correlated highly in our sample (r > .80), we combine them into a single well-being measure. Reliabilities are computed for the well-being measure of every country to ensure each meets the acceptable [alpha] [greater than or equal to] .70 cutoff (e.g., Cullen, Parboteeah, and Hoegl 2004). With regard to individual measures, the respondents were asked to state subjective levels of happiness on a four-point scale. To assess life satisfaction, respondents were asked to provide overall states with a ten-point scale ranging from "completely dissatisfied" to "completely satisfied." Because of differences between metrics, these items are standardized and then combined.

Individual-level Controls

Because we are interested in forces at the country level that influence individual well-being we employ a multilevel model. We also control for a number of individual-level factors, which is common practice in cross-national, multilevel well-being research (e.g., Zagorski, Kelley, and Evans 2007). Although literature suggests that personal income and social class self-assignments are important in evaluations of well-being (Coleman 1983; Henry 2005; Schaninger 1981), we are interested in how broader characteristics of the environment manifest in consumption along with well-being outcomes. Thus, we control for individual income and social class to isolate their effects. Income level is broken into deciles relevant to the local environment. Also included is a measure of social class, which asks people to note the level of social class that they identify with (e.g., middle class, working class and lower class). Findings also control for respondents' age, gender and years of education (e.g., Schimmack et al. 2002). Variables are obtained from the WVS and are listed in the measure Appendix 1 that comprises all study variables and data sources.

Of note, control variables featured in our analyses have been tested in various functional forms in past research. For example, income and social class have been tested as nonlinear and quadratic functions, and as continuous and categorical variables (Howell and Howell 2008). So too, age has been tested as a nonlinear function, with similarities reported between consumers at different age extremes (Blanchflower and Oswald 2004; Ferrer-i-Carbonell and Gowdy 2007). To be thorough, analyses that follow are computed using all possible functional forms of the control variables as suggested in the literature. Importantly, none of our findings, in terms of direction, significance, or strength, vary with the different functional forms. Thus, we report the least complex analyses that feature control variables as they appear in the WVS.

Higher-order Goods and Services

The World Consumer Lifestyles (WCL) Database provides expenditure data as a percentage of total household consumption. Categories include food, housing, health care, education, clothing and footwear, household goods and services, leisure and recreation and financial services. Research has examined the link between development and food expenditures, positing that as standard of living increases, expenditures on food relative to other consumption categories decreases (Banerjee and Duflo 2007). Our goal is to provide a more detailed and representative picture of consumption beyond simple comparisons between food and nonfood expenditures. Accordingly, we extracted consumption categories of leisure and recreation, housing and financial services as they are theoretically representative of higher-order fulfillment goods and services (e.g., Fournier, Antes, and Beaumier 1992). Although this measure has strong face validity, factor analysis of all consumption categories confirmed our assertion, as the three categories together (and absent any other consumption categories) loaded on a single factor. Factor loadings are provided in Appendix 1. Of note, a second category emerged in our factor analysis that reflects basic survival goods and services (food, clothing and footwear and household goods and services). When combined, the measure correlates strongly and negatively (r =-.80), with our higher-order fulfillment measure. (1)

Restriction

We use two widely accepted and validated measures of absolute restriction or poverty and relative restriction or inequality (Hill and Gaines 2007). To assess absolute poverty, we use the United Nations Human Development Index (HDI), which is a composite of three consumption-related variables of longevity/life expectancy and is a surrogate for available health care; knowledge/literacy, which measures access to and usage of educational opportunities; and standard of living/wealth, which is defined as GDP per capita. These indicators are reduced to a simple weighted scale that varies between zero and one, with one representing the highest human development attainment under current global conditions. We subtract this value from one so that higher values represented greater absolute poverty. To evaluate relative poverty, the Gini index is employed to measure the extent to which distribution of income/consumption among individuals deviates from an equitable sharing of resources within countries. Values close to zero suggest near equal consumption potential, while values close to 100 show vast disparities or high inequality among socioeconomic groups within a society.

Our thirty-eight-country sample provides a broad profile of nations with various combinations of absolute and relative restriction. For example, countries like South Africa rank highly in terms of absolute poverty and relative inequality. In contrast, Sweden, Norway and Australia experience low levels of absolute poverty and relative inequality because of high human development and redistributive economic policies. Many countries in our sample, however, experience disparities in rankings, such as the United States which ranks low in absolute poverty but high in relative inequality or restriction. An alternative picture is provided by many Eastern European countries that rank high in absolute poverty, but again, redistributive policies subject their citizens to lower relative inequality. Please see Figure 1 for a graphical depiction of each country's scores on these measures. This wide variety of nations with varying and conflicting degrees of each restriction provides a comprehensive test of our question and promotes the external validity of findings.

Country-level Controls: Culture

Culture variables are adapted from the Global Leadership and Organizational Behavior Effectiveness (GLOBE) study (House et al. 2004) using the "as is" indices reflective of actual societal practices. The GLOBE study assessed individualism/collectivism with four items loosely based on previous work by Hofstede (2001). The GLOBE study restructured the measure to focus on notions of whether societal values and practices promote individual vs. group interests (Gelfand et al. 2004). The measure was reliable at [alpha] = .70. Power distance assessed notions of societal hierarchy and cultural acceptance of authority, influence and status privileges (Carl, Gupta, and Javidan 2004). The GLOBE power distance measure also was inspired by Hofstede (2001), and it was reliable at [alpha] = .93.

Sample and Data Sources

To examine empirically our model of consumer behavior and restriction, we assemble a multilevel data set that includes individual consumer and country-level information. Individual-level data are derived from the most recent wave of a large-scale multinational study, the WVS (2005-2008). Country-level data are derived from the Euromonitor International (EI) and the United Nations Human Development Report (2007-2008). Details on each data source follow.

We obtain all individual-level measures, including individual subjective well-being items and individual demographic variables, from the WVS. The WVS is conducted by the University of Michigan Inter-University Consortium for Political and Social Research. Our study employs the current version of the WVS, the 2005-2008 wave, which included adults aged eighteen and over from fifty-six countries worldwide. WVS data are gathered in each country by local universities and/or social science research organizations through personal interviews using survey instruments that were translated and back-translated across countries to promote consistency. In each nation, respondents were recruited by local research organizations using a stratified systematic sample research design to ensure the sample was representative of the country's true population. Specifically, samples for each country were stratified by age, education, ethnicity and gender. Each nation produced a minimum of 1,000 completed responses. The average sample size for the countries included in our study is 1,480.

[FIGURE 1 OMITTED]

We are interested in how consumption patterns and country-level restriction influence individual outcomes, and sought statistics at the country level to juxtapose with our individual-level sample. Our goal is to maximize country coverage so as to represent as much of the world's population as possible; however, we were constrained to those countries with available data that overlapped with the WVS. Naturally, this is a limitation to our study, as with most cross-national research relying on secondary measures. Appropriate country-level data are available for thirty-eight nations (listed in Figure 1) that correspond with our individual-level sample. This produced an individual-level sample (level 1) of 56,261 consumers and a country-level sample (level 2) of thirty-eight countries. The mean age of respondents is 41-years-old and 52% are female. Respondents average a tenth-grade education. Fifty-five percent are married, with a mean of two children. Respondents report an average income in the fourth decile, with one the lowest and ten the highest category relative to local economic situations. The predominant self-reported social classification is "working class." Descriptive statistics within our subset of the WVS match those taken from the entire survey, revealing the representativeness of our thirty-eight-country sample.

Consumption data are obtained from the research arm at El through the WCL research program. These WCL data are derived from multiple sources gathered by research analysts located within each of the countries profiled to provide a complete and accurate picture of consumption for the nation in its entirety. Country statistics represent consumption by the full population, as measured as a percentage of total consumer expenditure. First, the WCL research program obtains purchase data onsite at retail locations across the featured nations. WCL researchers also spend time onsite in the form of store checks to verify validity of these figures and to look for shifting trends. Second, the WCL works in partnership with National Statistical Offices in each country to verify local behaviors and consumption trends. Third, data are obtained from international organizations including the World Bank and the World Health Organization to supplement local and national data for individual consumption categories. In data collection, all definitions were translated and back-translated to ensure accuracy and comparability. Euromonitor researchers attest to WCL data's international comparability (World Consumer Lifestyles Databook 2005).

Restriction data were available in the United Nations' Human Development Report 2007-2008, which coincides with individual-level data from the most recent wave of the WVS 2005-2006. Thus, all three data sources examin individual- and country-level conditions at the same point in time. Since the WCL data restricts our analysis to thirty-eight countries, we limit the extraction of our restriction and inequality data from the Human Development Report (HDR) to the same group of nations. As with the consumption data, the statistics featured in the HDR are collected and compiled by local and international data agencies. If discrepancies arise between national and international statistics organizations, the UNDP actively intervenes to resolve inconsistencies to produce the most accurate statistics possible (Human Development Report 2007-2008, p. 221). The HDR is known for its extensive country coverage, and is widely regarded as a representative, internationally comparable source for country-level information and data-derived indicators (Hill and Gaines 2007).

To control for potential effects of societal culture such as individualism/collectivism and power distance, we use measures from the GLOBE study (House, Hanges, Javidan, Dorfman, and Gupta 2004). Designed to conceptualize and validate country cultural dimensions, the GLOBE study is a long-term multi-phase program involving scholars from around the world. Grounded in the theoretical notions of foundational culture researchers (e.g., Hofstede 2001; Schwartz 1992; Trompenaars and Hampden-Turner 1998), House and colleagues conducted a cross-national, cross-industry study involving sixty-two countries with the aim of advancing and refining cultural variables. GLOBE measures better reflect contemporary theoretical culture arguments, and further, these measures overcome noted limitations of the Hofstede country data (gathered primarily in the late 1960s and early 1970s) as well as address recent statistical criticisms of his model and analyses (e.g., Earley 2006; Smith 2006). Because we are interested in individuals' perceptions of their state of well-being, we rely on the "as is" measures from the GLOBE study reflective of participants' views of the society as it actually exists.

ANALYSIS AND RESULTS

Cross-National Comparisons

Before making comparative assessments across the various countries, it is important to ascertain that differences between groups are not due to discrepancies in interpretation of the measurement instrument. Given the lack of multiple item measures available in the WVS corresponding to our theoretical framework, traditional measurement invariance approaches using confirmatory factor analysis are inappropriate (Steenkamp and Baumgartner 1998). Accordingly, we use methodological prescriptions for item bias analysis advocated by van de Vijver and Leung (1997) for evidence of cross-national comparability. Specifically, we test differential item functioning using a series of ANOVAs for each item to evaluate if individuals in the thirty-eight various nations interpreted these items differently. The level of absolute restriction and relative inequality and individual's scores summed across individual-level items serve as independent variables, and the individual item score as dependent variable to test for uniform bias. The interaction between summed score level and restriction is included to test for nonuniform bias. Neither main effects nor interaction effects for each study item is significant (all p-values at .20 or greater), supporting cross-national comparison.

Analytic Protocol

Because we are interested in relationships between well-being at the individual level and consumption, restriction and inequality as manifest at the country level, multilevel modeling is required. Specifically, variance may be partitioned between the individual-level and country-level variables (Hofmann 1997). Hierarchical linear modeling (HLM) with restricted maximum likelihood estimation using HLM 7 is employed (Raudenbush and Bryk 2001), which provides a platform for investigating relationships between variables at different levels of analysis. Specifically, HLM models both individual- and group-level residuals, recognizing the partial interdependence of individuals within the same group or, in our case, the same country. To accomplish this, HLM simultaneously estimates two models: the first modeling relationships within each individual-level unit, and the second modeling how those relationships within units vary between countries (Hofmann 1997). Put simply, we regress individual well-being on personal characteristics, and then save the estimated individual fixed effect or intercept so that we can regress that on important country characteristics. Model equations and details follow:

Baseline Model (Main Effects)

Level 1 (Individual level)

[WELL.sub.ij] = [[beta].sub.0j]+[[beta].sub.ij] + [INC.sub.ij] + [[beta].sub.2j] [SOCICLS.sub.ij] + [[beta].sub.3j] [AGE.sub.ij] + [[beta].sub.4j] [GEND.sub.ij] + [[beta].sub.5j] [EDUC.sub.ij] + [r.sub.ij]

Level 2 (Country level)

[[beta].sub.0j] = [[gamma].sub.00] + [[gamma].sub.01] [HIGHERORD.sub.j] + [[gamma].sub.02] [ABSRST.sub.j] + [[gamma].sub.03] [RELRST.sub.j] + [[gamma].sub.04] [INDCOL.sub.j] + [[gamma].sub.05] [PWRDIST.sub.j] + [U.sub.0j];

[[beta].sub.1j] = [[gamma].sub.10]+ [U.sub.1j]; [[beta].sub.2j] =[[gamma].sub.20]+ [U.sub.2j]; [[beta].sub.3j]= [[gamma].sub.30]+ [U.sub.3j]; [[beta].sub.4j] = [[gamma].sub.40] + [U.sub.4j];

[[beta].sub.5j] = [[gamma].sub.50] + [U.sub.5j]

Hypothesis Testing Model (with Interactions)

Level 1 (Individual level)

[WELL.sub.ij] = [[beta].sub.0j]+[[beta].sub.1j] [INC.sub.ij] + [[beta].sub.2j] [SOCICLS.sub.ij] [[beta].sub.3j] [AGE.sub.ij] + [[beta].sub.4j] [GEND.sub.ij] + [[beta].sub.5j] [EDUC.sub.ij] + [r.sub.ij]

Level 2 (Country level)

[[beta].sub.0j] = [[gamma].sub.00]+ [[gamma].sub.01] [HIGHERORD.sub.j] + [[gamma].sub.02] [ABSRST.sub.j] + [[gamma].sub.03] [RELRST.sub.j]+ [[gamma].sub.04] HIGHEROR * [ABSRST.sub.j] + [[gamma].sub.05] HIGHEROR * [RELRST.sub.j] + [[gamma].sub.06] ABSRST * [RELRST.sub.j] + [[gamma].sub.07] [INDCOL.sub.j] + [[gamma].sub.08] [PWRDIST.sub.j] + [U.sub.0j];

[[beta].sub.1j] = [[gamma].sub.10] + [U.sub.1j]; [[beta].sub.2j] =[[gamma].sub.20]+ [U.sub.2j]; [[beta].sub.3j] = [[gamma].sub.30] + [U.sub.3j];

[[beta].sub.4j] = [[gamma].sub.40] [U.sub.4j]; [[beta].sub.5j] = [[gamma].sub.50]+ [U.sub.5j]

where WELL represents the outcome measure, individual well-being, for individual i in country j. [[beta].sub.0j] is the intercept, and [[beta].sub.1j] through [[beta].sub.5j] are slopes estimated separately for each country, signified by the subscript j. Specifically, INC represents individual income category reported; SOCICLS signifies individual social class categorization; AGE represents individual's age; GEND represents gender, which is a dummy coded variable; EDUC is individual's age upon completion of education. Finally, for the level 1 equation, [r.sub.ij] is the residual, which is normally distributed with a zero mean and variance [[sigma].sup.2].

In our model, level 2 analyses use the intercept from level 1 analysis as a dependent variable. In these equations, [[gamma].sub.00] through [[gamma].sub.50] represent the second stage intercept terms, and [[gamma].sub.01] through [[gamma].sub.03] are the slopes relating the country-level variables in each of the three models to the intercept terms from the level 1 equation. Finally, [U.sub.0j] through [U.sub.5j] are the level 2 residuals, which are multivariate and normally distributed, with an expected value of zero and variance [[tau].sub.00] (Raudenbush and Bryk 2001). In each model for [[beta].sub.1j]- [[beta].sub.5j] in the level 2 equations, we do not include the country-level variables (basic survival and long-term betterment consumption or restriction) since we do not hypothesize country-level effects on the individual-level slopes (i.e., slopes-as-outcomes models for income, social class, age, gender, education). Rather, we investigate traditional intercepts-as-outcomes models. Specifically, at the country-level of analysis, in relation to our hypotheses HIGHERORD is consumption of higher-order goods and services. ABSRST represents the country's level of absolute restriction; RELRST is the country's level of relative restriction. We creat product terms for each using the same distinctions. Each model also controls for the potential role of societal culture including INDCOL representing individualism/collectivism, and PWRDIST symbolizing power distance. We test each model using a random coefficients approach.

Descriptive statistics are provided in Table 1. Since our variables of interest are measured at different levels of analysis (individual- and country-level), we attach the country-level items to each individual case to obtain cross-level correlations. We then counterweight by country sample size so that nations with larger sample sizes are not overrepresented. This approach to correlations with multilevel data is common in cross-national research (e.g., Martin et al. 2007); however, it is subject to a disaggregation bias described by Hofmann (1997). Thus, although correlations provide insights into the nature of variable relationships, the HLM analyses that follow provide a more conservative test and truer picture. Nonetheless, the correlation values between the country-level variables warranted further examination. Using ordinary least squares regression, we analyzed the country-level variables to examine collinearity diagnostics. The dependent variable in this analysis was well-being collapsed to the country mean. The n for this analysis was thirty-eight and collinearity diagnostics including tolerance and VIF coefficients all were well in the range of acceptability. This conservative test for multicollinearity gave us confidence in the remainder of our analyses and hypotheses testing. Please see Table 2 for these results.

Findings

Presentation of results is organized along the framework in Figure 2. Findings are interpreted in detail in the discussion, and allow for a revised model of consumption processes in light of inequality and restriction. Although we do not proffer main effects hypotheses for higher-order consumption and each form of restriction, we present this baseline model in Table 3. Intuitive to our theoretical development, greater consumption of higher-order goods and services significantly and positively ([gamma] = .101; p < .05) influences well-being. Country-level absolute poverty has a significant and negative effect ([gamma] = -.119; p < .001) on individual well-being as one would expect. Relative inequality has an unexpectedly positive effect on individual well-being ([gamma] = .197; p < .01). With regard to our hypothesized interactions, we do not support Hypothesis 1. Rather, the combined influence of higher-order goods and services consumption and relative restriction do not have a significant influence on individual well-being. Given this finding and the unexpected result for relative inequality's main effect, it appears that future work exploring the complex role of inequality or relative restriction is needed.

[FIGURE 2 OMITTED]

For Hypothesis 2, we find absolute poverty moderates the higher-order consumption-well-being relationship ([gamma] = -.118; p < .01) as expected. We show that as consumption of higher-order fulfillment goods and services increases, individual well-being dramatically increases, but only for individuals living in societies with low absolute restriction or poverty. For individuals in high poverty or high absolute restriction societies, there is no effect on well-being as a result of consumption. We graphically depict this relationship in Figure 2. The sharply increasing line demonstrates the improvement in individuals' self-reported well-being because of greater consumption of higher-order goods and services. The flat line represents high poverty or high absolute restriction societies. We observe low overall level of self-reported well-being regardless of ability to consume higher-order goods and services.

In Hypothesis 3 we consider the moderating relationship between societal absolute and relative restriction and individual well-being ([gamma] = .086; p < .05). Consistent with counterintuitive findings for relative restriction, interpretation of this result goes beyond our initial predictions. Increasing absolute poverty negatively impacts individuals' self-reported well-being. Yet this effect is more pronounced for individuals in low inequality countries as shown in the graphical depiction featured in Figure 3. Consumers living under high relative restriction experience less of a decline in well-being. Finally, it is noteworthy that neither societal culture variable included in the analyses as control variables is significant, revealing powerful effects of consumption and restriction on well-being.

DISCUSSION

Our premise is scholars primarily study ways consumers from developed countries navigate marketplaces characterized by relative affluence rather than absolute and relative poverty. Given the depth and breadth of poverty worldwide, we echo Chakravarti (2006) by arguing that models based primarily on the former present a distorted picture of how purchase decisions and their outcomes play out across vast socioeconomic chasms around the world. Our work expands such thinking by merging the conceptual underpinnings of the limited consumer scholarship on impoverished consumers (e.g., Hill 2001; Hill and Gaines 2007) with work in the larger social sciences that considers the impact of consumption within- and between-countries (e.g., Easterlin 2001; Shin 1980; Zagorski, Kelley, and Evans 2007). Importantly, it goes beyond studies to date through examination of absolute and relative poverty and their interactions with consumption and individual well-being. This focus suggests two categories of goods and services that meet different needs, and their emphasis shifts from basic survival to higher-order need fulfillment as development advances. Yet simple explanations fail to capture nuances associated with ways these forms of restriction manifest within the context of impoverished circumstances.

[FIGURE 3 OMITTED]

Our research explores three paths. The first involves Hypothesis 1, and focuses on the poverty-consumption linkage and well-being. Interestingly, relative restriction did not moderate the relationship between higher-order consumption and well-being. For Hypothesis 2, however, we find that absolute restriction does moderate the relationship between higher-order goods and services consumption and well-being. Our results thus show that ability to consume higher-order goods as part of one's personal consumption portfolio can improve well-being dramatically. Yet, as evidenced by support for Hypothesis 2 and graphically depicted in Figure 2, benefits of higher-order goods consumption only hold in societies with low absolute poverty or restriction. Figure 2 reveals the detriments to individual well-being in high poverty societies regardless of the nature of consumption.

Hypothesis 3 also presents a novel finding in that high absolute poverty leads to lower well-being, but, surprisingly, the impact of relative poverty is theoretically unexpected. Figure 3 shows that individuals living in countries with greater inequality or greater relative restriction report greater well-being regardless of the societal level of absolute poverty or restriction. In these societies, greater absolute poverty or restriction strongly reduces well-being. Clearly, the complex role of inequality or relative restriction on well-being warrants future investigation and scholarly research, particularly in the consumer sciences.

By looking at well-being as influenced by country-level and individual-level factors for thousands of consumers simultaneously across levels, we contribute to consumer research and research on poverty and well-being worldwide. Nonetheless, such an exploration is subject to limitations. Specifically, the use of secondary data constrained the questions, measures, and countries available to us. We also are mindful of interpretation difficulties when data, hypotheses and findings cross levels of analysis. Future research would benefit from marrying primary investigations at the individual level with what is known from these influential secondary data research programs.

The value of this investigation is measured by the extent to which our perceptions of consumer behavior shift relative to current thinking. Consumer research has taken note of real differences across peoples and nations due to globalization of theory and practice, but scholars have been slow to change primary conceptualizations as a consequence. As described earlier, distinctions between consumption environments characterized by too much instead of too little portend a fundamental but obvious difference that does nothing more than reduce the size rather than the nature of the context. The implicit belief is that substantial modifications to our notions of motivations, decision processes and outcomes associated with consumption are unnecessary, with few adjustments required to incorporate resource restrictions. Hopefully, findings from our study make the case for additional reflection on this position. We readily acknowledge that the consumers who have unfortunate material lives focused on acquisition of basic commodities vs. personal enrichment suffer emotional consequences. However, beliefs that rising well-being stagnates above some level of affluence is less important than recognizing that the poor continue to be less happy and satisfied with their lives.

Yet this result leaves us where we have always been--looking for ways to apply typical viewpoints in different resource contexts. Instead, poverty-as-restriction presents a novel vantage point from which to observe consumption. For instance, ability to consume higher-order goods and services can improve individual well-being, but the influence is not so simple. When low levels of absolute poverty exist (i.e., relative marketplace affluence), consumers who fulfill higher-order needs through consumption report greater well-being. The opposite is true for high levels of absolute poverty (relative marketplace impoverishment), with consumers experiencing low individual well-being regardless of the nature of consumption. In other words, higher-order goods and services do not abate the damaging effects of significant impoverishment.

Our results suggest that marketplace abundance in combination with relative affluence modifies substantially traditional consumer behavior characterizations. Marketplace abundance is based on availability of goods and services from the least to best possible assortment (Hill and Gaines 2007), and marketplace access focuses on relative equity within countries and ranges from low to high. Together they reveal a different landscape of how consumption occurs around the world than they do when separated out and partitioned into micro-level investigations. Thus, our study rejects implicit assumptions that consumer processes are uniform and can be summarized across nations and peoples. Scholars who examine global business practices are aware of such distinctions, but much work has concentrated on answering concerns about cultural disparities rather than marketplace nuances. Our findings and conclusions demonstrate that differences in consumption options and resulting well-being also are impacted by levels of absolute poverty and relative inequality, which modifies perceptions of the marketplace. Our understanding of the consumer culture requires that we ask "How am I doing?" which suggests a broader look at what the material world has to offer even if it does not offer it to all.

DOI: 10.1111/j.1745-6606.2012.01225.x

APPENDIX 1 STUDY MEASURES

Individual-level Variables

Well-being: World Values Study Group (2005-2006), Inter-University Consortium for Political and Social Research. Items standardized prior to combining ([alpha] = .80).

Happiness: Taking all things together, would you say you are: 1 = very happy; 2 = rather happy; 3 = not very happy; 4 = not at all happy. (reverse coded).

Life Satisfaction: All things considered, how satisfied are you with your life as a whole these days? Where 1 = completely dissatisfied and 10 = completely satisfied.

Individual Controls: World Values Study Group, World Values Survey (2005-2006), Inter-University Consortium for Political and Social Research.

Income: Grouped by decile; unique to each country, where 1 = lowest decile; 10 = highest decile.

Social Class: "People sometimes describe themselves as belonging to the working class, the middle class, or the upper or lower class. Would you describe yourself as belonging to the ..." (where 1 = upper class; 2 = upper middle class; 3 = lower middle class; 4 = working class; 5 = lower class; reverse coded)

Education: "At what age did you (or will you) complete your full time education?"

Age: "Can you tell me your year of birth, please? This means you are--years old."

Gender: male = 1; female = 0; Completed by interviewer.

Country-level Variables

Consumption: Euromonitor International, Worldwide Consumer Lifestyles Databook (2005). Data represent portion of consumer budget spent on the following categories. Please note that consumption of education and health services were removed from consideration due to overlap with elements of restriction variables. Theoretically derived consumption categories confirmed using factor analysis, with loadings in parentheses.

Higher-order goods and services consumption: housing (.68), leisure and recreation (.84), financial services (.65).

Restriction: United Nations Human Development Program, Human Development Report (2007-2008). (The report is delayed two years, and we are interested in 2005-2006 data corresponding with the World Values Survey.)

Absolute Restriction

Human Development Index: composite of longevity/life expectancy; knowledge/literacy; standard of living/GDP per capita. Values subtracted from 1 to represent restriction.

Relative Restriction

Gini Index: deviation of income sharing from equitable distribution, where values close to zero signal near equality and values close to 100 signal near perfect inequality.

Culture: GLOBE Cross-Cultural Research Program (House et al. 2004) (Unless indicated, all items 1 = strongly disagree, 7 = strongly agree) Individualism/Collectivism ([alpha] = .70)

(all items reverse-coded so greater values represent more individualistic cultures)

1. In this society, leaders encourage group loyalty even if individual goals suffer.

2. The economic system in this society is designed to maximize: (Individual interests/Collective interests)

3. In organizations, leaders encourage group loyalty even if individual goals suffer.

4. The pay and bonus systems in organizations are designed to maximize: (Individual interests/Collective interests)

Power Distance ([alpha] = .93)

1. In this society, followers are expected to: (Question their leader when in disagreement/Obey their leader without question)

2. In this society, power is: (Shared throughout society/Concentrated at the top)

3. In organizations, subordinates are expected to: (Question their boss when in disagreement/Obey their boss without question)

4. In organizations, a person's influence is based primarily on: (One's ability and contribution to the organization/The authority of one's position)

ACKNOWLEDGMENTS

The authors are grateful for the insightful feedback of Ken Manning and three anonymous reviewers. The financial support of the Naclerio family is warmly appreciated.

REFERENCES

Arndt, Jamie, Sheldon Solomon, Tim Kasser, and Kennon M. Sheldon. 2004. The Urge to Splurge: A Terror Management Account of Materialism and Consumer Behavior. Journal of Consumer Psychology, 14 (3): 198-212.

Banerjee, Abhijit V. and Esther Duflo. 2008. What is Middle Class about the Middle Classes Around the World? Journal of Economic Perspectives, 22 (Spring): 3-28.

Botti, Simona, Susan Broniarczyk, Gerald Haubl, Rou Hill, Yanliu Huang, Barbara Kahn, Praveen Kopalle, Donald Lehmann, Joe Urbany, and BrianWansink. 2008. Choice Under Restrictions. Marketing Letters, 19 (July): 183-199.

Carl, Dale, Vipin Gupta, and Mansour Javidan. 2004. Power Distance. In Culture, Leadership, and Organizations: The GLOBE Study of 62 Societies, edited by Robert J. House, Paul J. Hanges, Mansour Javidan, Peter W. Dorfman, and Vipin Gupta (513-563). Thousand Oaks, CA: Sage.

Chakravarti, Dipankar. 2006. Voices Unheard: The Psychology of Consumption in Poverty and Development. Journal of Consumer Psychology, 16 (4): 363-376.

Chirkov, Valery, Richard M. Ryan, Youngmee Kim, and Ulas Kaplan. 2003. Differentiating Autonomy from Individualism and Independence: A Self-Determination Theory Perspective on Internalization of Cultural Orientations and Well-Being. Journal of Personality and Social Psychology, 84 (January): 97-110.

Coleman, Richard P. 1983. The Continuing Significance of Social Class to Marketing. Journal of Consumer Research, 10 (December): 265-280.

Cullen, John B., K. Praveen Parboteeah, and Martin Hoegl. 2004. Cross-National Differences in Mangers' Willingness to Justify Ethically Suspect Behaviors: A Test of Institutional Anomie Theory. Academy of Management Journal, 47 (June): 411-421.

D'Ambrosio, Conchita, and Joachim R. Frick. 2006. Income Satisfaction and Relative Deprivation: An Empirical Link. Social Indicators Research, 81:497-519.

Davis, James A. 1959. A Formal Interpretation if the Theory of Relative Deprivation. Sociometry, 22: 280-296.

Diener, Ed., Shigehiro Oishi, and Richard E. Lucas. 2003. Personality, Culture, and Subjective Wellbeing: Emotional and Cognitive Evaluations of Life. Annual Review of Psychology, 54: 403-425.

Earley, P. Christopher. 2006. Leading Cultural Research in the Future: A Matter of Paradigms of Taste. Journal of International Business Studies, 37 (November): 922-931.

Easterlin, Richard A. 1974. Does Economic Growth Improve the Human Lot'? In Nations and Households in Economic Growth: Essays in Honor of Moses Abramowitz, edited by Paul A. David and Melvin W. Reder (89-125). New York: Academic Press.

--. 2000. The Globalization of Human Development. Annals of the American Academy of Political and Social Science, 570 (July): 32-48.

--. 2001. Income and Happiness: Towards a Unified Theory. The Economic Journal, 111 (July): 465-484.

--. 2004. The Economics of Happiness. Daedalus, 133 (Spring): 26-33.

--. 2009. Lost in Translation: Life Satisfaction on the Road to Capitalism. Journal of Economic Behavior & Organization, 71: 130-145.

Euromonitor. 2005. World Consumer Lifestyles Databook. London: Euromonitor International.

Fournier, Susan, D. Antes, and G. Beaumier. 1992. Nine Consumption Lifestyles. In Advances in Consumer Research, edited by John Sherry and Brian Sternthal (329-337). Provo, UT: Association for Consumer Research.

Gelfand, Michele J., Dharm P. S. Bhawuk, Lisa Hisae Nishii, and David J. Bechtold. 2004. Individualism and Collectivism. In Culture, Leadership, and Organizations: The GLOBE Study of 62 Societies, edited by Robert J. House, Paul J. Hanges, Mansour Javidan, Peter W. Dorfman, and Vipin Gupta (437-512). Thousand Oaks, CA: Sage.

Henry, Paul C. 2005. Social Class, Market Situation, and Consumers' Metaphors of (Dis) Empowerment. Journal of Consumer Research, 31 (March): 766-788.

Hill, Ronald Paul. 1991. Homeless Women, Special Possessions, and the Meaning of 'Home': An Ethnographic Case Study. Journal of Consumer Research, 18 (December): 298-310.

--. 2001. Surviving in a Material World: The Lived Experience of People in Poverty. South Bend, IN: University of Notre Dame Press.

Hill, Ronald Paul, William Felice, and Thomas Ainscough. 2007. International Human Rights and Consumer Quality of Life: An Ethical Perspective. Journal of Macromarketing, 27 (Fall): 370-379.

Hill, Ronald Paul, and Jeannie Gaines. 2007. The Consumer Culture of Poverty: Behavioral Research Findings and Their Implications in an Ethnographic Context. Journal of American Culture, 30 (March): 81-95.

Hill, Ronald Paul, and Mark Stamey. 1990. The Homeless in America: An Examination of Possessions and Consumption Behaviors. Journal of Consumer Research, 17 (December): 303-321.

Hofmann, David A. 1997. An Overview of the Logic and Rationale of Hierarchical Linear Models. Journal of Management, 23 (6): 723-744.

Hofstede, Geert. 2001. Culture's Consequences: Comparing Values, Behaviors. Institutions and Organizations Across Nations. Thousand Oaks, CA: Sage.

House, Robert J., Paul J. Hanges, Mansour Javidan, Peter W. Dorfman, and Vipin Gupta. 2004. Culture, Leadership, and Organizations: The GLOBE Study of 62 Soeieties. Thousand Oaks, CA: Sage.

Hsee, Christopher K., Yang Yang, Naihe Li, Luxi Shen. 2009. Wealth, Warmth, and Well-being: Whether Happiness is Relative or Absolute Depends on Whether It Is about Money, Acquisition, or Consumption. Journal of Marketing Research, 46 (June): 396-409.

Inglehart, Ronald. 2000. Globalization and Postmodern Values. Washington Quarterly, 23 (Winter): 215-228.

Inoguchi, Takashi and Doh C. Shin. 2009. The Quality of Life in Confucian Asia: From Physical Welfare to Subjective Wellbeing. Social Indicators Research, 92:183-190.

Martin, Kelly D., John B. Cullen, Jean L. Johnson, and K. Praveen Parboteeah. 2007. Deciding to Bribe: A Cross-Level Analysis of Firm and Home Country Influences on Bribery Activity. Academy of Management Journal, 50 (December): 1401-1422.

Ozanne, Julie, Ronald Paul Hill, and Newell Wright. 1998. Juvenile Delinquents' Use of Consumption as Cultural Resistance: Implications for Juvenile Reform Programs and Public Policy. Journal of Public Policy & Marketing, 17 (Fall): 185-196.

Raudenbush, Stephen W. and Anthony S. Bryk. 2001. Hierarchical Linear Models: Applications and Data Analysis Methods, 2nd edition. Newbury Park, CA: Sage.

Rindfleisch, Aric, James E. Burroughs, and Nancy Wong. 2009. The Safety of Objects: Materialism, Existential Insecurity, and Brand Connection. Journal of Consumer Research, 36 (June): 1-16.

Rucker, Derek D. and Adam D. Galinsky. 2008. Desire to Acquire: Powerlessness and Compensatory Consumption. Journal of Consumer Research, 35 (August): 257-267.

Runciman, William G. 1966. Relative Deprivation and Social Justice. London: Routledge & Kegan Paul.

Schaninger, Charles M. 1981. Social Class versus Income Revisited: An Empirical Investigation. Journal of Marketing Research, 18 (May): 192-208.

Schimmack, Ulrich, Shigehiro Oishi, Phanikiran Radhakrishnan, Vivian Dzokoto, and Stephan Ahadi. 2002. Culture, Personality, and Subjective Well-being: Integrating Process Models of Life Satisfaction. Journal of Personality and Social Psychology, 82 (April): 582-593.

Schwartz, Shalom H. 1992. Universals in the Content and Structure of Values. In Advances in Experimental Social Psychology, edited by Mark Zanna (1-65). New York: Academic Press.

Shin, Doh C. 1980. Does Rapid Economic Growth Improve the Human Lot? Some Empirical Evidence. Social Indicators Research, 8 (June): 199-221.

Smith, Peter B. 2006. When Elephants Fight, the Grass Gets Trampled: The GLOBE and Hofstede Projects. Journal of International Business Studies, 37:915-921.

Steenkamp, Jan-Benedict E.M. and Hans Baumgartner. 1998. Assessing Measurement Invariance in Cross-National Consumer Research. Journal of Consumer Research, 25 (June): 78-90.

Steward, Thomas Quincy. 2006. Reinvigorating Relative Deprivation: A New Measure for a Classic Concept. Social Science Research, 35: 779-802.

Suh, Eunkook M. 2002. Culture, Identity Consistency, and Subjective Well-being. Journal of Personality and Social Psychology, 83 (December): 1378-1391.

Suh, Eunkook M., Ed. Diener, Shigehiro Oishi, and Harry C. Triandis. 1998. The Shifting Basis of Life Satisfaction Judgments across Cultures: Emotions versus Norms. Journal of Personality and Social Psychology, 74 (February): 482-493.

Trompenaars, Fons, and Charles Hampden-Turner. 1998. Riding the Waves of Culture: Understanding Cultural Diversity, in Global Business. New York: McGraw Hill.

United Nations. 2007. Human Development Report 2007: Fighting Climate Change. New York: Palgrave Macmillan.

van de Vijver, Fons, and Kwok Leung. 1997. Methods and Data Analysis of Comparative Research. In Handbook of Cross-Cultural Psychology, Volume 1: Theory and Method, edited by John W. Berry, Ype H. Poortinga, and Janak Pandey (257-300). Needham Heights, MA: Allyn and Bacon.

Veblen, Thorsten. 1934. The Theory of the Leisure Class: An Economic Study of Institutions. New York: The Modern Library.

Viswanathan, Madhubalan, Jose Antonio Rosa, and Julie A. Ruth. 2010. Exchanges in Marketing Systems: The Case of Subsistence Consumer-Merchants in Chennai, India. Journal of Marketing, 74 (May): 1-17.

Wood, Stacey L. and R. Bettman. 2007. Predicting Happiness: How Normative Feeling Rules Influence (and even Reverse) Durability Bias. Journal of Consumer Psychology, 17:188-201.

World Values Study Group. 2005. World Values Survey, 2005-2006. Inter-University Consortium for Political and Social Research: University of Michigan.

Zagorski, Krzysztof, Jonathan Kelley, and M.D.R. Evans. 2007. Economic Development and Happiness: Evidence from 32 Nations. Section on Economic Sociology Roundtable, Annual Meeting of the American Sociological Association, New York.

Zhong, Jing Yang and Vincent-Wayne Mitchell. 2010. A Mechanism Model of the Effect of Hedonic Product Consumption on Well-being. Journal of Consumer Psychology, 20 (April): 152-162.

(1.) As a check on our data, we performed all consumption analyses using the basic survival category as the consumption moderator in place of the higher-order fulfillment category. We found similar, yet reversed, results, which are available upon request.

Ronald Paul Hill ([e-mail]) is Richard J. and Barbara Naclerio Chairholder in Business and Professor of Marketing and Business Law at the Villanova School of Business, and Kelly D. Martin ([e-mail]) is Assistant Professor of Marketing in the College of Business at Colorado State University in Fort Collins.

TABLE 1
Descriptive Statistics: Correlations, Means and Standard Deviations

Study Variables              Mean     SD

1. Well-being                9.66     2.92
2. Consumption higher order  0.39     0.10
3. Relative restriction      0.40     0.09
4. Absolute restriction      0.23     0.16
5. Individual/collectivism   4.27     0.44
6. Power distance            5.16     0.40
7. Income                    4.06     2.60
8. Social class              2.14     1.36
9. Age                       41.10    16.66
10. Gender                   0.48     0.50
11. Age complete education   16.02    9.36

Study Variables              1        2

1. Well-being
2. Consumption higher order  0.147
3. Relative restriction      0.120    -0.403
4. Absolute restriction      -0.159   -0.708
5. Individual/collectivism   0.028    0.296
6. Power distance            -0.105   0.032
7. Income                    0.156    0.075
8. Social class              0.068    0.021
9. Age                       -0.065   0.142
10. Gender                   0.008    -0.012
11. Age complete education   0.126    0.178

Study Variables              3        4

1. Well-being
2. Consumption higher order
3. Relative restriction
4. Absolute restriction      0.448
5. Individual/collectivism   -0.318   -0.172
6. Power distance            -0.227   -0.098
7. Income                    -0.113   -0.034
8. Social class              -0.225   -0.012
9. Age                       -0.173   -0.198
10. Gender                   0.016    0.017
11. Age complete education   -0.093   -0.262

Study Variables              5        6

1. Well-being
2. Consumption higher order
3. Relative restriction
4. Absolute restriction
5. Individual/collectivism
6. Power distance            -0.270
7. Income                    0.120    0.009
8. Social class              0.096    -0.099
9. Age                       0.020    0.033
10. Gender                   0.005    0.007
11. Age complete education   0.094    -0.071

Study Variables              7        8

1. Well-being
2. Consumption higher order
3. Relative restriction
4. Absolute restriction
5. Individual/collectivism
6. Power distance
7. Income
8. Social class              0.272
9. Age                       -0.052   0.010
10. Gender                   0.047    0.008
11. Age complete education   0.198    0.139

Study Variables              9            10

1. Well-being
2. Consumption higher order
3. Relative restriction
4. Absolute restriction
5. Individual/collectivism
6. Power distance
7. Income
8. Social class
9. Age
10. Gender                   0.005
11. Age complete education   -0.119    0.057

Notes: N = 56,261, level 1; n = 38, level 2. Level 2 countries
include Argentina, Australia, Brazil, Bulgaria, Canada, Chile,
China, Colombia, Egypt, Finland, France, Germany, India, Indonesia,
Italy, Japan, Jordan, Malaysia, Mexico, Morocco, Netherlands, Norway,
Peru, Poland, Romania, Russia, Slovenia, South Africa, South Korea,
Spain, Sweden, Switzerland, Thailand, Turkey, United Kingdom,
Ukraine, United States and Vietnam. Income and social class are
categorical variables; gender (male = 1) is dummy coded. Correlations
computed by attaching country-level variables to each individual-level
variable and counterweighting by country sample size to adjust for
countries with greater representation. Correlations .010 or greater
are significant at p < .05; correlations .012 or greater are
significant at p < .01.

TABLE 2
Collinearity Diagnostics for Country-Level Variables

                                     Model             Effects
                                     [gamma]           SE        t-value

Country-level predictors
  Consumption: higher-order goods    0.339             1.80      1.94*
  Relative restriction               0.590             1.46      4.22***
  Absolute restriction               -0.455            1.68      -2.71***
Country-level controls
  Individualism/Collectivism         0.108             0.14      0.79
  Power distance                     -0.235            0.12      -1.91*
                                     [R.sup.2] = .60

                                     Collinearity Statistics
                                     Tolerance         VIF

Country-level predictors
  Consumption: higher-order goods    0.41              2.46
  Relative restriction               0.63              1.58
  Absolute restriction               0.44              2.27
Country-level controls
  Individualism/Collectivism         0.67              1.50
  Power distance                     0.82              1.22

Notes: Analysis computed using ordinary least squares regression
(n = 38) by collapsing individual well-being values to the country
mean values for each of the thirty-eight countries. Parameter values
reflect standardized estimates.

*p < .05, **p < .01, ***p < .001.

TABLE 3
The Role of Country-Level Restriction and Consumption on
Individual Well-being

                                      Main Effects

                               [gamma]    SE     t-value

Country-level predictors
  Consumption: higher-order    0.101      0.05   2.23*
    goods
  Relative restriction         0.197      0.06   3.41**
  Absolute restriction         -0.119     0.03   -3.67***
  Consumption *
Relative
    restriction (HI)
  Consumption *
Absolute
    restriction (H2)
  Absolute * Relative
    restriction (H3)
Country-level controls
  Individualism/Collectivism   0.031      0.04   0.84
  Power distance               -0.061     0.04   -1.65
Individual-level controls
  Income                       0.256      0.04   7.17***
  Social class                 0.196      0.03   6.82***
  Age                          -0.077     0.02   -3.26**
  Gender                       -0.051     0.02   2.41 *
  Education                    0.041      0.01   2.91**
  Level 1                      [R.sup.2] = .31
  Level 2                      [R.sup.2] = .46

                                   Hypothesized Model

                               [gamma]    SE     t-value

Country-level predictors
  Consumption: higher-order    0.050      0.05   1.00
    goods
  Relative restriction         0.236      0.07   3.40**
  Absolute restriction         -0.151     0.04   -4.18***
  Consumption *
Relative                       -0.035     0.04   -0.87
    restriction (HI)
  Consumption *
Absolute                       -0.118     0.04   -2.99**
    restriction (H2)
  Absolute * Relative          0.086      0.04   2.08*
    restriction (H3)
Country-level controls
  Individualism/Collectivism   0.006      0.04   0.14
  Power distance               -0.016     0.03   -0.51
Individual-level controls
  Income                       0.255      0.04   7.06***
  Social class                 0.193      0.03   6.50***
  Age                          -0.077     0.02   -3.24**
  Gender                       -0.052     0.02   2.43*
  Education                    0.042      0.01   2.96**
  Level 1                      [R.sup.2] = 0.29
  Level 2                      [R.sup.2] = 0.51

Hierarchical linear model: level 1 = 56,261; level 2 = 38.
Parameter values reflect standardized estimates. Where greater
values indicate stronger individualist cultural values.

*p < .05, **p < .01, ***p < .001.