Consumption attitudes and behaviors in Asia: a “discovery-oriented” fresh look

Abstract

Purpose

This paper aims to identify distinct consumption patterns among Asian consumers and examine how these relate to cultural antecedents and key human values.

Design/methodology/approach

This study uses a large, representative sample of almost 7,000 Asian consumers in 10 culturally varying markets, using latent profile analysis (LPA) to identify the consumption profiles.

Findings

The findings empirically demonstrate that the two profiles are “inner-directed nationalistic frugals” (IDNF) and “outer-directed self-eekers” (ODSS). IDNF consumers spend more time and money on education and prefer ethnocentric consumption. ODSS consumers emphasize individuality, self-expression, seeking novelty and impressing others. Consumers with more collective values in Schwartz's typology tend to demonstrate the IDNF pattern; those with more individualistic values demonstrate the ODSS pattern. The distribution of IDNF and ODSS profiles is influenced by demographics, religion and geographical region: IDNF is greater than ODSS in Southeast Asia; ODSS is greater than IDNF in East Asia; IDNF is roughly equal to ODSS in Northeast Asia. IDNF tends to be found among older and more religious consumers, while ODSS is the opposite. Importantly, in the more religious Southeast Asian countries, even younger consumers are more IDNF than ODSS.

Research limitations/implications

This research uses an exploratory and discovery-oriented approach; future research can use more confirmatory approaches to systematically examine the relationship between cultural dimensions (e.g. individualism-collectivism) and consumption patterns.

Practical implications

For their brands to grow in Asian markets, marketing practitioners are advised to use multiple brands to segment Asian consumers based on their values, demographics, geographical location and what religious/faith traditions they follow.

Originality/value

This is the first paper to identify consumption profiles in Asian markets using LPA without prior conceptual biases and relate them to cultural values and demographic variables.

Citation

Batra, R., Li, D. and Chiu, C.-Y. (2022), "Consumption attitudes and behaviors in Asia: a “discovery-oriented” fresh look", Asia Pacific Journal of Marketing and Logistics, Vol. 34 No. 3, pp. 559-575. https://doi.org/10.1108/APJML-12-2020-0878

Introduction

Given the rapid economic growth in most Asian countries, marketing academics and practitioners seek to better understand the changing nature of consumer demand among Asian consumers. Existing research of this type has tended to examine small consumer samples in single countries. For example, Zakaria et al. (2021) studied Gen Y consumers in Malaysia; Han et al. (2021) compared older vs. younger consumers in China; Gupta (2011) contrasted age groups in India. We address this need on a much larger scale here by examining the nature of multiple distinct consumption patterns in a pooled representative field data sample of almost 7,000 consumers from 10 Asian countries. We then relate these distinct consumption patterns to cultural, economic and demographic correlations relevant to the Asian marketplace. We end by drawing important theoretical and managerial implications from our findings on how consumer attitudes and behaviors are changing in Asia.

Literature review and theoretical framework

The conceptual domains we cover align with variables that have previously been of interest to scholarly research on consumer values, attitudes, and behaviors in Asia. These include status-oriented, face building and “conspicuous” consumption (Wong and Ahuvia, 1998); the effects of power distance/hierarchy and masculinity (e.g. Hofstede, 2011) and “collective materialism” (Awanis et al., 2017) in which materialistic values serve the status needs of collective groups. Asian consumers have also been associated with other distinctive “values” (described as “guiding principles in life” by Schwartz, 1992), including religious/traditional (including Confucian) rather than secular values (Chen et al., 2019), as well as thriftiness (Barr, 2000; Kim et al., 1999). The table in Appendix 1 of the supplementary materials structures this literature with references.

It would, thus, seem natural to study today's Asian consumers using existing conceptualizations, scales and relationships for these constructs in the usual “hypothesis testing” manner. Recently, however, researchers have been asked to examine these critically – take a “fresh look” – in contexts different from the Western ones of prior research (Dubois and Ordabayeva, 2015). For instance, it has been argued that materialism may manifest itself differently and serve different end goals in Asian versus Western cultures, given the greater importance of religion and collective groups in most of Asia (Kassim et al., 2016; Awanis et al., 2017). In addition, Asian consumers have undergone significant changes in their values over the last few decades (Gupta, 2011; Han, 2017; Han et al., 2021), becoming more individualistic and cosmopolitan and less ethnocentric. These changes have been driven by major increases in industrialization and urbanization (Wu et al., 2019) and amplified by greater exposure to global consumer culture (Alden et al., 2006; Gupta, 2011). Such changes create the potential for the conceptual and nomological inapplicability of established constructs and relationships. For instance, what may be changing are not just the levels of individualistic (versus collective) values but also the very nature of these constructs (Vignoles et al., 2016).

Present approach and research propositions. There is a strong need, therefore, to study the nature of Asian consumption-relevant values and attitudes “afresh” unencumbered by prior formulations “standard measures” and assumed relationships. Consequently, we conduct this investigation in the spirit of what Wells (1993) called “discovery-oriented consumer research.” When the forces that shape consumers change in a major way, he writes, consumer researchers need to check afresh whether prior theories (therefore also their constructs and measures) still adequately capture real-world phenomena, instead of simply continuing to do hypotheses tests of abstract constructs developed for a different, earlier environment. There have also been calls to see if new cultural variables, not included in Hofstede's (2011) typology, matter today (Leong et al., 2008). Since our approach here is discovery-oriented, we will not offer the usual hypotheses to test. Instead, we offer these two research propositions:

Research proposition 1: In Asia, there are important consumption-relevant values, attitudes and behaviors that have not received adequate attention in prior scholarly research. In particular, we anticipate that consumers' religions/faiths deserve more attention in Asia for their “downstream” impact on consumption. Unlike in the mostly secular West, religious affiliations play a much larger role in large parts of Asia. Indonesia and Malaysia are heavily Islamic; India is mostly Hindu; Thailand and Japan have significant Buddhist influence; and China, S. Korea, Japan, Singapore and Hong Kong share a strong Confucian heritage (which has near-religious influence: Cohen et al., 2016). We expect to find significant associations of these religious differences with consumption patterns through their accompanied values. Importantly, by examining religious influences jointly with others, we will place them in context and examine mediators and moderators (Kassim et al., 2016, this journal).

Research proposition 2: We expect to find significant changes occurring across Asia in the levels and importance of key consumption-driving values manifest in differences across age cohorts. Earlier global research (Inglehart and Baker, 2000) leads us to expect that, given the rapid economic development in Asia, younger age cohorts today experience less economic and psychological insecurity during their formative years than their parents did. This should change their consumption-relevant attitudes, as should their greater contact with global consumer culture (Alden et al., 2006; Gupta, 2011; Han, 2017). However, societal, cultural and religious values should still moderate and constrain these generational changes in important ways.

Contributions. The first contribution to the “discovery-oriented” approach of the present investigation, therefore, is to establish the impact of the different religions/faiths discussed above on Asian consumption-relevant values. Prior cross-cultural investigations (Oyserman et al., 2002; Cohen et al., 2016; Hofstede, 2011) have either not studied the effects of these religious/faith differences or have not examined these simultaneously with other cultural and economic forces to identify interrelationships (Kassim et al., 2016). We also show the mediation of religious/faith effects on downstream attitudes/behaviors via the life values dimensions of (Schwartz, 1992; Schwartz et al., 2012), in particular “openness versus conservation” linked to types of religious affiliation. Thus, responding to the need raised by Leong et al. (2008), we identify openness-conservation as a construct especially important in Asia that has been neglected in prior cross-cultural marketing research.

Our second contribution is to show how Asian age cohorts differ in the relative importance they place on particular consumption-related values and attitudes, relating these to broader economic and sociocultural influences that moderate these differences. Despite earlier research on how individual/national cultural values change with economic development (Inglehart and Baker, 2000), there has been insufficient study of how younger age cohorts in Asia overall differ from older ones, across a broad array of consumption values and attitudes; for what reasons; and with what implications. Extent research is limited to single-country studies (e.g. Han et al., 2021; Zakaria et al., 2021; this journal) and/or examines just a limited range of constructs (e.g. Han, 2017 examines changing levels of cosmopolitanism and ethnocentrism in Japan and S. Korea).

This study also contributes in a methodological manner. It is unusually large and representative in its scope: we utilize a rare field dataset of almost 7,000 Asian consumers from 10 culturally varying countries, collected in a nationally representative manner. In approach, it differs from conventional work in its use of a fresh “discovery-oriented” lens (Wells, 1993), to see what “distinct patterns of consumption” emerge inductively, in an etic not emic manner. The analytical technique statistically identifies different consumption “profiles” among Asian consumers through a latent profile analysis (LPA) of the data. LPA is especially well suited to the detection of heterogeneous patterns of relationships (Stanley et al., 2017), allowing us to relate these to demographic, geographic and cultural (especially religious) correlations. These methodological choices allow us to derive valuable implications for both theory and practice. We identify constructs and processes neglected in prior research and show how international marketing practitioners can build stronger brand portfolios in Asia by broadening the appeal of their brands beyond the urban, educated elite.

Methodology

Variables

A complete list of the variables was used in our analysis, showing the specific items used; it can be found in Appendix 2 in the supplementary materials.

Consumer attitudes and behaviors

The variables that we examine include aspects such as novelty seeking, buying things to impress others, frugal vs. self-indulgent and luxury buying behaviors, preferring locally made over foreign-made products, learning and using technology in consumption, etc. We drew on measures from the literature on materialism (Richins, 2004), impression management (Fennis and Pruyn, 2007) and independent versus interdependent (collective) self-construal (Singelis, 1994), as well as some commercial market research studies.

Human values and Asian values

We also include variables pertaining to underlying human values from two major literature streams.

The first stream is the influential work of Schwartz and colleagues (Schwartz, 1992; Schwartz et al., 2012). Schwartz's circumplex structure includes four basic values: self-enhancement, self-transcendence, openness to change and conservation. We use these four Schwartz values groupings to examine their role in differentiating the consumption attitudes of different cultural (including religious) groups. The measures we use for Schwartz values are those used by the world values survey (Inglehart et al., 2014), which are shorter but still yield the structure that emerges from Schwartz’s own items (Held et al., 2009).

Second, we draw on the literature on the types of “Asian values” that are given higher weight among people living and socialized in Asia (e.g. Kim et al., 1999; Barr, 2000). Typically, such values include humility, conformity to norms, deference to authority and hierarchy, emotional self-control and self-discipline, thriftiness, the pursuit of achievement through hard work and education, collectivism and loyalty to the family, and the forgoing of personal freedoms for the benefit of one's social ingroups and religion, duty, and filial piety.

This literature on Asian values also points out that there are substantial variations within Asia. For instance, those in “Confucian societies” such as those found in China (including its Hong Kong region), Singapore, S. Korea, Japan naturally differ from those that are Islamic (Indonesia, Malaysia), Catholic (Philippines), Hindu (India) or Buddhist (Thailand, Japan). The relative salience of these values could well differ across levels of education, income and exposure to “Western” cultural influences (Alden et al., 2006). Thus, we will treat such variables as possible antecedents or correlates of the values variables in our data.

Survey measures

The survey questionnaire was administered in 13 regional languages (with careful back-translation checks) and English. The opening screener section captured country/region, gender, age, education and monthly household income or socioeconomic status (SES). At the end of the survey, data were obtained on employment and marital status, and religion.

The main sections of the questionnaire included five types of questions:

1. Items related to consumption behaviors and attitudes (see Appendix 2). Respondents indicated the extent to which they agreed with the statements (1 = disagree strongly, 4 = agree strongly). Prior research has shown the adequacy of four-point scales for the kind of multivariate LPA we will be conducting (Olsson, 1979).

2. Items (taken from the World Values Survey: Inglehart et al., 2014) demonstrated to be adequately similar (Held et al., 2009) to those used by Schwartz (1992) to measure openness, self-transcendence, conservation and self-enhancement values, which we tested as potential mediators of demographic effects. Respondents were shown a description of a person and asked if the person described was like them or not (1 = not at all like me, 4 = very much like me) on these values.

3. Because of their acknowledged importance in Asian culture (Kim et al., 1999), we asked respondents to tell us how important traditions and religion were to them. We measured this importance on 4-point scales (1 = not at all important, 4 = very important) using four items (α = 0.73): “following religion and tradition,” “tradition is important to this person; “following the customs handed down by one's religion or family” and “religious faith.”

4. Since consumption behaviors have been linked in prior literature to self-perceptions of economic and social advancement (Batra et al., 2000), we added a non-demographic control variable for the individual level of financial optimism. We used two items (1 = disagree strongly, 4 = agree strongly; alpha = 0.65): “My standard of living has been increasing quite a bit in recent years,” and “Five years from now, my family income will probably be a lot higher than it is now.”

Sample

The data were collected by a professional global market research firm from ten Asian countries/regions: India, (Mainland P.R.) China, Indonesia, Japan, Thailand, S. Korea, Malaysia, Singapore, Philippines and the Hong Kong SAR of China. These 10 countries include the five major economies in East and South Asia (China, India, Japan, South Korea and Indonesia), with substantial variation in the level of economic and social development across the set (2017 GDP-PPP per capita ranged from US$ 7,183 in India to US$ 93,905 in Singapore) [1]. The total sample size was 6,873 (see Table 1). The age range was 18–64 years (mean = 37.6); this age range covers over 70% of the population in the region. Sample quotas were established for each country to match census proportions on gender, age within 18–64 and socioeconomic status (SES) or income, thus making the percentage distributions across our LPA profiles projectable to the overall populations of these countries. In the four countries where internet penetration was greater than 70% (Hong Kong, S. Korea, Japan and Singapore), we used online panels to collect data, using quota sampling to ensure coverage from both large and small cities to match census proportions. In the six other countries, we used face-to-face interviews, typically utilizing random sampling (sometimes stratified, applying proportions based on population density). Again in each case, we selected respondents in both large and small cities to make the data as representative within-country and comparable across countries, as was practically feasible. For example, in Indonesia, we sampled from Jakarta, Surabaya, Medan and Yogyakarta; in Malaysia, from KL/PJ, Ipoh, Penang and Johor Bahru; seven cities in Japan, eight in India, six in China, etc. Within each city, rigorous sampling techniques were used to select city streets, households per street and respondents within a household. Of our respondents, 50% were male; 65.5% were married; 34.6% had high or very high education, and 46.1% were employed full time (see Appendix 3 in the supplementary materials).

Analysis technique

Latent profile analysis

Latent profile analysis (LPA), like latent class analysis (LCA), is a respondent-centered statistical technique used to probabilistically identify unobservable subgroups (latent “classes” or segments) of consumers that display differential relationships to attitudinal or behavioral variables of interest (Wedel and DeSarbo, 1994 provide an early review). The mathematical model used is given in Appendix 4. LPA is used when the observed variables are continuous (our case) rather than categorical. Unlike conventional factor analysis, LPA shows how the patterns of interrelationships (“profiles”) across the variables are a function of the unobserved heterogeneity in the population distribution. Compared to traditional cluster analysis, LPA allows researchers to make more systematic decisions about the number of latent profiles using formal criteria (Nylund et al., 2007). It also possesses methodological advantages over moderated regression or median-split analyses (Stanley et al., 2017). For all these reasons, we use LPA here to explore how multiple, distinct, data-based latent consumption profiles may exist among Asian consumers today. There are only a small number of published LPA applications in marketing (most use LCP); some are described in Appendix 4.

Analysis and results

We conducted the following analyses: (1) extracting the patterns (latent profiles) of consumption in our pooled 10-country data and (2) examining the influence on these profiles of our demographic variables, individual-level Schwartz values and the importance of religion and tradition in the respondent's life. Before conducting these analyses, since respondent usage of the measurement items might differ across consumers in the 10 countries, we converted the raw responses into more comparable data by standardizing each respondent's data within-respondent. Such within-country or within-respondent standardization (ipsatization) makes responses on measurement items amenable for cross-country analysis (Fischer, 2004). See Appendix 2 in the supplementary materials for how this addresses the issues discussed by Baumgartner and Steenkamp (2001).

Results

Profiles: two patterns of consumption

We ran a latent profile analysis (LPA) on the consumer attitudes and behaviors shown in Table 1 (the complete wording of these items is in Appendix 2 in the supplementary material) using Mplus 7.40. We ran LPA models with the number of profiles set from two to four. To determine the optimal number of latent profiles, we used two criteria: better statistical fit and greater substantive relevance to the questions of interest (Nylund et al., 2007). Regarding superior statistical fit, smaller values of the Akaike information criterion (AIC), the Bayesian information criterion (BIC) and the sample-size-adjusted Bayesian information criterion (SSA-BIC) indicate that more variance of the model relative to the number of parameters employed is being accounted for, and higher entropy suggests improvement in prediction (Vermunt and Magidson, 2002). Lo et al. (2001) test a bootstrap likelihood ratio test (BLRT) to evaluate a model with k profiles fitting the data better than a k−1 profiles model. As shown in Table 2, we examined solutions with 2, 3 or 4 profiles based on their fit indices (i.e. relatively lower AIC, BIC, SSA-BIC, a higher entropy, and significant LMR and BLRT). It might be expected that given the heterogeneity across Asia discussed earlier, the solution with 3 or 4 profiles would be needed to adequately capture the important variation in the data. However, a detailed comparison of the 2 versus 3 and 4 profile solutions (see Appendix 5 in the supplementary material) showed that the addition of the third or fourth profile would not add much to our substantive understanding and that the third and fourth profile solutions were questionable on both statistical and substantive grounds. Thus, given the importance of parsimony, we performed our remaining analyses using the 2-profile LPA solution.

In this two-profile solution, profile 1 consisted of 49.6%, and profile 2 consisted of 50.4% of the sample. We characterized these consumption profiles by examining the average scores across profiles of their ratings on our consumption measures. Since these scores were standardized within respondents, scores greater than zero indicate that respondents agreed with these consumption measures, while scores less than zero indicate that they disagreed with these consumption behaviors. The t-tests in Table 2 show that except for the measure “learn technology” (p = 0.92), the rating differences for all other measures across profiles 1 and 2 are statistically significant.

As seen from Table 2 and Figure 1, the two profiles do have some shared features: high agreement in both profiles with get more education, cautious spending, pay cash, value for the money, save money, health, prosocial, make good deals and have high quality goods. One interpretation is that these are quintessential and unchanging Asian consumption attitudes and behaviors, cutting across consumption subgroups, as discussed in the previously cited literature on Asian values (Barr, 2000; Kim et al., 1999).

However, the two profiles also have distinctive characteristics. First, consumers in profile 1 strongly agree that spending time and money on education is important (M = 0.58) and strongly disagree that money should be spent on non-practical things (M = −1.05). They are also much more frugal (high means for cautious pending, pay cash, value for price and save money). Second, they value locally made products and are against foreign products, demonstrating a tendency toward ethnocentric consumption. In contrast, consumers in the 2nd profile put relatively more emphasis on individual expression, treating oneself with buying, seeking novelty, impressing others and being involved with celebrity culture. They are also more favorable toward foreign products and international brands and less committed to buying local products.

Almost 70 years ago, in 1950, sociologists David Reisman, Nathan Glazer and Reuel Denny wrote a classic, highly influential book (updated edition: Riesman et al., 2001) on the changing American social character, documenting a shift in the way Americans sought to live, moving away from internalized behaviors learned from parents and elders (“inner direction”) toward emulating and seeking to impress “peer groups” (“outer- directed”). Given the rating differences above, we label the 1st profile as “inner-directed nationalistic Frugals” (IDNF); they clearly seem more “inner-directed,” in the classic typology of Reisman et al. (2001), than profile 2, which is apparently more “outer-directed”). We, therefore, label the 2nd profile as “outer-directed self-seekers” (ODSS).

Baseline influence of demographics, religions and tradition on consumption profiles

Using these variables as predictors, we estimated an ordinary least squares regression model to predict the probability that each individual belonged to the 1st IDNF profile (this probability was obtained from the LPA solution) as the dependent variable. As the probabilities of the two types of consumption profiles are exclusive and complementary, the estimates of coefficients of independent variables for profile 1 would be the same with those for profile 2 but in opposite valence.

In model 1, we examined the influence of demographics, religious affiliation and (importance of) tradition on the probability of belonging to IDNF profile 1. The following demographics were entered as independent variables: gender, age, education, employment, chief earner, income, marital status. Financial optimism was also entered as a control variable; another was Internet usage (hours per day). The five types of religions were entered as independent variables (via dummies): Hinduism, Islam, Catholicism, Christianity (other than Catholicism: there are many Evangelicals and other Protestants in Korea and Japan in particular), Buddhism and “no religion,” with the last becoming our base level in the analyses. Since one can be traditional without necessarily being religious, and to help calibrate the degree of religious intensity, we also included the interaction terms of tradition with each type of religion.

The results (see model 1 in Appendix 6 in the supplementary materials) show that age (b = 0.12), education (b = 0.05), Christianity (b = 0.04), Buddhism (b = 0.15), Catholicism (b = 0.11) and tradition (b = 0.07) were positively correlated with IDNF profile 1, while income (b = −0.09) and Hinduism (b = −0.13) were negatively correlated with profile 1. These results suggest that people who are older, more educated, have lower income, more traditional and believe in Christianity, Buddhism or Catholicism are more likely to be in profile 1, opposite of profile 2. However, consumers who believe in Hinduism tend to be in ODSS profile 2. Moreover, the interaction terms of tradition and Hinduism (b = 0.06), Islam (b = 0.22), Buddhism (b = 0.09) and Catholicism (b = 0.03) were also significantly correlated with profile 1. That is, people who are traditional and also believe in Hinduism, Islam, Buddhism or Catholicism are more likely to be in profile 1, which is the opposite of profile 2.

Relations with Schwartz values

Next, we additionally tested whether the Schwartz values affect the probability of belonging to a particular consumption profile beyond the variables just discussed for regression model 1. The analysis (see model 2 in Appendix 6 in the supplementary materials) shows that values of openness (b = −0.07) and self-enhancement (b = −0.13) were negatively correlated with IDNF profile 1, while conservation (b = 0.18) and self-transcendence (b = 0.12) were positively correlated with IDNF profile 1. The Schwartz values collectively improve the model fit significantly: the adjusted R² increases from 0.145 to 0.200. Previous research shows that conservation and self-transcendence values represent basic requirements of human existence in the pursuit of collective needs of groups, while openness and self-enhancement values represent the motivation to satisfy individualistic needs (Schwartz, 1992). These results suggest that consumers with more collective concerns tend to demonstrate an inner-directed nationalistic frugal consumption pattern (i.e. profile 1), while consumers with more individualistic concerns tend to demonstrate an outer-directed self-seeking consumption pattern (i.e. profile 2).

We also tested the degree to which the Schwartz values added in model 2 collectively mediated the effects of our demographic and other antecedent variables in model 1 in predicting the probability of profile 1. Complete Results are in Appendixes 7–8 in the supplementary materials. To summarize briefly, we found that the demographics-influencing-profile effects were partially mediated by the four Schwartz values we used, and the effects of tradition and Catholicism × tradition on profile 1 were fully mediated by the Schwartz values. The latter effect is not surprising since the Schwartz value-type of conservation covers similar conceptual domains as our separate variable of tradition and is empirically more strongly related to it (r = 0.57) than tradition is to self-transcendence (r = 0.23), self-enhancement (r = 0.05) or openness (r = 0.05).

Geographic, demographic and religious composition of each profile

To gain more insight into the geographic, demographic and religious composition of the consumers falling into each profile, we next analyzed the distribution of profile membership across countries, age groups, age groups within country and religion.

Distribution across countries

The distributions of profiles 1 and 2 in each country are presented in Figure 2. We identified three groupings of countries. The first grouping, in which the percentage of IDNF profile 1 was greater than ODSS profile 2, consists of Singapore, Malaysia, Thailand, Indonesia and the Philippines. The second grouping, in which the percentage of ODSS profile 2 was greater than IDNF Profile 1, consists of Hong Kong, China and India. The third grouping is one in which the percentage of profile 1 almost equals that of profile 2: Japan and S. Korea falls into this group. We note here the influence of geographical region on these groupings. The first covers the area of Southeast Asia, the second covers the area of East Asia, and the third covers the area of Northeast Asia, suggesting the influence of shared religious history across countries in the same geographic region.

Distribution across age groups

Our frequency distribution analysis also indicates that among older generations (40 years and above), the percentage of the IDNF Profile 1 is greater than of the ODSS Profile 2, while among younger generations (18–39 years old), the percentage of Profile 1 is less than Profile 2 (see Figure 3). Thus, subject to the caveat that our data are cross-sectional and not longitudinal, it does appear that the Asian consumers in the older age cohorts are more likely to be in the IDNF Profile 1, while the younger ones are more likely to be in the ODSS Profile 2.

Distributions of age groups within country

We also generated the frequency distribution of the two profiles across age cohorts within each country individually (see Appendix 9 in the supplementary materials and Figure 4). Some very interesting differences are evident. First, while the percentage of younger (18–29) consumers in the ODSS profile 2 is always greater than the ODSS percentages for the two older age groups, the absolute level of this percentage is not uniform. It is highest (67–76%) in China, India and Hong Kong, relatively more moderate (50–60%) in Singapore, Japan and S. Korea and clearly lower in Malaysia, Thailand, Indonesia, and the Philippines (30–46%). Second, the “rate of increase” in this percentage, as we move from the oldest to the youngest age groups, also varies in a very significant manner. This increase is the greatest in S. Korea, from 32.5% in the oldest to 60.6% in the youngest age group, showing a very dramatic change in life and consumption values in that country. In contrast, this rate of change is very slow in Indonesia, from 27.0% in the oldest group to only 29.5% in the youngest group, showing that “traditional/religious” values remain very powerful even among young consumers in that very populous country. (The Philippines also shows a relatively slow rate of change). This suggests that marketers in Asia cannot assume that these countries' populations will all inevitably and quickly fit the ODSS consumption profile. There are clearly large, important regions where more traditional values and consumption patterns will still hold sway for a long time, while other regions are making a much more rapid transition. Managerial implications of this important finding will be discussed below.

Distribution across religion

As depicted in Figure 5, the frequency distribution of consumption profiles by religion showed that among non-religious people, the percentage of ODSS profile 2 is greater than IDNF profile 1 (60.5 vs. 39.5%); similarly, among Hindus, the percentage of ODSS profile 2 is greater than IDNF profile 1 (62.2 vs. 37.8%). However, among other religions (e.g. Christianity, Taoism, Buddhism, Catholicism and Islam), the percentage of ODSS profile 2 is smaller than that of IDNF profile 1 (33–45%, vs. 54–67%).

One interpretation might be that these religions differ in how they view (and teach adherents about) the degree and nature of material consumption that is considered appropriate. Most Eastern/Asian religions emphasize the importance of frugality (e.g. “the person who lives completely free from desires, without longing…attains peace” in Buddhism). Hinduism, however, includes, as one of the four ends or pursuits of life, the pursuit of prosperity and material well-being (“artha”). Therefore, the pursuit of possessions and wealth is arguably more consistent with the Hindu way of life (Tharoor, 2018, p. 36) than it is for many other Asian religions, which may help explain why the level and growth of the profile 2 ODSS percentage are noticeably high in India.

Discussion

Our LPA analysis presents a very interesting snapshot of how Asian consumers today are caught in a world that is transitioning in a non-uniform and non-inevitable manner from one type of consumption profile to another. Given the heterogeneity in the “initial conditions” (in religions, culture, etc.) from which these ten Asian countries began their decades of rapid growth and the differences in these rates of recent growth, it is not surprising that the main contrast is between the IDNF profile 1 and the ODSS profile 2. The former is more driven by tradition and is more “inner-directed,” more nationalistic/ethnocentric and even more careful with their spending than the already high Asian average. The latter is more concerned with creating the right impressions (“outer-directed”), spending more to treat oneself and obtain novel products and experiences, favoring overseas brands. What is interesting from our analyses of this broad, rich field data set of almost 7,000 Asian consumers from 10 countries/regions is the fine-grained look we now get not only at these two profiles but also at what their demographic correlations are; the role of consumer values and religious faiths in shaping them and their geographical densities. These observations yield important implications for international marketing practice and theory.

Implications for theory and future research

In examining the predictive and mediating role of Schwartz values, we found that Asian consumers’ consumption patterns are related not only to the demographic (etc.) variables just discussed but also, quite significantly, associated with their Schwartz values-types. Consumers with collective concerns (i.e. emphasis on self-transcendence and conservation values) tend to be IDNF profile 1 consumers, while consumers with individualistic concerns (i.e. emphasis on openness and self-enhancement values) tend to fall into ODSS profile 2. The fact that the Schwartz measures proved useful predictors and (mostly partial) mediators attests to the explanatory and predictive power in marketing domains of the human values stream of research (e.g. Schwartz, 1992). These findings strongly support our first research proposition.

It is important to note in this context that the Schwartz value-types that added this explanatory variance included not only the self-transcendence and self-enhancement constructs (which should be conceptually related to the much-studied variables of collectivism-individualism and materialism) but also brought in the conceptually different constructs of openness and conservation that have been used much less in the cross-cultural marketing literature. This demonstrated importance of the openness and conservation two values-types clearly suggests that future research ought to investigate their consumption-shaping roles in more depth; these variables have not been studied much in prior cross-cultural research but deserve to be (cf., Leong et al., 2008). Some prior literature (e.g. Alden et al., 2006) has already shown that consumer exposure to global consumer culture, via mass media (e.g. the internet, movies. TV shows, etc.), as well as physical travel overseas and interactions with foreigners at home, increases consumers' openness to global influences and especially non-local brands. These authors did not incorporate into their theorizing and testing the constructs/values of openness and conservation, but our findings suggest that it is time the international marketing and cross-cultural consumer literature do so.

This importance of openness and conservation values is also evident in our finding that Indian-Hindu consumers are much more likely to be ODSS profile-2 consumers than those religious consumers in Indonesia, Thailand and the Philippines who believe instead of Islamic, Buddhist and Catholic faith traditions, respectively. Figure 5, in fact, shows that Hindus (thus, Indians) are even more likely to be ODSS profile 2 consumers than Chinese non-religious consumers. We have presented above some reasons why Hinduism, as a religion, may seemingly be much more accommodative of “materialistic” consumption (Gupta, 2011, p. 252; Tharoor, 2018) than other Asian religions and faiths, but this question is clearly worthy of more study. Related to the point just made, we find importantly that Hinduism seems more connected to Schwartz values of openness (see Appendix 7, b = 0.246) and self-enhancement (see Appendix 8, b = 0.234) than are Islam (b_openess = 0.133; b_{self-enhancement} = 0.155), Christianity (b_openness = 0.058; b_{self-enhancement} = 0.020, n.s.), Buddhism (b_openness = 0.073; b_{self-enhancement} = 0.024, n.s.), and Catholicism (b_openness = 0.099; b_{self-enhancement} = 0.035). Since openness and conservation are opposing values in the Schwarz circumplex values framework (Schwartz, 1992), the strong connection between Hinduism and openness/self-enhancement may be another possible reason why Hindus are more likely to be “outer-directed self-seekers,” while adherents of other religions seem to be more strongly linked to “conservation” values and more likely to fall into the “inner-directed nationistic frugals” consumption profile.

Scholarly research in the West on cross-cultural consumption values rarely gets into differences across religious faiths since most Western societies today are predominantly secular. A big implication of our results for theory-building research is that these religious differences are much more deserving of study in Asia, where they still matter more.

Implications for practice

These relationships discovered via our LPA profiles also have many implications for marketing practitioners seeking to grow in these Asian markets, which are now the main drivers of global market growth in many products and service categories.

Concerning the differences between the two profiles, the biggest differences in means across them come in their relative scores (from Table 2 and Figure 1) in how favorably they view international/global brands; their willingness to spend money on “non-practical” things; their celebrity consciousness; the desire to impress others; the search for novelty; the preference for well-known, trusted and luxury brands (“branding”); and the desire to express their individuality. Profile 2 ODSS consumers score much higher on all of these. IDNF profile 1 consumer in contrast score higher on desiring to spend money cautiously, paying in cash and saving money and getting the most value for the price, wanting to get more education and preferring locally made products instead of ones from abroad.

Clearly, these are rich findings that should be helpful in segmentation and targeting strategizing. Our data allow us to get a good sense of what types of people these two groups are (demographics), where they live and what religious/faith traditions they belong to. Demographically (see Appendix 6, model 1), profile 1 consumers are older in age, supporting our second research proposition. They are also higher in years of education but lower in income. They are more likely to be tradition-following believers in Islam, Buddhism, Catholicism and Hinduism (but not Hindus that are less traditional). Surprisingly, they are not (on average) less world-connected through the internet, nor are they less financially optimistic.

Figure 2 shows us that the more outer-directed, self-seeking Profile 2 consumers are present most (61–67%, as a % of the country) in China and Hong Kong and India (mostly East Asia), a little less so in the North Asian countries S. Korea and Japan (42–43%), and in Singapore and Malaysia (34–39%); and much less in the South-East Asian markets of Thailand, Indonesia and the Philippines (20–29%). The distinctions across these geographical clusters would seem to relate to shared cultural history (e.g. Japan and S. Korea); the prevalence of traditional religions (Thailand, Indonesia, Philippines) or their absence (China/Hong Kong); as well as levels of economic development (Singapore, Malaysia, Japan). As discussed above, India is an interesting case since it is less economically developed and seemingly quite religious (Hinduism), both of which would suggest a high % of profile 1. Yet, it is very high in profile 2 consumers, for possible reasons discussed earlier.

Regardless of their specific geographical location, however (Figure 3), in all these Asian countries considered together, the older you are in age, the more likely you are to be an IDNF profile 1 consumer; the younger you are, the more you fit the ODSS profile 2. This empirical evidence is consistent with the intuitive implication that global marketers of western origin brands should aim their brands and communications at younger cohorts. This result also suggests that as these age cohorts grow in size and proportion, these western-origin brands will inevitably take over most of the Asian marketplace, as traditional religions fade in importance and “conservation” values give way to “openness.”

While this may be true to some extent, our findings suggest caution. Even in younger age groups, the effects of religion/tradition that we saw in Figure 2 (Thailand, Indonesia and the Philippines) are still evidently strong since these countries have the lowest profile 2 % in younger cohorts. Appendix 9 shows this “religion/country effect” still matters in shaping the levels and rates of change for each of our three age cohorts (18–28, 29–39, and 40+). One can plausibly project that younger consumers in these more religious/traditional countries will still likely end up more as IDNF profile-1 than ODSS profile-2 consumers, even as they grow older.

A key managerial implication of the above is that marketers from global brand companies cannot assume that, over time, most or all consumers from all of Asia will become ODSS Profile-2 consumers. More in South-East Asia than elsewhere in the region, Profile-1 IDNF consumption values and preferences will still remain a significant chunk of the market. Thus, global marketers would be well advised to also have some brands that appeal to more traditional/religious consumers in their brand portfolios. Marketers and brand builders should not just be building brands for Asian markets that are anchored in Western and secular/modern values and symbolism. A large opportunity seems to also exist for global brand builders to consider building brands for such markets that symbolize and celebrate traditional values in these markets. This strategy helps to avoid ceding valued symbolic territory to local-origin brands such as Patanjali in India and Shanghai Jahwa and Shanghai Pehchaolin in China. Multinationals operating in Asian markets and seeking to win over not just younger and more elite/Westernized consumers but also “the mass market’s” need to consider building or acquiring some brands that symbolize traditional religious and communitarian values. They need to leverage not only their existing Western-market brands but also their brand-building skills and processes to create such locally relevant brands.

Limitations

Despite these important findings and implications, there are clearly also many limitations to this present study. Data can be analyzed from more countries and cultures. Future data collection can use more accepted, “standard” scales than the measurement items we used, even though most of our items came from prior studies and the World Values Survey. Future research can more systematically examine the role of “old” constructs such as materialism and individualism-collectivism while adding in the new constructs we have advocated related to openness and conservation, using more confirmatory approaches.

Figures

Figure 1

Latent profiles of consumption patterns

Figure 2

Distribution of consumption profiles across countries (in percentage)

Figure 3

Percentage distribution of consumption profiles by age

Figure 4

Percentage distributions of consumption profiles by age and country

Figure 5

Percentage distribution of consumption profiles by religion

Table 1

Fit statistics for consumption profile structures using LPA

No. of profiles	LL	FP	AIC	BIC	SSA-BIC	LMR (p)	BLRT (p)	Entropy	Profile size
2	−185682.322	67	371498.644	371956.613	371743.703	0	0	0.647	3,408, 3,465
3	−185174.374	90	370528.748	371143.93	370857.932	0	0	0.68	3,361, 916, 2,596
4	−184791.182	113	369808.363	370580.759	370221.671	0	0	0.622	1,363, 2,374, 786, 2,350

Note(s): LL = log-likelihood; FP = free parameters; AIC = Akaike information criteria; BIC = Bayesian information criteria; SSA–BIC = sample-size-adjusted BIC; LMR = Lo et al. (2001) test; BLRT = bootstrapped log-likelihood ratio tests

Table 2

Mean scores on consumption measures for the two profile solution

Variables	Profile 1		Profile 2		Rating difference on the measures of profile (1–2)
	Mean	p-value	Mean	p-value	Difference	p-value
get more education	0.579	0.000	0.237	0.000	−0.342	0.000
cautious spending	0.547	0.000	0.050	0.007	−0.497	0.000
pay cash	0.436	0.000	0.132	0.000	−0.304	0.000
value for the price	0.403	0.000	0.182	0.000	−0.221	0.000
saving money	0.364	0.000	0.011	0.527	−0.353	0.000
learning technology	0.203	0.000	0.201	0.000	−0.076	0.922
health	0.203	0.000	0.127	0.000	−0.002	0.001
prosocial	0.186	0.000	0.079	0.000	−0.107	0.000
making good deals	0.133	0.000	0.066	0.000	−0.067	0.009
against foreign products	0.079	0.000	−0.121	0.000	−0.200	0.000
high-quality goods	0.054	0.001	0.183	0.000	0.129	0.000
locally made products	0.026	0.125	−0.454	0.000	−0.480	0.000
express individuality	−0.150	0.000	0.282	0.000	0.432	0.000
treat self with buying	−0.160	0.000	0.172	0.000	0.332	0.000
do no't want to be seen as richer	−0.223	0.000	−0.150	0.000	0.073	0.007
seeking for novelty	−0.368	0.000	0.187	0.000	0.555	0.000
impress others	−0.431	0.000	0.131	0.000	0.562	0.000
branding	−0.445	0.000	0.048	0.000	0.493	0.000
artistic consumption	−0.689	0.000	−0.296	0.000	0.393	0.000
celebrity	−0.900	0.000	−0.244	0.000	0.656	0.000
pro-international brands	−0.970	0.000	−0.094	0.000	0.876	0.000
non-practical things	−1.047	0.000	−0.269	0.000	0.778	0.000

Note

1. Source: International Monetary Fund World Economic Outlook (April−2018), dated 05 Jun 2018.

Supplementary material

The supplementary files are available online for this article.