Generalized trust across six southeast asian societies an analysis of trust foundations

PROCEEDINGS THE 5th INTERNATIONAL CONFERENCE LANGUAGE, SOCIETY, AND CULTURE IN ASIAN CONTEXTS (LSCAC 2018) 25 - 26 May, 2018 Huong Giang hotel, Hue city, Vietnam PROCEEDINGS THE 5th INTERNATIONAL CONFERENCE LANGUAGE, SOCIETY, AND CULTURE IN ASIAN CONTEXTS (LSCAC 2018) Penulis : Woraya Som-In, dkk Editor : Prof Prabhakara Roa, dkk Desain Cover & Penata Isi Tim MNC Publishing Cetakan I, Mei 2019 Diterbitkan oleh : Media Nusa Creative Anggota IKAPI (162/JTI/2015) Bukit Cemara Tidar H5 No 34, Malang Telp : 0812.3334.0088 E-mail : mncpublishing.layout@gmail.com Website : www.mncpublishing.com ISBN 978-602-462-248-0 Hak Cipta dilindungi undang-undang Dilarang memperbanyak atau memindahkan sebagian atau seluruh isi buku ke dalam bentuk apapun, secara elektronis maupun mekanis, termasuk fotokopi, merekam, atau dengan teknik perekaman lainnya, tanpa izin tertulis dari Penerbit Undang-Undang Nomor 19 Tahun 2000 tentang Hak Cipta, Bab XII Ketentuan Pidana, Pasal 72, Ayat (1), (2), dan (6) PROCEEDINGS GENERALIZED TRUST ACROSS SIX SOUTHEAST ASIAN SOCIETIES: AN ANALYSIS OF TRUST FOUNDATIONS An Nguyen Huu Department of Sociology – Hue University of Sciences Email: Annguyenxhh2001@gmail.com Abstract: Generalized trust in people has been widely viewed as the key component of social capital, one which contributes to the smooth operation of a society Recently, scholars have employed a new measure of generalized trust – trust in unknown people – as a viable alternative to previous measures Applying this measure, this paper provides an empirical test of six theories on the origins of generalized trust as suggested by Delhey and Newton (2003) These theories associate generalized trust with personality, success and well-being, voluntary organizations, social networks, community support, and societal conditions The test uses data from the 5th and 6th wave of the World Values Survey and covers six Southeast Asian societies - Vietnam, Thailand, Indonesia, Malaysia, the Philippines, and Singapore Results of an ordered logistic regression analysis reveal voluntary organization theory, network theory, and societal theory are more powerful than community theory and individual theories in predicting generalized trust Keywords: Generalized trust, social capital, personality, well-being, social networks I Introduction In recent decades, the concept of generalized trust has attracted a huge attention from scholars of various social science disciplines Generalized trust is considered as a key component of social capital and conceptualized differently from particularized 296 LSCAC 2018 International Conference trust and reciprocity networks formed with family and friends (Natalia Letki, 2018), the former (generalized trust) is defined as an optimistic attitude in interaction with more or less distant strangers, as an opportunity rather than a threat (Uslaner, 2002) PROCEEDINGS trust based on Fukuyama’s conception of “radius of trust” While the latter refers to There is a wide consensus that generalized trust in others facilitates individual interactions and creates a solid basis for sustained cooperation which contributes to the smooth running of a society (Almakaeva et al, 2014) Generalized trust promotes connectedness, engagement, tolerance and democracy in communities (Stolle and Hooghe, 2004 in Delhey et al, 2011), serving as the basis of reciprocity, social connectedness, peaceful collective action, inclusiveness, tolerance, gender equality, confidence in institutions, and democracy itself (Helliwell and Putnam 2004; Inglehart 1999; Putnam 2001; Uslaner 2002 in Delhey et al, 2011) Putnam (1993) found that among citizens, trusting easily is conducive to healthier societies and more efficient economic growth, making people happier (Resskens, 2013) Owing to the importance of the concept, debates on the factors determining generalized trust have grown the academic literature A pioneering study, conducted by Delhey and Newton (2003), developed six theories of social trust and empirically tested these using data from the Euromodule survey in seven countries, namely East and West Germany, Hungary, Slovenia, Spain, Switzerland and South Korea Both authors found that societal conditions, social networks, and success and well-being are three theories that largely explain the performance of trust Theories of personality, community, and voluntary organisations rather poorly In turn, the correlation between membership in voluntary associations and trust is not high as expected In higher-trust societies, personal success and well-being were found to be the strongest sources of trust, while in the lower-trust societies, either social networks or societal conditions plays an important role in generating trust Demographic variables (gender, age, and education) have relatively weak relationships with trust (Delhey and Newton, 2003) 297 LSCAC 2018 International Conference PROCEEDINGS Scholars have expressed concerns about the validity of measuring generalized trust The conventional measurement of generalized trust drawing on the oft-used question “Generally speaking, would you say that most people can be trusted or that you need to be very careful in dealing with people?” leads to methodological doubt regarding two points: (1) The dissimilar imagination of the notion of “most people” across cultures; and (2) the cautious rather than trusting attitude in the second part of the question Delhey, Newton and Welzel (2011; 2014) have shown that the notion of “most people” was differently imagined among surveyed countries, in which, in some countries, reflects in-group trust rather than out-group trust This results in the question about the reliability of findings of the determinants of generalized trust in previous studies which measured the variable based on the question Accordingly, developing a new and valid measurement of generalized trust is in need Based on the traditional definition considering generalized trust as an attitude towards people in general and the inclusion of two dimensions of trust in unknown people and out-group trust, trust in strangers has been recently proposed as an alternative indicator of generalized trust by scholars (Anna Almakaeva et al, 2014) This paper takes the above methodological consideration to investigate the origins of generalized trust in six Southeast Asian societies - Vietnam, Thailand, Indonesia, Malaysia, the Philippines and Singapore These are developing countries and have different cultural contexts from developed societies that have been the focus in previous research This study wishes to contribute empirical evidence to enrich the theoretical foundation of generalized trust The paper is organized as follows Following this introduction, the next section reviews six theories on the origins of trust A discussion of the data, variables and other methods used in the study comes next, followed by a report on the empirical findings The final section discusses these results and offers conclusions II Six theories of the origins of trust 298 LSCAC 2018 International Conference two broad schools of thought: trust as an individual property and trust as a product of social systems The individual school consists of two approaches, namely (1) personality theory and (2) success and well-being theory In turn, the social school of PROCEEDINGS Delhey and Newton (2003) presented six theories of the origins of trust classified into trust covers four perspectives: (1) Voluntary organization theory, (2) Social network theory, (3) Community theory, and (4) Societal conditions theory Personality theory argues that trust is a psychological trait that individuals learn during the socialization process It is primarily formed in early childhood and slowly changes or even persists in later life According to this perspective, trust should be related to personality characteristics Osakrsson et al (2012), for example, found that positive associations between social trust and extraversion, personal control, and intelligence Similar associations appear in a study by Patrick Sturgis, Sanna Read and Nick Allum (2009) demonstrate generalized trust to be a function of individual intelligence Newton and Delhey (2003) affirm that social trust can be linked with personality variables, especially personal optimism and the sense of control over one’s life However, their test in seven countries (Germany, Hungary, Slovenia, South Korea, Spain, and Switzerland) shows a significantly positive association between generalized trust and optimism as well as a significant but very weak relationship appears between self-control and trust only in Switzerland The success and well-being theory sees trust as a result of positive life experiences For this perspective, generalized trust is expected to be higher among people with more experiences of success and happiness The positive link between generalized trust and subjective well-being has been found in a huge number of studies (Helliwell & Putnam 2004; Hooghe & Vanhoutte 2011; Klein 2013; Leung et al 2011; Yip et al 2007; Allum et al 2010; Elgar et al 2011; Freitag & Buhlmann 2009; Puntscher et al 2014; Rodríguez-Pose & von Berlepsch 2014; Robbins, 2011; Tokuda et al 2010; van der Veld & Saris 2012; Yamaoka 2008 in Almakaeva et al, 2014) However, the relationship is not confirmed in the case of Ghana, a developing country with very high distrust level, in a study carried out by Sulemana (2014) Socioeconomic 299 LSCAC 2018 International Conference PROCEEDINGS variables capturing the level of success and well-being such as education and income have also been employed to unravel the determinants of generalized trust Several studies have found a positive influence of education and income on generalized trust (Brehm and Rahn 1997; Alesina and La Ferrara 2000, Rahn et al., 2009; Sturgis et al., 2009; Park & Subramanian, 2012; Robbins, 2011; Geys, 2012; Paxton, 2007; Delhey & Welzel, 2012; Newton & Zmerli, 2011 in Almakaeva et al, 2014) However, the link is not confirmed in the work of Soroka et al (2007) and in all participating countries of Newton and Delhey’s (2003) study Voluntary organization theory emphasizes the role of participation in voluntary associations in generating trust Theoretically, formal associations are ideal social spaces from which generalized trust emerges through interaction with unknown people and members of out-groups (Putnam, 2001) Nevertheless, the theory has been criticised on both theoretical and empirical grounds (Newton and Delhey, 2003) Empirical tests yield inconsistent results on the effects of organizational membership on generalized trust: the results have been positive impact in some societies but not in others The inconsistencies, it is argued, stem from at least four factors: (1) the application of different types of voluntary association in previous studies; (2) the issue of causality where more evidence support the notion that generalized trust facilitates civic participation rather than the other way around; (3) the underestimated role of non-institutional forms of civic participation; and (4) lack of using moderators at the societal level in the analyses (Almakaeva et al, 2014) As well, Newton and Delhey (2003) also failed to establish the relationship between civic engagement and generalized trust in Slovenia, Spain and South Korea While the voluntary organization theory focuses on active participation in formal association in generating trust, social network theory points to the importance of informal social relations and daily interaction in creating trust According to the theory, family members, friends and neighbours lie at the center of the trusters’ radius of trust There are also ad hoc and sporadic participation of people in the network who, according to Gundelach and Torpe (1996:31) “gather in local bars and pubs, at work, 300 LSCAC 2018 International Conference baby-sitting circles, and local action groups that tend to cluster around schools, community centres, and residential areas” (Gundelach and Torpe 1996: 31; see also Parry et al 1992: 86–7; Foley and Edwards 1997 in Newton and Delhey, 2003) Thus, PROCEEDINGS in book-reading groups and support groups, and among those who form car pools, the informal social network comprises the known people of the trusters In turn, community theory argues that characteristics of local communities explain variations in trust The argument originates in the work of House and Wolf (1978) which show a negative association between town size and trust (Newton and Delhey, 2003) Putnam (2000) reaches a similar conclusion: the smaller the community, the greater the sense of generalized trust The impact of community characteristics on generalized trust has been little taken into account in social capital studies, thus, the evidence of the relationship is still modest (Anna Almakaeva et al, 2014) Newton and Delhey (2003) propose that community size, satisfaction with the community and the sense of community safeness at night should be determinants of trust Taking a broader view, societal theory examines the impact of the whole countries and nation states on trust Some scholars (e.g Newton and Delhey 2003) even contend that this theory reasonably explains why people in wealthier, democratic countries and nations with greater income equality are high trusters as compared to the others Indeed, in their original analysis, Newton and Delhey (2003) posited that societal conditions are among the strongest predictors of generalized trust But it doesn’t appear in all analyzed countries One such social condition is people’s subjective assessment of the intensity of conflict in their respective country Both authors found the expected negative impact of conflict assessment on generalized trust, but only in of the countries included in the study However, this finding is not in line with results from cross-cultural studies which show a positive link between institutional and generalized trust (Allum et al., 2010; Brehm & Rahn, 1997; Freitag, 2003; Freitag & Traunmüller, 2009; Newton & Zmerli, 2011; Robbins, 2011; Robbins, 2012; Rothstein & Stolle, 2008; van Oorschot et al., 2006 in Anna Almakaeva et al, 2014) 301 LSCAC 2018 International Conference PROCEEDINGS In sum, empirical tests of the six theories on the origins of trust yield mixed results This study offers another test, this time by focusing on six societies in Southeast Asia Will the findings be more consistent with a regional focus? Will the predictors of general trust in these six countries parallel those found in Newton and Delhey’s study? What would be the theoretical implications of these results? Before presenting the regional findings, a discussion of the study’s variables and methods is in order III Data, variables and methods 3.1 Data This paper primarily relies on the fifth and sixth wave of WVS survey, a data set which has been widely used in academic studies Six waves were carried out between 1981 to 2014 The fifth wave was conducted in 2005–2009 for 57 countries while the sixth was done from 2010 to 2014 for 60 countries All surveys entailed face-to-face interviews, using a standardized sampling population of adult citizens aged 18 years and over Two reasons justify the use of WVS data sets from the two waves One is that the fifth and sixth wave contain questions measuring trust in strangers as a new indicator of generalized trust The other reason is that these data are the latest waves of the WVS survey and thus offer more current material for testing the theories, Six Southeast Asian countries were involved in WVS for these two phases: Indonesia, Malaysia, Philippines, Singapore, Vietnam, and Thailand Malaysia and Thailand were included in both waves, Indonesia and Vietnam participated in the fifth phase, while the Philippines and Singapore took part only in wave The sample sizes per country and wave are specified in the table below: Table 1: Sample size of the study Countries Wave Indonesia 2,015 Malaysia 1,201 Philippines Singapore Vietnam 1,495 302 LSCAC 2018 International Conference Wave 1,300 1,200 1,972 - Total 2,015 2,501 1,200 1,972 1,495 PROCEEDINGS have shown to be consistent correlates of trust These three variables will thus be included in all regression models used in testing the six theories 3.3 Method Given the ordinal nature of the dependent variable trust in unknown people, this study employs ordered logistic regression to test the six theories on the origins of trust Data analysis will proceed in two steps First, the six theories of generalized trust are tested separately to detect the power of each theory Second, all six theories are included in one model to be tested against each other In both steps, the analysis is conducted country by country Table 2: Generalized trust in six Southeast Asian Countries Trust completely and somewhat (%) Singapore 31.90 Vietnam 22.08 Indonesia 19.52 Philippines 18.60 Thailand 16.12 Malaysia 10.76 Valid N 1,977 2,501 1,200 1,972 1,431 2,717 IV Results Table shows the level of generalized trust across six Southeast Asian countries In general, the majority of individuals in these countries show more distrust than trust towards strangers Singapore has the highest level of generalized trust with 31.9 percent of respondents trusting people they meet for the first time In contrast, Malaysia has the lowest percentage of people (10.76%) reporting a generalized trust strangers (10.76%) Vietnam occupies the second highest position with 22.08 percent followed by Indonesia (19.52%), Philippines (18.6%), and next to last, Thailand (16.12%) 306 LSCAC 2018 International Conference are presented country by country from Table to Table Table shows the regression result when the theories are tested against each other in a single model Gender, age and level of education are control variables included in all models PROCEEDINGS The tests of six theories of the origins on trust across six Southeast Asian societies Table points out that the sole indicator of personality theory, the locus of control, affects levels of generalized trust only in Indonesia and the Philippines, both in the positive direction, It is not statistically significant, in the other four countries However, as Table reveals, the effect of the locus of control on generalized trust disappears when all variables are tested in one model The success and well-being theory is tested on three indicators: subjective well-being, financial household satisfaction and self-rated social class Table indicates that subjective well-being exerts a significant impact on generalized trust only in Indonesia In turn, satisfaction with one’s household finance significantly generates generalized trust in Singapore and Thailand but not in the other four countries Subjective social class, the third measure, shows a positively significant influence on generalized trust in three of the six countries: Indonesia, Vietnam, and Thailand In the full model shown in Table 9, the success and well-being theory well explains generalized trust in Thailand Interestingly, while financial household satisfaction and social class also increase trust in Thailand, subjective well-being decreases the level of trust In contrast, following Table 9, the theory does not receive a statistically significant outcome in the Philippines when combined with the other theories In turn, generalized trust in both Vietnam and Indonesia is evidently enriched by an increase in social class Subjective well-being produces generalized trust in Malaysia, while financial household satisfaction pronounces effect in Singapore It is evident in Table that the active membership in voluntary organizations is a significant source of generalized trust across societies Note, however, the association is positive for all societies except Vietnam where generalized trust is reduced with active membership in voluntary organizations In the full model, shown in Table 9, this factor remains significant in three of the six countries, namely 307 LSCAC 2018 International Conference PROCEEDINGS Indonesia, Singapore and Vietnam, in which the association between active membership and generalized trust still stays negative and statistically significant in the case of Vietnam In turn, network theory greatly explains and highly impacts on generalized trust in all six countries when tested separately (see Table 6) and has the same performance in the full model (see Table 9) The town size variable from community theory also has a significant impact on generalized trust in the Philippines and Vietnam: in the negative direction for the Philippines but in the positive direction for Vietnam (see Tables and 9) The results for the Philippines show that people living in small towns trust strangers more than those living in larger areas The positive association for Vietnam contradicts the conventional hypothesis towns larger in size have more trustful persons Societal theory, as indicated by a measure of confidence in social institutions, performs very well in explaining the origin of trust when tested separately Table shows that confidence in institutions is a statistically significant factor in producing generalized trust across societies Nevertheless, Table reveals that the positively significant association between the two variables only appears in the case of Indonesia and the Philippines but disappears in the rest of four countries The results suggest that the impact of confidence in institution on trust is moderated by factors from other theories In Table 9, demographic variables show significant associations generalized trust in three of the six countries: Indonesia, Malaysia and the Philippines Gender is a significant predictor in Indonesia and Malaysia where men trust more than women Age positively correlates with generalized trust in the Philippines, but in Malaysia the relationship is negative People with higher education likewise display a higher level of generalized trust in Indonesia, while the converse applies in the Philippines Taking the above results into consideration, several patterns on the sources of generalized trust across six Southeast Asian societies can be drawn: 308 LSCAC 2018 International Conference and well-being theory performs better but depends on distinct factors Subjective wellbeing yields converse effects in Malaysia and Thailand, financial household satisfaction has positive effects on generalized trust in both Singapore and Thailand PROCEEDINGS Individual theories rather poorly in explaining generalized trust The success Subjectively-ranked social class is the best indicator of the theory when explaining trust and has a significant effect on generalized trust in three countries: Indonesia, Vietnam and Thailand Generalized trust is well explained by factors stemming from the social system Of social theories, the theory of voluntary organization, network theory and societal theory are more powerful than community theory in predicting generalizied trust The community factor does not appear to be powerful in explaining trust The impact of town size on generalized trust also has a significant effect, at least in the Philippines and Vietnam Trust in known people is the strongest source of generalized trust and has a universal effect on trust even when tested with indicators from the other theories The association between active membership in voluntary organization and generalized trust is inconsistent across the countries, with significant effects in three countries and no effects in the remaining three countries Active organizational membership, however, decreases generalized trust in Vietnam Confidence in institutions produces generalized trust across all countries Table 3: Ordered logistic regression with trust in people met for the first time as a dependent variable and locus of control from personality theory Indonesi Malaysia Philippine Singapor Vietnam Thailand a s e Estimate Estimate Estimate Estimate Estimate Estimate (SE) (SE) (SE) (SE) (SE) (SE) Personality theory Locus of 0.12*** -0.01 0.05* 0.05 -0.05 0.02 control (0.02) (0.02) (0.02) (0.03) (0.03) (0.02) Demographic characteristics Gender 0.16 0.14 0.23* 0.14 0.03 0.12 (0.09) (0.08) (0.11) (0.09) (0.11) (0.08) 309 LSCAC 2018 International Conference PROCEEDINGS Age 0.006 -0.01* 0.004 0.004 0.004 -0.001 (0.003) (0.003) (0.003) (0.003) (0.004) (0.003) Education 0.06** 0.06** -0.09*** 0.01 -0.01 -0.06** (0.02) (0.02) (0.02) (0.02) (0.04) (0.02) Pseudo R 0.0126 0.0046 0.0104 0.0016 0.0023 0.0027 Log -2396.307 -1259.172 -2716.490 likelihood 1803.475 *** *** 2127.529 1239.00 ** ratio *** N 1820 2423 1173 1841 1343 2534 * p