Social Capital and Trust in South-east Asian Cities

Urban Studies, Vol. 41, No. 4, 853–874, April 2004

Social Capital and Trust in South-east Asian Cities

Jeffrey P. Carpenter, Amrita G. Daniere and Lois M. Takahashi

[Paper first received, April 2003; in final form, August 2003]

Summary. This paper conducts a comparative analysis of social capital and environmentalmanagement in two rapidly growing regions in south-east Asia, Bangkok and Ho Chi Minh City.Theoretical and empirical triangulation are used to clarify the nature of social capital and theways that gender, in particular, affects trust and co-operation. Theoretical triangulation isaccomplished by drawing on conceptual arguments made in economics, geography and urbanplanning to develop a theoretical framework explaining social capital and environmental man-agement in south-east Asia. Empirical triangulation is accomplished through a multimethodanalytical approach, including survey methods and experimental games conducted with the samepopulations. The results of the comparative analysis among squatter residents in Bangkok andHo Chi Minh City are presented with a focus on gender differences in trust and co-operation andthe paper concludes with a summary of the results and recommendations for policy and futureresearch.

Introduction

The recent economic crisis in south-east Asialeft public institutions with even fewer re-sources than in the past to address the re-gional problems of environmentaldegradation and insufficient basic infrastruc-ture. These two brief vignettes represent lifein two rapidly developing cities in south-eastAsia—Bangkok, Thailand and Ho Chi MinhCity, Vietnam.

Early one morning, a young woman strug-gles down the lane that skirts her communitylocated in the suburbs of Bangkok. She iscarrying a toddler in one arm and holdingseveral knotted plastic bags full of waste-

paper and plastic in the other. She collectedthe garbage last evening from the swampyspaces surrounding her one-room houseas part of the community-organised effortsto limit the spread of rats and other pestsin the neighbourhood. She has promisedher next-door neighbour, one of thecommunity leaders, that she will do herpart although she notices that most of thehomes on the way to the communal trashbin remain completely surrounded by allkinds of garbage. While many of the neigh-bours appear to be good friends and arewilling to help care for her child

Jeffrey P. Carpenter is in the Department of Economics, Middlebury College, Middlebury, VT 05753, USA. Fax: 802 443 2084.E-mail: [email protected]. Amrita G. Daniere is in the Department of Geography and Planning, University of Toronto, 100 StGeorge Street, Toronto, Canada, M5S 3G3. Fax: 416 946 3886. E-mail: [email protected]. Lois M. Takahashi is in theDepartment of Urban Planning, University of California at Los Angeles, Box 951656, Los Angeles, CA 90095–1656, USA. Fax:310 206 5566. E-mail: [email protected]. The authors express their gratitude and appreciation to Dr Anchana NaRanongat the National Institute for Development Administration in Bangkok, Thailand, Dr Nguyen Quang Vinh and Ms Van Thi Ngoc Lanof the Institute for Social Sciences in Ho Chi Minh City without whom this research could not have been conducted. The authorsalso thank Molly Davidson-Welling, Nadia Abu-Zahra, Kate Swanson and Jill Wigle of the Department of Geography at theUniversity of Toronto for their substantial contributions as well as Michael Timberlake and two anonymous reviewers. The authorsare grateful to the University of California Pacific Rim Research Program, the Social Science and Humanities Research Councilof Canada and the National Science Foundation for funding this research. The authors remain responsible for all errors andomissions.

0042-0980 Print/1360-063X On-line/04/040853–22 2004 The Editors of Urban StudiesDOI: 10.1080/0042098042000194142

JEFFREY P. CARPENTER ET AL.854

while she does piecework on the floor of herhome, surprisingly few of these generouspeople seem interested in cleaning the slumor actively participating in group efforts toimprove community health and hygiene.Consequently, she worries a great deal aboutmalaria, dengue fever and rabies affectingher family.

On a rainy summer afternoon in Ho ChiMinh City (Saigon), several middle-agedwomen (grandmothers all) work hard toready the community meeting area for thenext meeting of the local mothers’ group.The group is planning to meet with People’sCommittee members and representatives ofDOSTE (the local Department of Science,Technology and the Environment) to discussimproving the flow and quality of waterreaching homes situated in the core of thehistorical city centre. Water quality and ac-cess have decreased sharply during the recenturban boom and the low-income women liv-ing in the area are struggling to get enoughwater to clean their homes, cook their foodand bathe their children. The long-standingand well-organised mothers’ group has askedfor and received a meeting opportunity withlocal bureaucrats to achieve some kind ofresolution. Many of the members are willingto donate time and energy (as well as what-ever money can be spared) to implementrepairs and renovations to their water pipes.While few expect that changes will comequickly, many believe that, by standing to-gether, they may be able to improve theirquality of life in the near future.

Although neither of these vignettes repre-sents actual individuals, they are compila-tions of the challenges faced daily bylow-income communities in developingcountries across the Pacific Rim. More im-portantly, these vignettes highlight the im-portance of trust, social integration andco-operation in improving environmentalconditions for communities across the devel-oping world. With decreasing capacity bygovernments to cope with rapid populationgrowth, economic change and environmentaldegradation, governments and internationaldonor agencies (such as the World Bank)

have increasingly looked to the participationof local communities to develop and im-plement environmental management policies.A primary element of such strategies is iden-tifying and expanding the local capacity oflow-income, tenure-insecure neighbourhoodsto engage in such policy design and im-plementation. However, policy-makers andresearchers typically have a poor understand-ing of whether communities are willing toparticipate in improving access to environ-mental services as well as how factors, suchas social networks and social capital, mightinfluence resident participation in the designand delivery of urban services, such as watertreatment and delivery or waste collectionand disposal.

To address this issue, this paper conducts acomparative analysis of social capital andenvironmental management in two rapidlygrowing regions in south-east Asia—Bangkok and Ho Chi Minh City. Theoreticaland empirical triangulation are used to clar-ify the nature of social capital and the waysthat gender, in particular, affects trust andco-operation.1 Theoretical triangulation is ac-complished by drawing on conceptual argu-ments made in economics, geography andurban planning to develop a theoreticalframework explaining social capital and en-vironmental management in south-east Asia.Empirical triangulation is accomplishedthrough a multimethod analytical approach,including survey methods and experimentalgames conducted with the same populations.The remainder of the paper is organised inthe following fashion. The next section pro-vides a brief overview of the social capitalliterature and outlines a conceptual frame-work drawing from these disparate ap-proaches. Next, a brief comparison is madeof the environmental, social and politicalconditions in Bangkok and Ho Chi MinhCity. The paper then summarises the twoempirical methods (survey methods and fieldexperiments) used to elaborate the form andfunction of social capital and how these re-late to environmental management decisions.The results of the comparative analysisamong squatter residents in Bangkok and Ho

SOCIAL CAPITAL IN SOUTH-EAST ASIAN CITIES 855

Chi Minh City are then presented, with afocus on gender differences in trust and co-operation. Finally, the paper concludes witha summary of the results and recommenda-tions for policy and future research.

Social Capital: Form and Function

Scholarly Debates

Although the social capital approach hasgained increasing prominence among schol-ars and policy-makers, there is still widedebate about its form, function and import-ance. Proponents of social capital present itas a critical element missing in modern so-ciety, leading to minimal civic participationand even ineffective governance (for exam-ple, Coleman, 1990; Putnam, 1993, 2000;Woolcock, 1998). Simply put, social capitalis a resource available to individuals thatemanates from group interaction because oftrust, reciprocity and co-operation. The ex-pansion of positive social capital, especiallyfor communities with few economic and pol-itical resources, consequently results in en-hanced economic and political performanceand improved quality of life. Like otherforms of capital (i.e. human, financial, physi-cal), social capital is productive, providingmaterial benefits to individuals who are con-nected to other individuals and groups (i.e.networks and relationships) (Bourdieu, 1985;Coleman, 1990; Putnam, 1993). These socialnetworks and relationships as concretely ex-pressed as the reciprocal (and at times asym-metrical) exchange of material goods,services, information and money, as well asless tangible support such as mutual assist-ance and emotional support (Kim et al.,1997).

Researchers in varying disciplines havestarted with this premise and explored howrelationships among individuals, groups andorganisations relate to economic and politicaloutcomes. Economists and political scientistshave used the social capital approach to in-vestigate communal action as an outcome ofnetwork interaction or relationships ratherthan as aggregations of individual utility-

maximisers (see, for example, Ostrom et al.,1994). Geographers, in addition to the em-phasis on communal action and social struc-ture, have focused on the spatial dimensionsof social capital. Using this spatial lens, so-cial capital is seen to be very dependent onthe interconnectedness of residents inspecific geographical locales, where spaceboth facilitates and constrains the initiationand sustainability of such connections(Gertler, 1997). Public policy and urbanplanning researchers have used social capitalarguments to assess the potential for com-munity-building and capacity-enhancingstrategies in arenas such as economic devel-opment and housing production (for exam-ple, Woolcock, 1998). For those studying thedeveloping world, the social capital approachhas been seen as particularly important forindividuals with constrained material re-sources, as a necessary precursor to commu-nal action and potential community- andregion-wide social, economic, political andeconomic improvements (see, for example,Narayan and Prichett, 1999).

Social capital, both in its form and itsimpact, is likely to differ among places, pop-ulations and communities. There is specu-lation—for example, that social capital levelsin particular urban Asian environments, suchas New Delhi or Jakarta, might be very low(Woolcock, 1998; Isham and Kakhonen,2001). Many researchers argue that one rea-son for the lack of economic growth or prog-ress in some developing countries is a lack ofgeneral and positive social capital (Wool-cock, 1998; Narayan and Pritchett, 1999;Grootaert and van Bastelaer, 2001). In gen-eral, scholars argue that countries, regions orcommunities with greater social capital willbe or are better positioned to take advantageof economic and social opportunities.

Critics of the social capital approach high-light two major flaws in this approach. First,scholars point to Putnam’s and Coleman’sconceptualisation as overreliant on consen-sus-building (i.e. non-confrontation strate-gies)2 and as assuming that low-incomehouseholds and neighbourhoods lack trustand co-operative relationships (for example,


DeFilippis, 2001; Mayer, 2003). Such criticspoint out that trust relationships exist inmany low-income communities, but thatthese social networks are unable to accessfinancial capital and political influence, andthat a focus on non-confrontation as the pre-eminent tactic to accomplish diverse com-munity goals dismisses the importance ofadvocacy and grassroots mobilisation. Sec-ondly, critics of social capital argue that, ascurrently conceived, social capital is both acause and an outcome—that is, researcherslook for evidence of social capital in com-munities that have beneficial outcomes. Thistheoretical conflation of cause and effectleads to methodological challenges. Whilescholars have increasingly refined the con-cept to clarify its form and function, thereremains confusion about its characteristicsand its impact. For example, scholars fromdiverse disciplines tend to use different dataand analytical methods to measure the exist-ence of social capital and its effect onspecific outcomes (Durlauf, 2002).

Reformulating Social Capital

Given this development of the concept andits varying critiques, this paper offers a con-ceptual approach that incorporates gender asa critical dimension of social capital andenvironmental decision-making especially inrapidly developing urban contexts. The un-derlying assumption is that environmentalprotection practices among urban squattersare likely to have been developed as a conse-quence of overlapping social and institutionalnetworks. The form of social capital consistsof the connections of community memberswith other community members and withcommunities, groups and institutions outsidethese communities (the bridging and bondingrelationships or linkage and social integrationargued by social capital scholars). However,in addition to the existence of these networkswithin and without communities, the func-tion of social capital lies in its effectivenessin bringing resources and material environ-mental improvements to communities.Consequently, those communities that pos-

sess or cultivate social capital can effectivelygarner the collective action to maintain orbuild environmental or infrastructure projectsthat benefit their and other communities,such as communal water supplies. Thosecommunities without sufficient social capitalmay have developed trust, reciprocity andeven co-operation (the forms of social capi-tal), but will find it hard to provide suchservices (the function of social capital).Thus, social capital is defined here as effec-tively functioning social networks, with highlevels of trust, symmetrical reciprocity andco-operation within and without the com-munity, resulting in the construction and sus-tainability of communal environmental orinfrastructure projects.

There is much debate in the literature re-garding if and how gender affects the will-ingness of individuals to co-operate ondifferent projects or endeavours. Tradition-ally, economists have argued that gender isnot a relevant consideration in the design ofeconomic policies because all human beingsbasically respond in the same way to incen-tives. Some economists, feminist critiques ofhomo economicus to the contrary, have notfound measurable differences between thegenders when it comes to playing co-oper-ation games (Ferber, 1995; Brown-Kruse andHummels, 1993; Solow and Kirkwood,2002). Many other social scientists, however,have long believed that context matters andthat, in particular, women tend to be moreinterested in investing in household andcommunal welfare than men, which is onereason that numerous micro-credit organisa-tions target their loans to women, particularlyin developing countries (Anderson et al.,2002; Graham and Manning, 2000; Amin etal., 1998). Many researchers note further thatwomen tend to be the primary decision-makers in the household concerning waterand waste management (for example, Whiteet al., 1972; Daniere and Takahashi, 1999)and that communal activities such as com-munity cleaning tend to be seen as extensionsof domestic responsibilities (Milroy and Wis-mer, 1994). More recently, some economistshave come to agree that gender may well be


a key policy variable and economic experi-ments are being conducted to elaborate therole of gender in public and private spheres(Croson and Buchan, 1999; Andreoni andVesterlund, 2001; Barr and Kinsey, 2002).However, more research is needed to articu-late the linkages among social capital, genderand environmental or infrastructure manage-ment, especially in rapidly developing urbancontexts.

The underlying premise in this paper isthat gender influences both the form andfunction of social capital. If social capitalconsists of an effectively functioning socialnetwork for environmental and infrastructureimprovements, then because the socio-spatialcharacteristics of networks tend to differamong women and men, the form and func-tion of social capital is also likely to differamong women and men, as well as acrossplaces and communities. These gender dif-ferences in networks and in social capital canbe traced to the gender relations withinhouseholds and communities. With respect toenvironmental management, social capital islikely to vary across places because the so-cio-political-cultural context will influencethe gender roles in households and communi-ties related to environmental and infrastruc-ture decision-making. However, even giventhese distinct contexts, gender relations andforms and functions of social capital, wemight expect several general tendencies.Women typically have fewer resources avail-able to them than men and would conse-quently depend more on resources accessedthrough social networks than men. Thiswould suggest that women in general wouldbe more likely to engage in trust, reciprocityand co-operative relationships than men, al-though this may not necessarily result ingreater social capital since our definition in-cludes the outcome of environmental andinfrastructure improvements. The literatureon social capital has paid scant attention togender, although much research on com-munity participation, grassroots mobilisationand advocacy has highlighted the role ofwomen as instigators and sustainers of localenvironmental improvements, especially in

terms of environmental justice movements inthe US (for example, Pulido, 1998).

Bangkok and Ho Chi Minh City: AnOverview

While the political, economic and social con-ditions of Bangkok and Ho Chi Minh Citydiffer significantly, the two cities share im-portant similarities including rapid urbangrowth, an increasingly modern economybased on manufactured or processed exports,and rapidly degrading urban environments.In the case of Thailand, despite the continuedeffects of the region-wide economic crashsustained in 1997, the metropolitan area ap-pears to be growing more rapidly than theThai economy overall (approximately 2 percent in 2001). Environmental degradationhas been especially rapid; widespread lack ofaccess to drinkable water, clean air and ad-equate sanitation has resulted in considerablerisks to public health in Bangkok, as in manyAssociation of South East Asian Nations(ASEAN) cities (Daniere et al., 2002). Viet-nam, also a member of ASEAN, was slightlyless affected by the Asian economic crisiscompared with other nations, related to itshaving lower levels of foreign investment(compared with such nations as Indonesia,Malaysia and Thailand). GDP grew by 5 percent in 2001 and the Vietnamese economycontinues to expand steadily. Economicgrowth, as in Thailand, is concentrated inVietnam’s urban areas, particularly Ho ChiMinh City and Hanoi (Drakakis-Smith andDixon, 1997; Parenteau, 1997; Gainsbor-ough, 2002).

Bangkok is widely seen by scholars andpolicy-makers as a stark example of uncon-trolled and unsustainable growth and con-comitant environmental degradation.Thailand is known for having one of themore laissez-faire governments in the world,at least in practice if not in terms of theactual letter of the law. Industrial firms com-monly ignore Thai regulations regarding en-vironmental protection. However, recentmedia accounts have indicated a growingpublic dissatisfaction with environmental


conditions, expressed through public protestsfocused on air and water pollution. Localmobilisation such as urban squatter associa-tions, loosely organised through NGOs and/or community groups, can in specificcircumstances—particularly with the assist-ance of well-placed contacts in the BangkokMetropolitan Authority or one of the na-tional-level Ministries—achieve substantialimprovements in local quality of life. Squat-ter settlements have obtained piped waterconnections, structural upgrades or daycarecentres as well as improved land tenure ar-rangements through local organisation andlobbying of key bureaucrats.

Ho Chi Minh City, although slightly morethan one-third the size of Bangkok, has ex-perienced extremely rapid expansion of itsurban area, leading to its characterisation asthe world’s next ‘Bangkok’ (Drakakis-Smithand Dixon, 1997). Compared with Thailand,the communist political regime in Vietnamretains much more centralised control, inmost cases, over the environmental and pub-lic health practices of local industry and theconstruction and maintenance of basic urbanservices such as sewers, roads, water pipesand solid waste collection. At the same time,economic conditions are changing so quicklythat, especially in Ho Chi Minh City, pol-lution is growing and squatter settlements arerapidly expanding. Where once the localPeople’s Committees in charge of each ward(Ho Chi Minh City is divided into largedistricts that in turn are composed of smallwards, some as small as a large city block)were able to monitor and control the numberof people and households within their par-ticular areas, this is less true today. Manysettlements are becoming overcrowded; thereis more informal housing being built alongthe open spaces within and exterior to tra-ditional residential areas; and the supply ofpublic infrastructure lags far behind the cre-ation of housing, particularly in neighbour-hoods with low-income residents. In general,however, it is evident that the municipalgovernment in Ho Chi Minh City is more orless following a plan to supply differenttypes of infrastructure to various parts of the

metropolitan area, particularly planned high-income areas. It is less clear how responsivethe metropolitan government is to the needsof low-income communities, as representedby NGOs and even ward-level People’sCommittees.

While there are obviously many differ-ences in the politics, economies and culturesof the two cities, the challenges faced aresomewhat similar because of two factors:Vietnam is likely to continue to depend ondoi moi (the liberalisation process) to helpjump-start its economy; and both Thailandand Vietnam are increasingly exposed to for-eign investment and the consumer culture ofthe West apparent in most globalizing cities(Kim et al., 1997). Consequently, urban en-vironmental degradation is likely to deepen,requiring new and cost-effective environ-mental management strategies to sustaineconomic growth. Community participationremains one of the factors believed by inter-national donor agencies, NGOs and govern-ment agencies to be vital to the potentialsuccess of urban environmental policies. Afundamental challenge to understanding andeffectively incorporating community actionremains the lack of knowledge about theinterests and motivations guiding local resi-dents.

Methodology

The most prevalent approach to studying so-cial capital is to analyse survey data (Put-nam, 2000; Onyx and Bullen, 2000; Ishamand Kahkonen, 2001; Daniere et al., 2002).This mode of analysis depends on individualresponses to survey questions on trust, co-op-eration and relationships to test hypothesesrelated to the nature of social networks orco-operative behaviour on specific outcomes.Economic games have also been increasinglyused to measure directly levels of trust andco-operative behaviour. Such games or ex-periments record the observed behaviour ofparticipants in terms of strategy exhibitedduring the game and use these responses toexpress a measure of trust within a specificcommunity or group of people.3 This paper


uses these two methods to conduct a compar-ative analysis of Thai and Vietnamese urbanresidents to triangulate empirically the formand function of social capital in these twocities.4

We focused on five low-income communi-ties in Bangkok, Thailand, and Ho Chi MinhCity, Vietnam.5 In the past three years, wehave studied the role of social capital inenvironmental management within five dif-ferent slum communities in these two cities.In doing so, we conducted household surveyson demographics and social networks andin-depth qualitative interviews with key com-munity informants and, most recently, wecollected data via incentive compatible fieldexperiments designed to estimate the amountand intensity of trust and social capital. Thispaper compares the results of our findingsfrom more conventional survey-based re-search on social capital (implemented in2000) to the data collected from experimen-tal games played in the same 10 communitiesduring the summer of 2002.

Bangkok Communities

In Bangkok, the communities were selectedthrough the use of a sampling frame devel-oped from a somewhat dated study ofBangkok slums and the local expertise of theproject team (Setchell, 1992). While not ran-dom, the selected communities are generallyrepresentative of the broad range of squatterareas in terms of size, history, location andenvironmental conditions existing throughoutthe Bangkok Metropolitan Region. Theselection criteria included low average percapita or household income (as traditionallymeasured) within the slum or neighbourhoodand disparate locations in order to ensure thatthere would be some variation among thefive chosen communities in terms of accessto services (for example, Daniere and Taka-hashi, 1999). We contacted the neighbour-hood organisation or community leaders ineach of the neighbourhoods to request per-mission to work with the community. It wasunderstood that we hoped to return to thecommunities for the next three years to

gather information about social relations andnetworks, environmental management, waterand sanitation practices and some socioeco-nomic data. Each of the five communities weapproached agreed to work with us and weonly surveyed households and individualswho, when randomly selected, agreed to par-ticipate.

In the case of experimental games, poten-tial participants were informed about the op-portunity through leaflets and communityannouncements. Participants were asked tomeet at a central location, generally the localcommunity or daycare centre, if they wantedto volunteer for the games. The players wererandomly selected from among the volun-teers gathered at the site.

Ho Chi Minh City Communities

To select communities in Ho Chi Minh City,we needed to rely even more on local knowl-edge and connections. The communitiescould not be selected from a sampling framebut rather were proposed by the People’sCommittees responsible for specific city dis-tricts. We approached five different districtsand asked them to nominate one or two oftheir most impoverished wards or communi-ties for inclusion in the project. The Commit-tees were generally quite co-operative andwere very knowledgeable about the mostimpoverished communities within their geo-graphical boundaries. We normally followedthe suggestions of the People’s Committees.Since the research project has the support ofthe Vietnamese government and one of ourproject team members is a government em-ployee at the Institute for Social Sciences (anacademic think-tank), we were able to obtainpermission to work in a variety of wardswithin the city and to conduct the householdsurvey relatively free of control or interfer-ence. It is possible, of course, that we weredirected to showpiece communities although,given our extensive travels throughout thecity, this seems very unlikely to us and to ourVietnamese colleagues.

Playing or conducting experimental gamesin Ho Chi Minh City was somewhat more


complicated than in the Bangkok communi-ties. We chose to conduct the games in oneor two of the meeting rooms within the Insti-tute for Social Sciences as it proved to be theleast disruptive to communities, as well asthe option the most free of party (or People’sCommittee) monitoring. Participants fromthe different communities who volunteeredfor the experiment were transported by vanor taxi to the centre, played the game andwere then given transport back to their com-munities. The participants were all volun-teers who were able to spend approximatelytwo or three hours away from their com-munities. As in Bangkok, potential partici-pants were informed that the game wouldresult in guaranteed and immediate earnings;there was no lack of volunteers.

Household Survey

The first data-gathering effort took place inthe summer of 2000 and involved surveying500 households in each city. We designedand translated two almost identical surveyinstruments (one in Thai and the other inVietnamese) to administer in each of the 10selected neighbourhoods (5 in each city).While very similar, the surveys used in eachcountry nonetheless reflected differences inculture and practice between the two cities.6

Graduate students or staff recruited from thelocal authors’ universities/research staff ad-ministered the surveys. These enumeratorswere provided with intensive training in howto administer the survey by the principalinvestigators and relied on procedures usedin two previous household surveys conductedwith high levels of success in south-eastAsian urban communities (Daniere andTakahashi, 1999; Crane and Daniere, 1996).

Approximately 15 per cent of all house-holds within each community for a total of1000 household representatives (500 percity) were surveyed during the summer. Ineach household, the individual most respon-sible for water, sanitation and solid wastemanagement was asked to participate in thesurvey. In many cases, the respondent wasfemale, as women are generally responsible

for these household decisions (Daniere andTakahashi, 1997). The surveys were de-signed to assess social networks, health be-haviours and environmental practices andconsisted of three sections: household en-vironmental attitudes and practices, includingquestions about the time spent in water- andsanitation-related activities, the type of facili-ties used and monthly expenditures; house-hold health behaviours such as frequency ofillness and visits to medical facilities as wellas knowledge regarding the relationship be-tween the environment and health; and, theform and function of the household’s socialnetworks, such as the number and type ofpeople on whom the household relied fordaily assistance, the connections of house-hold members to others outside the house-hold for various purposes, the existence ofand participation in community or local or-ganisations, and trust in various externalagencies and levels of government.7

Experimental Games

The field experiments, or games, were con-ducted during the summer of 2002 withmembers of the same 10 communities (5communities from each city) where thehousehold surveys had been conducted twoyears earlier. Initially, we conducted thegames both with 24 individual workers andlabourers in and near the Institute for SocialSciences in Ho Chi Minh City and withstudents enrolled in NIDA in Bangkok. Thelatter experiments acted as both pilot testsand controls for the field experiments andallowed us to examine how the behaviour ofpeople in communities who typically facereal social dilemmas, such as the mainte-nance of clean pathways and sewers, com-pare with the responses of typical workers orstudent participants.

The games that we played with workers,students and community members were sim-plified versions of basic voluntary contribu-tion games as described in Davis and Holt(1993) that have already been tested in manyNorth American and other locations.8 Eachgame consisted of 10 rounds of play divided


in 2 separate sections, or treatments, of 5rounds each. During the games, participantswho volunteered to play were assigned togroups of 4 people who played together forall 10 rounds. In each community, we con-ducted 6 games so that, in all, 24 individualsparticipated from each site. We played 2games with only male participants, 2 gameswith only female participants and 2 gamescomposed of 2 men and 2 women. We wereable to gather responses from 240 people (2cities, 5 sites per city, 24 people per site). Insome cases, we played 1 game at a time butin the majority of cases we were able to run2 games at a time in which case the 2 female,male or mixed gender games were playedconcurrently. The 6 games usually took be-tween 5 and 7 hours to play for a particularcommunity because of the significant organi-sational requirements of the game. Finally,we set the stakes of the game (depending oncurrency denominations) so that the potentialearnings of game participants if they all co-operated fell somewhere between a half anda whole week’s wages (approximatelyUS$44 in Thailand and US$12 in Vietnam).

When we began each game, we sat partic-ipants at some distance from each other andrequested that they not talk to or communi-cate with each other until we were com-pletely finished playing all 10 rounds of thegame. At the beginning of each round (orperiod), each player was given 10 bills orcoins (5 Baht coins in Thailand and 1000Dong bills in Vietnam). The experimenter satclose to the players but behind a blind or achalkboard, or a blanket suspended from theceiling; in some cases, the experimenter satoutside the room. During each round, everyplayer would go behind the blind twice toeither contribute or receive money. In eachcase, the blind was constructed so that play-ers could make their decisions privately.

First section: voluntary communicationmechanism (rounds 1–5). On the first tripbehind the blind, participants decided howmany of the 10 bills (or coins) to allocate to

a common cause (or public good) and howmany to keep for themselves. Once everyplayer had been behind the blind for the firsttime, as they were promised, the exper-imenter counted and doubled the sum of allthe contributions. Each person then returnedbehind the blind to receive an equal share ofthe doubled amount. At that time, eachplayer would see a listing of the differentcontributions of each of the 4 players (in-cluding themselves) but were not told whichplayer contributed which amount. This pro-cedure was followed for 5 periods or rounds.Notice that this provides the incentives of asocial dilemma because individuals can al-ways do better by free-riding off the contri-butions of the other group members.9

Second section: social sanctioning (rounds6–10). At the end of round 5, we changed therules to add the following, simple, socialsanctioning mechanism. It is important thatwe did not tell our participants that we wouldchange the rules between rounds 5 and 6because we did not want them to conditiontheir early play on the future rule changes.We told the participants that for one-fifth (20per cent) of a bill or coin (i.e. 1 Baht or 200Dong), any participant could have a picturedisplayed (see below) at the beginning of thenext round.

In the instructions, we said that the pictureis meant to express one’s dissatisfaction withthe contribution choices of one or more ofthe participants in the group. In this case,while it is costly to show disapproval, anybenefit of the sanctioning must be due tosocial pressure because the sanction has nomaterial fine attached to it.

The social sanctioning game began eachperiod, as in the previous 5 rounds, with anendowment of 10 bills (or coins) and, on the


Figure 1. Average contribution per round, by country.

first trip behind the blind, each participantfaced the same 2 choices—to allocate someor all of the money to the public good or tokeep it all. The only difference occurred onthe second trip behind the blind, when theparticipant was informed how much the otherparticipants contributed and picked up herearnings from the public good. At that point,each participant could elect to pay one-fifthof a bill or coin visually to sanction the restof the group at the start of the next round.

We were interested in noting the numberof times that a participant was willing to payto show her dissatisfaction as economists usethis as a measure of the propensity to punishothers, which can be considered an elementof trust and co-operation. At the completionof both sections, following the tenth round,all the participants were interviewed andgiven another payment for staying for theinterview. A brief survey was conducted dur-ing this final interview and included demo-graphic questions as well as someinformation regarding individual values/norms and their environmental behaviour.

Results

Overall Trust

Perhaps the most important finding from ourresearch is that both methods, but particu-larly the game methodology, revealed veryhigh levels of trust and co-operation amongparticipants in both cities and among all thecommunities studied. As can be seen fromboth Table 1 and Figure 1, the contributionper round of the experimental game averagedmore than 70 per cent of the amount it waspossible to contribute for much of the game.Such high levels of contribution are veryunusual and almost without precedent, par-ticularly in North America and Europe wherethe average mean contribution is generallybetween 40 and 60 per cent of the possiblemaximum.

In general, we find that the Vietnameseparticipants are more likely than Thai partici-pants to contribute larger amounts to thecommon pot. This is particularly true in thefirst 5 rounds of the game. Assuming eachround of players to be independently drawn


Table 1. Overall game results (mean contribution per round)

Round

Cities 1 2 3 4 5 6 7 8 9 10 Total

Bangkok (N � 120) 4.7 5.9 6.3 6.7 7.1 6.4 7.2 7.5 7.5 8.0 6.7Ho Chi Minh (N � 120) 7.1 7.1 7.2 7.3 7.4 7.1 7.6 7.7 7.9 7.8 7.4Average of both 5.9 6.5 6.8 7.0 7.3 6.8 7.4 7.6 7.7 7.9 7.1T-test scores 0.00 0.00 0.02 0.10 0.52 0.05 0.20 0.45 0.30 0.58 0.22

Table 2. Trust index calculated from household survey, 2000

Cities Trust index Standard deviation

Bangkok 2.27 0.98Ho Chi Minh City 1.04 1.37

Total average 1.66 1.18

from the same distributions (which may ormay not reflect the underlying reality of thecommunities in question), a t-test rejects thehypothesis that the mean contributions ofresidents from Bangkok and residents of HoChi Minh City are equal at beyond the 5 percent level for rounds 1, 2, 3 and 6. Thedifference between the two cities’ partici-pants is not significant on average as can beseen in the total column. In essence, theVietnamese contributions were fairly steadythroughout the course of the game, which isquite atypical for a majority of this type ofgame. Most researchers report a decrease incontributions over each successive round ofthe experiment, despite the introduction of asanction opportunity (Carpenter, 2002;Solow and Kirkwood, 2002). It is also appar-ent that the participants of both cities weremotivated to increase their contributions fol-lowing the imposition of sanctions; however,it is hard to tell whether punishment mat-tered, considering that the trend before pun-ishment was to increase one’s contribution.

Results from the household survey, col-lected from respondents in these same com-munities two years earlier, found similar, butless striking, evidence of co-operation. Thesurvey contained a number of questions re-garding trust and co-operation, including theNational Opinion Research Centre’s GeneralSocial Survey that has included questions on

trust since 1972.10 Table 2 presents the meanvalues of a trust index that we constructedusing responses to the trust question to sum-marise each individual’s basic level of trust(Daniere et al., 2002).11

While we found evidence of a high level oftrust using the survey, it appears that respon-dents in Vietnam are somewhat less trustingof their neighbours and community membersthan respondents in Thailand. The averagevalue of the trust index is only equal to 1.04in Vietnam, but is 2.27 in Thailand. Exceptfor being more likely to say that they feel thattheir immediate neighbours are like familythan Thai households (14.9 per cent of Viet-namese respondents compared with only 6.5per cent of Thai respondents), Vietnameserespondents are less likely to trust their neigh-bours across the spectrum of trust questions.

A possible explanation for this trend, ac-cording to our research collaborators, is thatVietnamese communities are poorer andmore politically charged than Thai communi-ties. It is certainly true, based on the socio-economic data we collected at the same time,that the expenditures per household are some60 per cent less among Ho Chi Minh Cityrespondents than in the Bangkok neighbour-hoods that we surveyed. Similarly, there is amuch more established and enforced politicalsystem at work in Ho Chi Minh City com-munities than one finds in Bangkok. Our data


Table 3. Social integration results from household survey, 2000 (percentages)

Cities Q35 Belong to Q45 Participatory Q33c Participate ingroup projectcommunity groups decision-making

Bangkok 13.4 (65) 35.7 (36) 71.7 (256)Ho Chi Minh 30.6 (153) 54.0 (81) 62.1 (195)

Total average 56.0 67.4

Table 4. Game results by gender (mean contribution per round)

Round

1 2 3 4 5 6 7 8 9 10 Total

Female (N � 120) 5.6 6.4 6.6 6.9 7.3 6.6 7.1 7.3 7.6 7.8 6.9Male (N � 120) 6.2 6.6 6.8 7.1 7.2 6.9 7.6 7.9 7.8 8.0 7.2Average of both 5.9 6.5 6.7 7.0 7.3 6.8 7.4 7.6 7.7 7.9 7.1T-test scores 0.18 0.69 0.67 0.73 0.86 0.39 0.16 0.04 0.57 0.39 0.47

collection—for example, engendered muchmore monitoring, interference and commentin Vietnam than it did in Bangkok where wewere essentially ignored by local, regionaland national authorities.

Interestingly, while trust levels inBangkok, according to survey results, arehigher in Bangkok than in Ho Chi Minh City,we did find that other dimensions of socialcapital appear to be higher in Ho Chi MinhCity than in Bangkok. Several questions inthe survey were designed to measure or as-sess the degree of social integration withincommunities. The social integration ques-tions measured the connection between thelevel of intracommunity networks and en-vironmental or infrastructure practices. As anexample, respondents were asked about thewillingness of their household and othercommunity members to raise and contributesignificant resources for environmental im-provements.

An important point is that respondents inHo Chi Minh City are more than twice aslikely to belong to a community group thanBangkok respondents (Table 3). More than30 per cent of all Ho Chi Minh City respon-dents belong to community groups comparedwith a mere 13.4 per cent of all Bangkokrespondents. Furthermore, the groups are de-

scribed by Vietnamese participants as rela-tively inclusive and open in their operatingprinciples compared with the groups de-scribed by Bangkok respondents. The im-portance of such features of group dynamicsis often highlighted in the social capitalliterature as helping to create situations thatgenerate positive social capital (Isham andKakhonen, 2001).

Another indicator of social integration isthe percentage of households who said theyparticipated in community improvementprojects in the past. In this case, Bangkokrespondents overall claim somewhat higherlevels of actual participation than the peoplewe surveyed from Ho Chi Minh City. Ap-proximately 62 per cent of all respondents inHo Chi Minh City reported that they partici-pate in community projects compared withalmost 72 per cent of all Bangkok respon-dents.12 Interestingly, neither group member-ship nor participation in community effortswithin the past year differed significantlyacross gender lines. What is evident is thatresponses to trust questions do not capture allsocial capital dimensions and, also, there isinconsistency in responses related to othertypes of social capital question in terms ofwhich cities and gender groups exhibit moresocial capital potential.


Figure 2. Average contribution per round, by gender.

The Role of Gender

Table 4 and Figure 2 present the averagecontributions of all participants in each roundand in total, broken down by gender. Com-bining all groups, from both countries, itappears that male participants contributedslightly more, on average, than female partic-ipants although this is not a statisticallysignificant difference except for round 8. Ascan be plainly seen from Figure 2, both maleand female participants increase their contri-butions to the common pot over the course ofthe game with a slight dip in contributionsimmediately following the introduction ofthe sanction option. Again, this is a veryunusual pattern as, in most examples in theliterature, participants tend to decrease theircontribution fairly steadily during the courseof the game.

Table 5 presents data for both cities bro-ken down by gender. The most obviousfinding here, and one that is supported by ourinitial premise, is that female participants inHo Chi Minh City contribute a significantlyhigher level of dong to the common poolthan male participants. As shown in Figure 3,female contributions are higher than male

contributions for the entire 10 rounds of play.In addition, after sanctioning is introduced inround 6, there is a marked increase in malecontributions for a short period that thenflattens out somewhat. Again, what is partic-ularly striking for the Vietnamese case is thataverage contributions are so consistentlyhigh across the board. In most instances ofsimilar games played in other experiments,average contributions hover between 40 and60 per cent of the social optimum whenpunishment that actually imposes a monetarycost on the punishee is allowed and decreasesto 10 or 20 per cent when no punishment isallowed, while in Vietnam the average con-tributions are close to 75 per cent.

Finally, we see in the final portion ofTable 5 and Figure 4, that Thai male partici-pants are more likely than Thai female par-ticipants to contribute significant amounts tothe common pot. While in both cases contri-butions increase more or less steadily overthe course of play, male participants contrib-ute significantly more (Figure 4). As is thecase in Vietnam, these gender differences arestatistically significant at the 5 per cent level.Interestingly, the data from the earlier house-hold survey suggested that no real trend ex-


Table 5. Game results by city and gender (mean contribution per round)

Round

1 2 3 4 5 6 7 8 9 10 Total

BangkokFemale (N � 60) 3.8 4.9 5.4 6.1 6.5 5.6 6.4 6.6 6.8 7.3 5.9Male (N � 60) 5.6 6.8 7.1 7.3 7.8 7.3 7.9 8.4 8.3 8.7 7.5Average of both 4.7 5.9 6.3 6.7 7.2 6.5 7.2 7.5 7.6 8.0 6.7T-test scores 0.00 0.00 0.00 0.02 0.01 0.00 0.01 0.00 0.00 0.00 0.00

Ho Chi Minh CityFemale (N � 60) 7.5 7.9 7.8 7.8 8.1 7.7 7.8 8.0 8.4 8.2 7.9Male (N � 60) 6.7 6.4 6.5 6.8 6.7 6.5 7.3 7.5 7.3 7.4 6.9Average of both 7.1 7.2 7.2 7.3 7.4 7.1 7.6 7.8 7.9 7.8 7.4T-test scores 0.11 0.00 0.00 0.06 0.00 0.01 0.24 0.29 0.01 0.05 0.08

Figure 3. Average contribution per round in Vietnam, by gender.

ists regarding trust between male and femalerespondents, particularly when the data fromthe two countries are combined.

In Table 6 and Figure 5, we disaggregatethe data even more. As noted earlier, weconstructed different types of group to playthe game in each of the 10 communities. Wehad 2 groups each of all-female participants,2 groups each of all-male participants and 2groups each of mixed-gender participants. InTable 6, rather than aggregating all the fe-male participants and all the male partici-

pants together, we divide the data further bynoting whether the participants were in anall-male, all-female or mixed-gender group.

The most striking point to be gleaned fromthis table is that, while the gender differencebetween male and female participants inVietnam holds true for segregated groups, itvirtually disappears in the mixed groups. InBangkok, in contrast, the gender differencesnoted earlier are apparent in both segregatedand mixed-gender groups. Figure 5 sum-marises the contributions of all mixed groups


Figure 4. Average contribution per round in Thailand, by gender.

in both countries disaggregated by male andfemale participants. In the case of Bangkok,the participants in the mixed groups havetotal higher contributions across the roundsplayed than participants in the correspondingsegregated groups. In other words, playingtogether is associated with both groups con-tributing more than when they play in single-sex groups (i.e. females in mixed gamescontribute on average 6.1 per round vs 5.9per round in segregated groups and malesfrom mixed groups contribute an average of8 per round compared with 7.5 when theyplay in segregated groups). The Vietnamesesituation is similar in that the female andmale participants in mixed groups con-tributed higher total average contributionsthan participants in all-female or all-malegroups. Interestingly, Ho Chi Minh City fe-male participants in mixed groups are thehighest bidders or contributors of all thedifferent groups we examined.

Summary and Conclusions

An important finding elaborated in this paperis the very high, by Western standards, levels

of trust and co-operation, as estimated by theaverage contribution levels during the courseof the games, found in both Bangkok and HoChi Minh City participants. As noted earlier,most researchers have found that participantscontribute between 40 and 60 per cent of thesocial optimum (or, in terms of our game,between 4 and 6) and that the average contri-bution decreases with rounds played. Wefound that average contributions were be-tween 60 and 80 per cent and that playersgenerally increased their contributions orbids over the 10 rounds. One possible expla-nation for this is that the participants in bothcities come from relatively stable and verylow-income communities. Most of thesecommunities have a history of working to-gether on minor and major initiatives to im-prove their communal quality of life.Furthermore, while some of the communitiesare quite populous (up to 1500 people in thecommunity), many of the individuals knoweach other to some degree. Thus, both thesize and the nature of the communities fromwhich the players are drawn might influencethe measured level of trust or co-operation,and are likely to exceed that observed among


Table 6. Game results by group, country and gender (average contribution per round)

Round

1 2 3 4 5 6 7 8 9 10 Total

By group and genderFemale (N � 80) 5.4 6.1 6.4 6.8 7.0 6.5 6.8 7.1 7.6 7.7 6.7Male (N � 80) 5.9 5.9 6.3 6.6 6.7 6.7 7.3 7.6 7.4 7.9 6.8Mixed (N � 80) 6.5 7.4 7.5 7.6 8.1 7.3 8.1 8.2 8.1 8.2 7.7Mixed female (N � 40) 6.3 7.2 7.2 7.2 7.9 7.0 7.9 7.7 7.7 7.9 7.4Mixed male (N � 40) 6.8 7.5 7.9 8.0 8.3 7.5 8.4 8.7 8.6 8.5 8.0

Total average 5.9 6.4 6.7 7.0 7.3 6.8 7.4 7.6 7.7 7.9 7.1

Bangkok, by group andgenderFemale (N � 40) 3.6 4.7 5.1 6.1 6.2 5.7 6.2 6.5 7.0 7.5 5.9Male (N � 40) 5.4 6.0 6.7 7.2 7.5 7.1 7.7 8.0 8.1 8.7 7.2Mixed (N � 40) 5.0 6.5 7.0 6.8 7.7 6.6 7.6 8.0 7.5 7.8 7.1Mixed female (N � 20) 4.1 5.4 6.2 6.1 7.0 5.4 7.0 6.7 6.3 6.9 6.1Mixed male (N � 20) 6.0 7.6 7.9 7.5 8.4 7.7 8.2 9.3 8.6 8.7 8.0

Total average 4.7 5.7 6.3 6.7 7.2 6.5 7.2 7.5 7.6 8.0 6.7

Ho Chi Minh City, bygroup and genderFemale (N � 40) 7.1 7.4 7.7 7.5 7.8 7.2 7.4 7.6 8.2 7.9 7.6Male (N � 40) 6.3 5.8 5.8 6.0 5.9 6.2 6.8 7.2 6.7 7.0 6.4Mixed (N � 40) 8.0 8.2 8.0 8.4 8.5 8.0 8.6 8.4 8.7 8.6 8.3Mixed female (N � 20) 8.4 9.0 8.2 8.3 8.7 8.6 8.7 8.7 9.0 8.9 8.7Mixed male (N � 20) 7.6 7.4 7.9 8.5 8.2 7.3 8.5 8.1 8.5 8.3 8.0

Total average 7.1 7.1 7.2 7.3 7.4 7.1 7.7 7.9 7.9 7.9 7.5

Note: Total averages calculate the combined averages of female, male and mixed categories.

students at major universities or other similartypes of unconnected social group (Kollock,1998).

In terms of the noted differences in genderbetween countries, we have several compet-ing explanations as to why female partici-pants are more trusting and co-operative thanmale participants in Ho Chi Minh City butless co-operative and trusting than male par-ticipants in Bangkok. In Vietnam, whilehousehold management tasks tend to be gen-der-driven—i.e. if there is a woman ofsufficient age and experience in the house-hold, she will be in charge of running thehousehold on a day-to-day basis—womenand men in a household do also sharebudgeting responsibility and decision-mak-ing. In addition, most women in Vietnamearn income for work outside or inside the

home and are, consequently, material con-tributors to household income and wealth.Women have been targeted historically bythe Communist Party in Vietnam to work oncommunal projects and are supposed to betreated, in law anyway, as being equal tomen. There is an impressive legacy ofrhetoric in Vietnam extolling gender equalityand the active participation of women in theeconomy and community noted by a numberof commentators (Coit, 1998). This legacy orphilosophy may explain the high rates ofcontribution among the Vietnamese partici-pants overall and the particularly high levelof trust and co-operation that we found onthe part of Vietnamese female participants.

The household division of labour inBangkok is somewhat different. Tradition-ally, men in Thailand are supposed to give


Figure 5. Average contribution per round, mixed by country and gender.

everything they earn to their female partnerso that she, with her heightened sense ofhousehold needs and priorities, can managethe household budget. While this pattern isbecoming less prevalent, it is a well-acceptedThai practice that many claim to follow. Infact, many men apparently hand over a size-able portion of their earned income to theirpartners but retain a small part for gamblingand gaming purposes. Gambling is endemicin Thailand and is one of the most popularpastimes for men, in particular. While wehave no survey evidence on this point topresent at this time, it may well be that whilethe participants in Bangkok who played thegame were similar to Vietnamese in that theywere residents of long-standing and rela-tively co-operative communities, they dif-fered in their gender relations and habits.

The Thai female participants played con-servatively and were less trusting, perhapsstemming from their personal experience inbalancing meagre household budgets. Theirplay exhibited much more risk-aversion thaneither Thai male or Vietnamese male or fe-male participants. At the same time, it ispossible that Thai male participants who mayhave looked upon the funds provided as ‘play

money’ actually used it for gambling pur-poses in that the game can be perceived, bypeople who gamble regularly, as a gamblewith a high probability of paying off or, inother words, as a very good bet. Gambling isunusual and generally scorned in Vietnam,which might explain the lack of this kind ofattitude among either Vietnamese male andfemale participants.

The two sets of results drawn from thesurvey and the experimental games providean opportunity to assess the complex natureof trust, reciprocity and co-operation amongthese Thai and Vietnamese participants. Ac-cording to the survey, respondents inBangkok say that they are more trusting oftheir neighbours than in Vietnam, indicatingthat social capital is more solidified amongBangkok respondents. However, the be-havioural measures of trust (through the ex-perimental games) indicate that the surveyquestions may not provide an accurate por-trait of the level of trust within a community(self-report is likely to inflate the degree oftrust). These self-reports may be seen insteadas providing a sense of the socially accept-able norms of behaviour within communities.That is, the responses reflect the prevailing


norms of behaviour within specific communi-ties across places. While it is important forour definition of social capital to understandthe direct linkages among trust, reciprocity,co-operation and environmental improve-ments, it is also vital to explore whether thesenetworks are empowering or coercive. Thesocial capital approach assumes that resourcesacquired through networks are positive andempowering but, as Bourdieu (1985) has ar-gued, such networks can also reinforce exist-ing relations of power, resulting in coerciveand disempowering, although perhaps eco-nomically advantageous, circumstances.

The experimental game results indicatethat the Vietnamese participants contributedsignificantly more to a common pool ofmoney from which all would reap benefitsthan the Thai participants. These results areconsistent with the surveyed reported behav-iour of Ho Chi Minh City respondents vis-a-vis Bangkok respondents in terms of socialintegration—i.e. group membership andcommunity participation. The game resultsmeasure a different component of socialcapital than the trust questions on the survey;the experimental game assesses face-to-faceinteraction under relatively controlled cir-cumstances, providing additional insight intohow participants might behave when facedwith the social dilemmas inherent in commu-nal environmental and infrastructure projects.

The two empirical analyses together pro-vide a more complete portrait of social capi-tal than either method alone. The surveyresults provide insight into the stated prefer-ences of respondents, indicating the prevail-ing norms of behaviour especially in thesetwo cultures, where conflict and disagree-ment are typically avoided. The experimentalgame results provide data on potential behav-iour given a narrowly defined set of rules,actions and outcomes. These results provideinsight into potential actions by participantsgiven the complexity of actual projects andalso indicate the potential high level of com-munity support that might be gained if com-munities in both places were made aware ofindividual resource gains through social net-work interaction and communal action.13

There are several research and policy im-plications that may be drawn from this analy-sis. First, in terms of implications forinvestigating social capital, especially forscholars interested in developing countries,the analysis made clear that the use of amultimethod approach is critical in highlight-ing the complex dimensions of trust andco-operation. Using a combination of the twomethods (surveys and experimental games)tended to compensate for weaknesses in ei-ther method alone and also highlighted po-tential areas for policy intervention. Thegender differences in trust and co-operationindicated by the survey and the experimentalgame results suggest that there are aspects ofsocial and cultural norm structure (for exam-ple, gambling, household budget decisions)that play central roles in the nature ofcommunal action. For example, researchershave suggested that women in the developingworld are decision-makers in household ac-tivities concerning water, solid waste andsanitation, further indicating that womenmay therefore be the key actors in anycommunal activity associated with com-munity environmental improvements. How-ever, this analysis indicates that trust andco-operative action may be more complexthan the aggregation of household action tolarger groups. Instead, the findings indicatethat participants’ stated beliefs and attitudesmay differ radically from their co-operativebehaviour and that such behaviour may de-pend on the gender relations of the specificgroup (i.e. mixed-gender groups versus sin-gle-gender groups). While social capital maybe seen in the literature as generally non-gen-der specific, this analysis points to the needfor detailed articulation of trust and co-oper-ation across many dimensions of social rela-tionships (age, race/ethnicity, household size,etc.). Much more work is necessary to deter-mine how broader social relationships, suchas gender, point to viable strategies orsignificant obstacles to harnessing and sus-taining the form and function of social capi-tal in distinct places and communities.

Related to this, a second set of implica-tions involves how social capital should be


conceptualised and framed given a multi-method empirical approach. Multimethod ap-proaches have embedded within themdisparate and at times contradictory concep-tual underpinnings, making interpretationsand conclusions challenging. For example, inthis analysis, survey methods and experimen-tal games resulted in seemingly contradictoryresults concerning the gender differences intrust and co-operative behaviour betweenThai and Vietnamese participants, and be-tween male and female participants. Wewould argue, however, that each methodhighlights significant but distinct elements ofsocial capital that provide components to acomprehensive understanding of this con-cept. It is difficult to assess which result ismore accurate, the survey results or the ex-perimental games, since accuracy is depen-dent on the conceptual assumptions madeprior to the analysis. However, there is alarger issue that emanates from a multi-method approach, that both results are accu-rate but that they measure varyingdimensions of social capital that are relevantfor developing distinct policy strategies forcommunity action. Consequently, the chargefor future research is to continue to use mul-timethod approaches to develop a more com-prehensive portrait of social capital, trust andco-operation.

Finally, the third set of implications isrelated to developing effective and appropri-ate policy interventions that will incorporatethe norm structures and daily practices ofcommunity members, to improve local en-vironmental conditions and quality of life forlow-income residents. The analysis indicatedthat residents in low-income settlementsacross the two cities are likely to participatein communal action around environmentalimprovements and that, more importantly,there remains individual commitment tocommunal goals through time (as indicatedby activities throughout the rounds of thegames) and across place (as indicated by thesimilar patterns across neighbourhood partic-ipants in each city). There are specific policylevers that should be investigated further,such as mixed-gender groups in designing,

promoting and implementing communalprojects, and the relatively untapped poten-tial exhibited for communal action whenprojects are depicted as public goods withindividual and communal benefits. Com-munity participation will remain a criticalcomponent of all environmental managementpolicies throughout the developing world.Continued investigation of the nature andproblematic of communal action is necessaryfor such policies to be both cost-effective andappropriate.

Notes

1. Decisions within households concerning wa-ter and solid waste are usually made bywomen, especially in the developing world(Daniere and Takahashi, 1999), and as exten-sions of such domestic responsibilities,women are also often the instigators of com-munity cleaning (Milroy and Wismer, 1994).However, more research is needed to articu-late the linkages among social capital, gen-der, and environmental or infrastructuremanagement within neighbourhoods, es-pecially in rapidly developing urban contexts(Andreoni and Vesterlund, 2001; Croson andBuchan, 1999).

2. The social capital theorist Bourdieu (1985)makes clear the conflict and normalisationinherent in resources available through socialnetworks and groups.

3. A key feature of experiments, for example, isthat participants have some incentives to be-have selfishly because their behaviour islinked with potential earnings and, typically,more self-interested behaviour pays more.Notice this also means that behaving non-selfishly in an experiment is different fromsaying one would behave unselfishly on asurvey. For example, in social dilemmagames, it is materially costly to co-operatebecause free-riding always yields a higherpay-off. Hence, forgoing the free-rider’s in-crement means one is essentially willing topay a cost to co-operate and this cost is notincurred when filling out a survey. Con-versely, surveys are able to control the fram-ing or contextual issues of the situation muchmore effectively than experiments and canusually, therefore, be generalised to broaderpopulations. Furthermore, in terms of im-plementation, surveys may be administeredto all sorts of groups and communities, be-yond ‘captive’ students in economic experi-ments (Carpenter, 2002).


4. Few studies utilise the two methods togetherto study the same community or group al-though there are a number of reasons why itmight make sense to rely simultaneously onboth types of approach when evaluating theimpact of social capital in a population (Car-penter, 2002). Many researchers have arguedthat self-reported responses to behaviouralquestions having to do with social dilemmasituations (for example, questions on trust) insurveys do not match actual behaviour(Glaeser et al., 2000). On the other hand, thedata from experiments are criticised becauseof their questionable external validity (forexample, how can the results of experimentsconducted with students in a homogeneouslaboratory be generalised beyond theconfines of the university or student body?).

5. Details regarding the location and character-istics of these communities are availablefrom the authors.

6. Some questions differed in tone or exacttranslation because native speakers felt thatthe language used or the approach taken inEnglish or the other language was not appro-priate for that given context. Details regard-ing the survey and implementationprocedures can be found in Daniere et al.(2002).

7. Following the United Nations (1987), wedefined households to mean any group ofpeople who lived together and pooled theirresources to manage their day-to-day exist-ence.

8. See Ledyard (1995) for a survey of the re-sults from the voluntary contribution mech-anism. The instructions to participants areavailable from the authors upon request.

9. The pay-off function for each player is sim-ply

Payoffj � (10 � xj) �

2�xj

4

where, xj is individual j’s contribution to thepublic good (and a behavioural measure ofthe propensity to co-operate in a social di-lemma). This is a social dilemma becauseputting any money in the public good returnsonly half that amount to the participant; thismeans free-riding is the dominant strategy.But, if all participants allocate everything,they each receive 20 back which is largerthan the 10 they receive if each participantkeeps it all, so contributing is sociallyefficient.

10. The questions included items such as:

Q28: “If your household was short of food

and/or money, could you ask people in thisneighbourhood for help?”Q29: “Do most people try to be helpful totheir neighbours?”Q30: “Would you ask neighbours to keep aneye on your house when you are away?”Q31: “How do you feel about your next-doorneighbour?”.

11. 11. The trust index (trustind) is an additiveindex (range: 0–4) composed of the follow-ing items

Q28: If your household was short of foodand/or money, could you ask people in thisneighbourhood for help? If “yes”, thentrustind increases by one; if “no” or “don’tknow” or “refused”, then trustind remainsunchanged.Q29: Do most people try to be helpful totheir neighbours? If “yes”, then trustind in-creases by one; if “no” or “don’t know” or“refused”, then trustind remains unchanged.Q30: Would you ask neighbours to keep aneye on your house when you are away? If“yes”, then trustind increases by one; if “no”or “other”, then trustind remains the same.Q31: How do you feel about your next-doorneighbour? If participant responds, “likefamily,” then trustind increases by one; anyother response does not affect the value oftrustind.

12. This question was asked only of respondentswho said that they were aware of communityimprovement projects within the past year. Inthe case of Bangkok, some 357 respondents(or 71.4 per cent) of respondents claimed toknow of such projects compared with only314 respondents (or 62.8 per cent) of all HoChi Minh City respondents.

13. At the same time, however, we should notethat the monetised experimental games areheavily imbued with Western values andmay actually teach participants even moremistrust of academic researchers. Finding analternative way to conduct an experimentalgame that relies on more culturally appropri-ate values may be the best way to proceed.We are indebted to an anonymous reviewerfor this insight.

References

ALRECK, P. L. and SETTLE, R. B. (1995) The Sur-vey Research Handbook, 2nd edn. Boston, MA:Irwin McGraw-Hill.

AMIN, R., BECKER, S. and BAYES, A. (1998) NGO-promoted microcredit program and women’sempowerment in rural Bangladesh: quantitativeand qualitative evidence, Journal of Develop-ing Areas, 32(2), pp. 221–236.


ANDERSON, C. L., LOCKER, L. and NUGENT, R.(2002) Microcredit, social capital and commonpool resources, World Development, 30(1),pp. 95–105.

ANDREONI, J. and VESTERLUND, L. (2001) Whichis the fair sex? Gender differences in altruism,The Quarterly Journal of Economics, 116(1),pp. 293–312.

BARR, A AND KINSEY, B. (2002) Do men reallyhave no shame. Working Paper WPS2002.05,Center for the Study of African Economics.

BOURDIEU, P. (1985) The forms of capital, in: J.G. RICHARDSON (Ed.) Handbook for Theory andResearch for the Sociology of Education,pp. 241–258. New York: Greenwood.

BROWN-KRUSE, J. AND HUMMELS, D. (1993) Gen-der effects in public goods contribution: doindividuals put their money where their mouthis?, Journal of Economic Behavior and Organi-zation, 22, pp. 255–267.

CARPENTER, J. (2002) Measuring social capital:adding field experimental methods to the ana-lytical toolbox, in: S. RAMASWAMY, J. ISHAM

and T. KELLY (Eds) Social Capital, EconomicDevelopment and the Environment, pp. 119–137. Cheltenham: Edward Elgar Publishers.

COIT, K. (1998) Housing policy and slum upgrad-ing in Ho-Chi-Minh City, Habitat Inter-national, 22(3), pp. 273–280.

COLEMAN, J. (1990) Foundations of Social The-ory. Cambridge, MA: Harvard UniversityPress.

CRANE, R. and DANIERE, A. (1996) Measuringaccess to basic services in global cities: de-scriptive and behavioral approaches, Journal ofAmerican Planning Association, 62, pp. 203–221.

CROSON, R. and BUCHAN, N. (1999) Gender andculture: International experimental evidencefrom trust games, American Economic Review,89, pp. 386–391.

DANIERE, A. and TAKAHASHI, L. M. (1997) En-vironmental policy in Thailand: values, atti-tudes, and behavior among the slum dwellers ofBangkok, Environment and Planning C, 15,pp. 305–327.

DANIERE, A. and TAKAHASHI, L. M. (1999) En-vironmental behavior in Bangkok, Thailand: aportrait of attitudes, values, and practices,Economic Development and Cultural Change,47(3), pp. 525–557.

DANIERE, A., TAKAHASHI, L. and NARANONG, A.(2002) Social capital, networks and communityenvironments in Bangkok, Thailand, Growthand Change, 33(4), pp. 453–484.

DAVERN, M. (1997) Social networks and econ-omic sociology: a proposed agenda for a morecomplete social science, The American Journalof Economics and Sociology, 56, pp. 287–302.

DAVIS, D. and HOLT, C. (1993) Experimental Eco-

nomics. Princeton, NJ: Princeton UniversityPress.

DEFILIPPIS, J. (2001) The myth of social capital incommunity development, Housing Policy De-bate, 2(4), pp. 781–806.

DRAKAKIS-SMITH, D. and DIXON, C. (1997) Sus-tainable urbanization in Vietnam, Geoforum,28, pp. 21–38.

DURLAUF, S. (2002) On the empirics of socialcapital, The Economic Journal, 112, pp. F459–F479.

FERBER, M. A. (1995) The study of economics: afeminist critique, American Economic Review,85, pp. 357–361.

GAINSBOROUGH, M. (2002) Understanding com-munist transition: property rights in Ho ChiMinh City in the late 1990s, Post-CommunistEconomies, 14, pp. 227–243.

GERTLER, M. (1997) The invention of regionalculture, in: R. LEE and J. WILLS (Eds) Geogra-phies of Economies, pp. 47–58. London: Ed-ward Arnold.

GLAESER, E., LAIBSO, D., SCHEINKMAN, J. A., andSOUTTER, C. L. (2000) Measuring trust, Quar-terly Journal of Economics, 65, pp. 811–846.

GRAHAM, P. E. and MANNING, L. M. (2000)Women’s poverty and entrepreneurship: repli-cating microloan programs in the developedworld, Journal of Economics, 26(2), pp. 23–33.

GROOTAERT, C. and BASTELAER, T. VAN (Eds)(2001) The Role of Social Capital in Develop-ment: An Empirical Assessment. Cambridge:Cambridge University Press.

ISHAM, J. and KAKHONEN, S. (2001) What deter-mines the effectiveness of community-basedwater projects: evidence from Central Java,Indonesia, in: C. GROOTAERT and T. VAN BASTE-

LAER (Eds) The Role of Social Capital in Devel-opment: An Empirical Assessment, pp. 123–154.Cambridge: Cambridge University Press.

KIM, W. B., DOUGLASS, M. and HO, K. C. (1997)Culture and the City in East Asia. Oxford:Oxford University Press.

KOLLOCK, P. (1998) Transforming social dilem-mas: group identity and cooperation, in: P.DANIELSON (Ed.) Modelling Rationality, Moral-ity and Evolution, pp. 186–210. Oxford, UK:Oxford University Press.

LEDYARD, J. (1995) Public goods: a survey ofexperimental research, in: J. KAGEL and A.ROTH (Eds) The Handbook of ExperimentalEconomics, pp. 111–194. Princeton, NJ:Princeton University Press.

MAYER, M. (2003) The onward sweep of socialcapital: causes and consequences for under-standing cities, communities and urban move-ments, International Journal of Urban andRegional Research, 27(1), pp. 110–132.

MILROY, B. M. and WISMER, S. (1994) Communi-


ties, work and public/private sphere models,Gender, Place and Culture, 1(1), pp. 71–90.

NARAYAN, D. and PRITCHETT, L. (1999) Cents andsociability: household income and social capitalin Rural Tanzania, Economic Development andCultural Change, 47, pp. 871–888.

ONYX, J. and BULLEN, P. (2000) Measuring socialcapital in five communities, The Journal ofApplied Behavioral Science, 36, pp. 23–42.

OSTROM, E., GARDNER, R. and WALKER, J. (1992)Covenants with and without a word: self-gover-nance is possible, American Political ScienceReview, 86, pp. 404–417.

OSTROM, E., GARDNER, R. and WALKER, J. (1994)Rules, Games and Common-pool Resources.Ann Arbor, MI: University of Michigan Press.

PARENTEAU, R. (1997) Habitat et Environnementau Vietnam: Hanoi et Ho Chi Minh-Ville. Paris:Editions Karthala.

PORTES, A. (1998) Social capital: its origins andapplications in modern sociology, Annual Re-view of Sociology, 24, pp. 1–24.

PULIDO, L. (1998) Community, place, and ident-ity, in: J. P. JONES, H. NAST and S. ROBERTS

(Eds) Thresholds in Feminist Geography,pp. 11–28. Lanham, MD: Rowman & Lit-

tlefield.PUTNAM, R. (1993) Making Democracy Work:

Civic Traditions in Modern Italy. Princeton,NJ: Princeton University Press.

PUTNAM, R. (2000) Bowling Alone: The Collapseand Revival of American Community. NewYork: Simon and Schuster.

SETCHELL, C. (1992) Final report of the GreaterBangkok slum housing market study. Preparedfor the Regional Housing and Urban Develop-ment Office, United States Agency for Inter-national Development, Bangkok, Thailand.

SOLOW, J. and KIRKWOOD, N. (2002) Group ident-ity and gender in public goods experiments,Journal of Economic Behavior and Organiza-tion, 48, pp. 403–412.

UNITED NATIONS (1987) Statistical Yearbook.New York: United Nations, Statistical Office.

WHITE, G. F., BRADLEY, D. J. and WHITE, A. U.(1972) Drawers of Water: Domestic Water Usein East Africa. Chicago, IL: University ofChicago Press.

WOOLCOCK, M. (1998) Social capital and econ-omic development: toward a theoretical syn-thesis and policy framework, Theory andSociety, 27, pp. 151–208.

Social Capital and Trust in South-east Asian Cities

Documents

Transcript of Social Capital and Trust in South-east Asian Cities