Consumer Savings Psychology

Consumer Savings Psychology: A Structural Equation Modelling Approach And Analysis Of Vietnamese Saving And Spending Behaviour William James Ardrey, University of Western Australia Abstract As economies grow, consumer spending behaviour tends to evolve much more rapidly compared with investing and savings behaviour However, after consumption brings consumers and households to a certain level of income and perceived life quality, consumption decisions become considered concurrent to preferences around investing and savings choices This is a study, conducted within Vietnam’s rapidly emerging economy in recent years, which uses marketing research methods to collect data on the relationship between consumer psychology and savings patterns Based on marketing and economic psychology literature, a model of savings behaviour is hypothesized, followed by data collection and analysis from over 650 Vietnamese consumers Following data analyses using the Partial Least Squared methodology within the Structural Equation Modelling paradigm, a more parsimonious model is presented and validated to correctly explain and predict savings behaviour in a Southeast Asian transition economy Conclusions and limitations of the study are then discussed, and the more parsimonious model evaluated as a basis for bank marketing strategy development Introduction Despite an ambitious reform effort, the formal banking systems in Vietnam and neighbouring Southeast Asian Transition Economies (SEATE’s) continue to struggle with competition from foreign banks, informal and unlicensed credit and saving schemes, and with non bank savings such as hoarding US dollars, gold or real estate holdings outside of the banking system Below, data collected as part of an investigation into the general problem of low savings in Vietnam are analysed from a quantitative perspective The quantitative methodology was built subsequent to earlier library immersion, input from SEATE bank experts, and instrument development and pre-testing Analysis of results facilitates the investigation of hypotheses relating to psychological profiles of Vietnamese consumers likely to invest savings in banks (Sikor 1996) Literature review Economic reforms alone not guarantee that consumers will save in banks During the 1997-1999 Asian Financial Crisis, many SEATE policy makers relied on advice from the International Monetary Fund and other purely economic theories to shock emerging banking systems to compete under market forces (Antonides 1998) Unfortunately, such reforms have failed in practice to promote a market orientation within the state-dominated banking systems in Vietnam and neighbouring countries, thus consumers actively avoid banking institutions as a repository for the savings created by economic liberalization Mobilising domestic savings from retail consumers depends increasingly on understanding the wants and needs of these consumers and designing investment services to meet these needs In SEATE’s such as Vietnam, these consumers are likely to reside in countryside villages In underdeveloped countries, attempts at converting a communist banking system to a market system to facilitate the movement of capital and finance has created numerous political, ideological and economic conflicts (Tarling 1992; Templer, 1999) ANZMAC 2005 Conference: Consumer Behaviour 13 While considerable effort has been invested into understanding why SEATE consumers—especially Vietnamese consumers—spend money, a gap in the literature clearly exists when seeking to understand consumer savings in these challenging environments A neglected but important possible way out of this impasse for Vietnam and its neighbours can be found in employing marketing principles and marketing research techniques (Meidan 1996) Advances in marketing and consumer behaviour can assist in understanding the behaviours of consumers involved in the process of investing the surplus savings made possible by economic renovation and transition (termed “doi moi” in Vietnam) (Lin and Chen 1997) For policy makers, bankers and individual consumers, a marketing perspective most frequently conceptualised within the ambit of “macro marketing” or social marketing can offer a necessary supplement to economic reforms targeted at the banking sector, its customers and regulators (Baker and Haslem 1973; 1974a; 1974b; Brailsford and Heaney 1997; Lim 1996; Lin 1992) Marketing and consumer behaviour in transition economies has received considerable attention, particularly in SEATE’s, as the process of economic development coupled with the recent financial crises places immense pressure on the socio-political and financial institutions in an already turbulent environment (Shultz and Ardrey 1997; Shultz, Pecotich, and Le, 1994) Marketing activities, while encouraged by the transition process in both Eastern Europe and SEATE’s, are undertaken in Asia within an increasingly challenging environments as transition economies seek to catch up with the rest of the developed world A recognised area of special importance in this context is that of financial services marketing, and the marketing of savings products at a time when most marketing initiatives tend to promote spending and consumption rather than savings Financial services involve activities, benefits and satisfactions connected with the sale of money which offer users and customers financial-related value (Meidan, 1996) Financial services face a number of unique challenges associated with their complexity, which complicates retail financial services marketing Bank marketing problems are compounded in transition economies by the legacy of central planning, unstable national currency, low consumer knowledge, lack of developed banking and marketing infrastructure, as well as an over-reliance on macro economic policy combined with poor understanding of customer needs (Ennew, Wright, and Kirnag, 1996) Successful bank marketing in SEATE’s such as Vietnam is also hindered by low marketing skills, as Vietnam’s financial sector is the last to receive attention under “doi moi” style economic reforms Indeed, during the 1997-1999 period when this study was conducted, it became illegal to publicly criticise the financial reform process and the conduct of the stateowned banking system in Vietnam (Templer, 1999) Successful marketing in the financial sector is also hindered by consumers unsure about how banks can benefit their local situations Conceptual development The Model A (attached) hypothesises four independent variables include risk, knowledge, commitment and trust which drive savings behaviour, at four levels, namely the national, banking system, workplace and local levels The dependent variable is intention or choice of savings with a bank (Fishbein 1967; Raaij and Gianotten 1990) Based on research by Winton & Nelson (1978), financial consumers can be profiled based on ‘intent to bank,’ and their investment choices can be predicted and explained Thus, based on this research, it is hypothesized in Model A below that the independent variables of trust, risk perception, commitment and financial knowledge help create profiles of consumers holding banking ANZMAC 2005 Conference: Consumer Behaviour 14 intention Dependent variables involve investment choices, based on psychological decision styles (Baker and Hazlem 1974) An alternative model was also possible, and indeed a more parsimonious model (Model B) could be generated to better explain explicit and latent relationships One of the key differences involves the dependent variable, which details the interaction between investor intent and investment choice In Vietnam, the banking system is stratified among government owned state banks (eg VietComBank), government linked joint venture banks (eg Indovina Bank), foreign banks (eg Citibank) as well as smaller ‘joint stock bank’ entrepreneurial competitors, all of whom hold banking licenses Importantly, Southeast Asian business systems also experience informal or village run banks or savings cooperatives which are not licensed (Templer 1999) When banking systems become unstable, due to inflation, bank runs or other uncertainties, ‘informal’ savings methods are often adversely impacted, and carry no insurance or other protection In the 1996 period, Vietnam’s financial sector experienced exactly this sort of instability Procedurally, a questionnaire based on the economic psychology literature, a questionnaire was developed to measure these variables (Evans, Holcomb and Chittenden 1989; Durvasala, Lysonski and Andrews 1993; Everdingen and Raaij 1998; Faber and O’Guinn 1988, Morgan and Hung 1994a; 1994b; McGoldrick and Greenland 1994) It was administered in two waves to consumers in Vietnam in the 1998-1999 period to collect 654 valid responses Appropriate market research methods were employed to deal with translation, cross cultural and related issues, as well as to protect informants from any negative impact from government or banking organisations (Brislin 1973; 1979; 1980; Moorman, Zaltman and Deshpande 1992; Shore and Vankatachanlam 1995; Vandermehden 1974; Verbrugge 1995; 1999; Vu and Speece 1996) Upon collection, validation of the instrument was undertaken To validate the scales used to measure these constructs, co-efficient alpha was again used for both independent variables, and dependent variables, described below (Nunnally, 1967) Reliability of measures was found to be excellent in most cases, above the 70 required by most market research literature (Zikmund, 1998) National risk (.85), Banking risk (.84) and Workplace risk (.78) exhibited high co-efficient alpha scores Local or individual investor risk tolerance was somewhat problematic (.56), but under the circumstances considered acceptable to proceed Commitment measures were also acceptable at all levels National commitment (.84), Banking commitment (.80) and Workplace commitment (.81) exhibited high reliability Individual commitment (.68) was again lower than desired, but considered acceptable Trust measures were also satisfactory from a reliability standpoint National level trust (.72), Banking trust (.72), Workplace Trust (.74) and Individual trust (.83) were all considered satisfactory As expected, based on the pre-test, knowledge measures also performed acceptably Subjective Knowledge reliability was determined to be satisfactory (.80), as was Objective Knowledge (.79) Because True-False items in the objective knowledge scale were dichotomous, the Kudder-Richardson Alpha was used for reliability verification (Nunnally, 1967) Dependent variables were grouped into four categories based on factor analysis first of the perceived risk of, and intention to use, various banking products and again on the measurement of intention to use such products Factor analysis is generally used to summarize information contained in a large number of variables into a smaller number of factors, and factor analysis is a commonly used tool to prove validity and to discern the ANZMAC 2005 Conference: Consumer Behaviour 15 underlying dimension of responses (Zikmund, 1998) Such factor analysis indicated the existence of four factors, wherein “banking” intention is contained in the first two factors of formal banking and foreign banking Based on work by Fishbein and others, the marketing and bank marketing literature has established intention as an acceptable predictor of final behaviour Data analysis Below, relationships hypothesised by Model A are tested, and a rival Model B emerges as a new model better supported by the data For data analysis, the Partial Least Squares approach of Structural Equation Modeling was used to test the relationship between the four independent variables (risk, trust, commitment and knowledge) and the dependent variables (bank intent, bank action) (Kroonenberg, 1990; Lohmoeller, 1989; Wold, 1981) In general, Structural Equation Modeling has advantages when evaluating service marketing relationships because within the same decision context, a variance-covariance matrix can be used to estimate all direct and indirect relationship between variables It is also a technique appropriate to consumer profiling in cross cultural research (Durvasula et al., 1993; Joreskog, 1971) However, due to the stringent distributional assumptions needed by typical LISRELbased, for this study the Partial Least Squares (PLS) estimation procedure was ultimately selected to validate theoretical hypotheses Detailed data analysis will be presented separately, but basically indicated that Model A—and its hypothesized linear and direct relationships—were unsupported, and Model B proved superior LISREL was used as a program easily applied to PLS estimation (Joreskog and Sorbom, 1993) This LISREL program permits data analysis and parameter estimates to be undertaken easily (Bagozzi and Yi, 1988) It also grants the researcher the ability for more restricted testing, for example examining multiple inter-related dependence relationships, and including observations from the factor model into the causal model (Chin, 1998b; Chin and Fry, 1998; Kroonenberg, 1990) The PLS procedure was deemed appropriate in this case given this banking research is the first work of its kind done in the Consumer Behaviour field in Vietnam, and because of the difficulty of measuring certain constructs in Vietnam described in the literature (Vu and Speece, 1996) These difficulties led to some violations of the violations of the strict distributional assumptions of some dependent variables required by “harder” models (Falk and Miller, 1992) PLS also held many advantages for three reasons First of all, it was possible to create formative indicators to determine the existence of latent variables Latent variables can be the effects, as much as the cause, of formative indicator variables Second, A PLS model is formally specified by two sets of variables An outer model, wherein relationships between the latent and manifest variables are specified, is one set of linear relations wherein the interpretation is similar to that of principal component loadings (in the reflective case), or of canonical correlations (in the formative case) An inner model is another set of relationships, where hypothesized relationships between the latent variables are specified and whose interpretation is similar to interpretations for standardized regression co-efficients (Fornell and Cha, 1994) The PLS computer program, as implemented by Lohmoller (1981), was further developed by Chin and Fry (1998) as a tool to systematically and concurrently evaluate the properties of the outer and inner models, including relationships among latent and formative variables PLS can be used for theory confirmation, and also to develop new theoretical models via trimming and expansion procedures From this analysis, a more parsimonious Model B was determined to be indicative of the data as analysed ANZMAC 2005 Conference: Consumer Behaviour 16 Results and discussion Mean savings behaviour in government banks were noted to be far less than the other savings options, especially local national banking choices This is described in Table Table 1: Mean Savings Intention Intended Choice Formal Nat’l Bank Foreign Bank Workplace Bank Local Bank N 654 654 654 654 Mean (US$ Saved) 331.44 59.85 68.05 532.62 Standard Deviation 308.63 173.95 171.35 311.81 Moreover, while Model A was determined to be a useful and general guide for the study, a more parsimonious Model B Clearly, both Model A and Model B make the initial contribution involving the extension of marketing to savings as well as spending behaviours In addition to presenting a parsimonious and valid model of savings behaviour applicable to an emerging economy, this study also contributes to the literature through demonstrating the benefit of “soft modeling” included in this approach to LISREL estimation procedures Thus, this further develops the assertion by Wold (1981) recommending avoidance of “hard” assumptions of normality and large sample sizes as a general technique for estimating path models, in this study relationships in exploratory research involved latent constructs indirectly observed by multiple indicators This analytical approach circumvents the multivariate assumptions of normality inherent in experimental testing, validating that PLS is best suited to such exploratory research, where relationships are analysed during the early stages of investigation where measurement scales, sample sizes and residual distributions all represent challenges to the researcher (Wold, 1981) ANZMAC 2005 Conference: Consumer Behaviour 17 Appendix Model A: Vietnam Consumer Behaviour for Savings (Model A) Psychological Factors Subj Knowledge (5A)* Obj Knowledge (5B)* National Risk (2B3)* Bank Risk (2B4)* H1+ Knowledge Saver H15+ H20+ H21+ Foreign Action H16+ H3+ H4+ H17+ Formal Natl Intent H22+ Formal Natl Action Patriotic Choice H5+ Indiv Risk (2B1)* Foreign Intent Foreign Choice H2+ Patriotic Saver Work Risk (2B2)* Savings Choices Investment Profile Investor Profile H23+ H24+ Formal Natl Account (6C)* H6+ H7+ H8+ Natl Commit (3A)* Bank Commit (3C)* H25+ H9+ H10+ Group Work Saver H18+ Group Work Choice Work Commit (3B)* H26+ H11+ Indiv Commit (3D)* National Trust (4A)* H13+ H27+ H12+ Bank Trust (4A)* Local Saver H19+ H28+ Local Choice Workplace Intent Workplace Action Local Intent Local Action Work Trust (4A)* Indiv Trust (4A)* H14+ ANZMAC 2005 Conference: Consumer Behaviour H29+ Informal Account (6C)* 18 Appendix Model B: Vietnam Consumer Behaviour for Savings (Model B) Psychological Factors Subj Knowledge (5A)* Obj Knowledge (5B)* Bank Commit (3C)* H1+ Knowledge Saver H20+ H15+ Foreign Choice H2+ H6+ Savings Choices Investment Profile Investor Profile UH3+ UH6+ H21+ Foreign Action UH8+ H22+ H17+ Patriotic Choice Patriotic Saver Foreign Intent Formal Natl Intent H23+ Formal Natl Action UH1+ UH4+ Work Trust (4D)* H11+ Group Work Saver H18+ UH5+ Indiv Risk (2B1)* H12+ Local Saver H19+ H26+ Group Work Choice Workplace Action UH7+ Local Choice Local Action H28+ Indiv Trust (4C)* H14+ ANZMAC 2005 Conference: Consumer Behaviour 19 References Antonides, G., 1998 Psychology in Economics & Business Kluwer, London Bagozzi, R., and Yi, Y., 1988 On the Evaluation of Structural Equation Modelling Academy of Marketing Sciences 16, 1, 74-94 Baker, H., and Haslem, J., 1973 Information Needs of Individual Investors Journal of Accountancy, 136(Nov), 64-69 Baker, H., and Haslem, J., 1974a The Impact of Investor Socio Economic Characteristics on Risk & Return Preferences Journal of Business Research, 2(10), 469-476 Baker, H., and Hazlem, J., 1974b Towards the Development of Client Specific Valuation Models Journal of Finance, 29(Sept), 1255-1263 Brailsford, T., and Heaney, R., 1997 Investments: Concepts & Applications in Australia Harcourt, Sydney Brislin, R., 1979 Back Translation for Cross National Research Journal of Cross National Research, 1(1), 185-216 Brislin, R., 1980 Translation and Content Analysis of Oral and Written Material In H Triandis & L Berry (Eds.), Handbook of Cross Cultural Psychology (Vol 2, ) Allyn, Boston Brislin, R., Lonner, W., and Thordike, R., 1973 Cross Cultural Research Methods John Wiley, NYC Chin, W., 1998a PLS Approach to SEM Erlbaum, New Jersey Chin, W., and Fry, T., 1998b PLS Graph 2.9.03.04 University of Calgary Faculty of Management, Calgary Durvasula, S., Lysonki, S., and Andrews, J., 1993 Cross Cultural Generalisability of a Scale for Profiling Consumers Decision Making Styles Journal of Consumer Affairs, 27(1), 33-75 Ennew, C., Wright, M., and Kirnag, K., 1996 The Development of Bank Marketing in Eastern Europe: The Case of Slovakia Services Industry Journal 16, 4, 443-457 Evans, D., Holcomb, J., and Chittenden, W., 1989 The Relationship between Risk Return Preferences and Knowledge in Experimental Financial Markets Journal of Behavioural Economics, 18(1), 19-40 Everdingen, F.V., and Raaij, F.V., 1998 The Euro: How Consumers Perceive the Single European Currency Paper presented at the European Marketing Association Conference, Stockholm ANZMAC 2005 Conference: Consumer Behaviour 20 Faber, R., and Oguinn, T., 1988 Compulsive Consumption and Credit Abuse Journal of Consumer Policy, 11(1), 97-109 Fishbein, M., 1967 The Relationship between Belief, Attitude, Intention and Behaviour Readings in Attitude Theory and Measurement, 199-223 Fornell, C., and Cha, J., 1994 Partial Least Squares In Bagozzie, R., (Ed.), Advanced Methods in Market Research Basil Blackwell, Oxford Joreskog, K., 1971 Simultaneous Factor Analysis in Several Populations Psychometrika 36, 4, 409-426 Kroonenberg, P., 1990 Review of Latent Variable Modelling with PLS Journal of the American Statistical Society 9, 909-915 Lim, L., 1996 The Evolution of the Southeast Asian Business System Journal of Asian Business, 12, 51-74 Lin, W., and Chen, T., 1997 Management of Developing Economies in Economic Transition Praeger, Westport Lin, Z., 1992 The Banking Structure of China In Kaufman, B., (Ed.) Banking Structure in Major Countries Kluwer, Amsterdam Lohmoeller, J., 1989 Latent Variable Modelling with PLS Physica, Heidelberg McGoldrick, J., and Greenland, S., 1994 Retailing of Financial Services McGraw Hill, New York Meidan, A., 1996 Marketing Financial Services Macmillan, London Morgan, R., and Hunt, S., 1994a The Commitment Trust Theory of Relationship Marketing Journal of Marketing, 58(6), 20-38 Morgan, R., and Hunt, S., 1994b Relationship Marketing in an Era of Network Competition Marketing Management, 3(1), 18-28 Moorman, C., Zaltman, G., and Deshpande, R., 1992 Relationships between Providers and users of Market Research: The Dynamics of Trust within and between Organisations Journal of Marketing Research, 29(3), 314-328 Shore, B., and Vankatachanlam, A., 1995 The Role of National Culture in Systems Analysis and Design Journal of Global Information Management, 3(3), 5-14 Shultz, C.J., and Ardrey, W.J., 1997 Asia’s Next Tiger: Vietnam is fraught with Promises & Perils for Marketers Marketing Management (Winter) 27-35 Shultz, C.J., Pecotich, A., and Le, K., 1994 Changes in Marketing and Consumption Behaviour in the Socialist Republic of Vietnam Research in Consumer Behaviour 7, 1, 225257 ANZMAC 2005 Conference: Consumer Behaviour 21 Sikor, 1996 Economic and Environmental Dynamics of Reform in Vietnam, Asian Survey 6, 1, 601-617 Tarling, N., 1992 The Cambridge History of Southeast Asia Cambridge University Press, Cambridge Templer, R., 1999 Shadows & Wind, A View of Modern Vietnam Penguin, New York Vandermehden, F., 1974 Southeast Asia: The Legacy of Nationalism and Colonialism 19301970 Thames Hudson, London VanRaaij, F., and Gianotten, H., 1990 Consumer Confidence, Savings and Credit Journal of Economic Psychology, 11(1), 269-290 Verbrugge, J., 1999 Event Studies in Finance and Marketing: A Roadmap for Finance and Marketing Managers The Journal of Market Focused Management, 2(2), 10-24 Verbrugge, J., Whidbee, D., and Friedmann, R., 1995 Marketing Financial Services to All Levels of Affluence Psychology and Marketing, 12(8), 703-719 Vu, T., and Speece, M., 1996 Marketing Research in Vietnam Journal of International Marketing and Marketing Research, 21(3), 145-160 Winton, D., and Nelson, F., 1978 Launching New Financial Services to Consumers Journal of the Market Research Society, 20(1), 30-39 Wold, H., 1981 Model Construction and Evaluation when Theoretical Knowledge is Scarce, In Kimenta, J., (Ed.) Evaluation of Econometric Models Academic Press, NYC Zank, N., 1991 Reforming Financial Systems Greenwood, Westport, Connecticut ANZMAC 2005 Conference: Consumer Behaviour 22 ... invested into understanding why SEATE consumers—especially Vietnamese consumers—spend money, a gap in the literature clearly exists when seeking to understand consumer savings in these challenging environments... (Meidan 1996) Advances in marketing and consumer behaviour can assist in understanding the behaviours of consumers involved in the process of investing the surplus savings made possible by economic... analysed ANZMAC 2005 Conference: Consumer Behaviour 16 Results and discussion Mean savings behaviour in government banks were noted to be far less than the other savings options, especially local