The role of perceived quality of personal service in influencing trust and satisfaction with banks Anders Carlander, Ph.D.a,*, Amelie Gamble, Ph.D.a, Tommy Gärling, Ph.D.a, Jeanette Carlsson Hauff, Ph.D.b, Lars-Olof Johansson, Ph.D.a, Martin Holmen, Ph.D.b aDepartment of Psychology, University of Gothenburg, P.O. Box 100, S-405 30, Gothenburg, Sweden bSchool of Business, Economics and Law, University of Gothenburg, P.O. Box 600, S-405 30, Gothenburg Sweden Abstract Trust is of paramount importance to banks. Previous research has shown that trust increases with repeated personal contacts. We investigate if this applies to the customer-employee relationship in banks. Data from an on-line survey of 293 customers of Swedish retail banks are used to construct indicator measures. By means of structural equation modeling we find that trust in the bank is influenced by perceived quality of personal service through employees’ perceived competence, perceived benevolence, and perceived transparency, and that satisfaction with the bank is influenced by perceived quality of personal service through perceived competence, perceived benevolence, perceived transparency, and trust. © 2018 Academy of Financial Services. All rights reserved. JEL classification: M Keywords: Retail banking; Personal service; Satisfaction; Trust 1. Introduction A trend during the last decades is a declining trust in other people as well as institutions (Twenge, Campbell, and Carter, 2014). Following this trend, trust in banks has likewise declined since the 1970s, further exaggerated by the financial crisis in 2008 (Edelman, 2017; Gallup, 2013). If trust is declining in banks, an important question is what role trust has for consumers’ financial decisions. The financial sector is often described as a “trust business” because of the * Corresponding author. Tel.: �46-31-786-30 92; fax: �46-31-786 46 28. E-mail address: anders.carlander@gu.se (A. Carlander) Financial Services Review 27 (2018) 83-98 1057-0810/18/$ – see front matter © 2018 Academy of Financial Services. All rights reserved. credence qualities of the offered services, in conjunction with the frequently long time span until delivery of benefits. As a consequence, it has been argued that trust is of paramount importance for the consumer of financial services (Sekhon, Ennew, Kharouf, and Devlin, 2014). Financial decisions have increased in importance to individuals—a trend that has aug- mented the need for research that addresses individuals’ ability to make such decisions (Gough and Niza, 2011). Given the complexity of many financial decisions and the often replicated finding that individuals perceive they lack adequate knowledge (e.g., Lusardi and Mitchell, 2007), their coping strategies have naturally become an important focus. One strategy preferred by many customers is to personally interact with bank employees to obtain financial advice to compensate their lack of knowledge (e.g., Hermansson, 2015). It has been demonstrated that trust increases with repeated personal contacts (e.g., Jones and George, 1998; King-Casas et al., 2005), and we assume that this also applies to the relationship between bank customers and bank employees. Based on other research showing that attitudes towards companies are influenced by and influence attitudes towards employees (cf. Plitt, Savjani, and Eagleman, 2015), a plausible further assumption is that customers’ trust in a bank changes with their trust in its employees with whom they interact. Our aim in this study is to attempt to empirically demonstrate that the quality of personal service offered by bank employees influence trust in their bank through the most common determinants of trust. Hence, as detailed below, we investigate whether perceived quality of personal service influences trust in banks through bank employees’ perceived competence, perceived benevolence, and perceived trans- parency. Because another important question is whether personal service to some extent affects satisfaction with banks as previous research has shown (Levy and Hino, 2016), we also inves- tigate the possibility that this relationship is mediated by trust that in previous research has been shown to have a direct effect on satisfaction (Geyskens, Steenkamp, and Kumar, 1999). If personal service is important for bank customers’ trust, it implies access to an important tool for increasing trust or counteracting a declining trust. Training of bank employees is one avenue to achieve this—preferably by explicitly targeting trust-building factors such as competence and benevolence. Transparent communication with customers is yet another practical tool, proven in previous research to be effective (Anderson and Weitz, 1992). The role of personal service for trust may also be important to take into account in digital contexts, in particular when considering that younger people are more skilled using new technology and less likely to want a personal relationship with the bank (KPMG Nunwood, 2016). The remaining of this article is organized as follows. First, we review relevant previous research on trust, personal service, and satisfaction. Then we propose our hypotheses. After this we describe the method and results of an internet survey to bank customers. Finally, we discuss the survey results and implications for theory and practice. 2. Review of previous research 2.1. Trust Trust is a multifaceted construct with a multitude of different definitions (Kantzberger and Kuntz, 2010). A common feature is that trust is associated with a willingness to depend on 84 A. Carlander et al. / Financial Services Review 27 (2018) 83-98 another party (Hong and Cha, 2013; Sekhon et al., 2014). One implication is that people substitute lack of knowledge for trust in experts (Carlander, 2015; Dia, 2011; Siegrist and Cvetkovich, 2000) such that trust will play the role of coping with uncertainty (Colquitt et al., 2012). Knowledge of determinants of trust is essential for banks to proactively address the issue of increasing trust through personal service. Conceptualizations of trust determinants draw on variants of competence, benevolence, and integrity (e.g., Ennew, Kharouf, and Sekhon 2011; in their study including also shared values and communication). In this study we similarly propose that competence, benevolence, and transparency are primary determinants of customers’ trust in banks, and that these determinants are all affected by personal service as follows. (1) Competence as a determinant of trust in banks refers to trustees’ knowledge and skill needed to complete a specific task (Malhotra and Lumineau, 2011). Financial services are by many customers perceived to be complex and difficult to understand. Bank employees’ competence may, therefore, be valuable to customers and contribute to building trust in the bank. (2) Benevolence is the expectation that the trustee acts in the trustor’s best interests (Sirdeshmukh, Singh, and Sabol, 2002). Similarly, integrity and shared values (e.g., Ennew, Kharouf, and Sekhon 2011) refer to a sense of moral, ethics, and consistency beyond the sole interests of the trustee. Benevolence may be important for customers’ trust in a bank since mass media coverage of high profits, high bonus payments to employees, and financial turmoil signal that banks are not acting in the customers’ best interests. (3) Transparency (e.g., Kanagaretnam, Mestelman, Naiar, and Shehata, 2010) may be defined as providing sufficient information to satisfy the requirements of a reasonable person (Rawlins, 2009). There has been a growing interest in the value of transparency and transparent communi- cation for attempts to increase trust (Rawlins, 2009). Transparent information is also believed to increase rational decision making in financial markets (Lusardi and Mitchell, 2007). 2.2. Personal service The importance of personal service has been claimed to increase in services with a high level of intangibility (Berry, 2000) as well as a people-intensive character (Sirdesmukh, Singh, and Sabol, 2002). Being a customized type of service high in credence qualities, financial services have further been described as part of a service industry where the relation between employees and customers is particularly important (Yim, Chan, and Lam, 2012). In line with this, dissatisfaction with in-branch banking is one factor that has been shown to push customers into alternative service distribution channels such as internet banking (Devlin and Yeung, 2003). It has been recognized that a high quality of personal service is instrumental for creating and maintaining satisfied customers (Crosby, Evans, and Cowles, 1990). Previous studies have also shown that the quality of personal service is a direct determinant of satisfaction (Krishnan et al., 1999) as well as of perceived service quality (Taylor and Baker, 1994). In Bloemer, de Ruyter, and Peeters (1998) perceived service quality was furthermore both directly and indirectly related to bank loyalty through satisfaction. We argue that an additional benefit of personal service is that it influences the determi- nants of customers’ trust in a bank, and through trust satisfaction with the bank. The 85A. Carlander et al. / Financial Services Review 27 (2018) 83-98 importance of face-to-face meetings when building trust has been documented (Howcroft, Hewer, and Durkin, 2003). In personal contacts employees may act in a way such that they are perceived to be competent by giving customers good advice, perceived to be benevolent by prioritizing customers’ interests, and be transparent by disseminating relevant information to customers in an understandable way. As a consequence, customers may both perceive that the banks’ personal service is of high quality and develop trust in the bank, which is one of the channels through which their satisfaction with the bank is influenced. 2.3. Satisfaction A positive relation between trust and satisfaction has been empirically verified in several meta-analyses (Geyskens, Steenkamp, and Kumar, 1999). Particularly relevant are the results from a survey focusing explicitly on banks, which found that trust is the most important factor affecting customers’ satisfaction (Chakravarty, Feinberg, and Widdows, 1997). The difference between the two constructs has been suggested to rely on the duration of the relation, with trust being a key variable in novel settings, and satisfaction more important in maintaining an existing relationship (Selnes, 1998). This is consistent with our conjectures that trust in a bank has the role of reducing uncertainty, and that trust subsequently is a determinant of satisfaction with the bank. Additionally, because trust is believed to increase with time (Jones and George, 1998), as well as become more relational and emotionally oriented over time (Lewicki and Bunker, 1995), customers� satisfaction would increase with length of their relation to the bank. Customers are, therefore, likely to be loyal to a bank that they trust, a result that has been empirically documented (Shainesh, 2012). 2.4. Hypotheses Previous research has shown that determinants of trust include competence, benevolence, and transparency (e.g., Ennew, Kharouf, and Sekhon, 2011). Our rationale for including these determinants of trust in banks is customers’ demand for competent advice by bank employees, customers’ demand for a benevolent advisor to prevent them from being cheated by the bank, and customers’ demand for transparency by the bank employees to facilitate accurate communication in contacts with the bank. Therefore, we propose Hypothesis 1: The quality of personal service increases trust in banks through its deter- minants perceived competence, perceived benevolence, and perceived transparency. A direct effect of quality of personal service on satisfaction with banks has been documented (Levy and Hino, 2016; Moin, Devlin, and McKechnie, 2015). However, we expect that this relationship will be weaker when measuring the indirect effect of trust. We thus expect that the relationship between quality of personal service and satisfaction is mediated by trust and its determinants perceived competence, perceived benevolence and perceived transparency. Hypothesis 2: Perceived quality of personal service has an indirect effect on satisfaction through trust and its determinants perceived competence, perceived be- nevolence, and perceived transparency. 86 A. Carlander et al. / Financial Services Review 27 (2018) 83-98 3. Study 3.1. Overview To test Hypotheses 1 and 2 an on-line survey questionnaire was administered to bank customers asking them questions that provide indicators of the constructs entailed by the hypotheses. We then use covariance-based structural equation modeling (SEM) to estimate the expected direct and indirect effects associated with the latent constructs perceived competence, perceived benevolence, perceived transparency, trust, and satisfaction (see Fig. 1). In SEM (see, e.g., Bollen, 1989; Kaplan, 2009) the regression coefficients of a theoret- ically specified system of linear regression equations are simultaneously estimated. This is made by comparing covariances between constructs derived from the regression equations to observed covariances computed from the indicators. Several measures of goodness-of-fit of the model are computed. 3.2. Method Questionnaire data were collected in an on-line survey administered in January 2013 to participants randomly selected from a customer database maintained by the Swedish bank Skandiabanken. Skandiabanken is an internet-only bank providing special services. Only a minority of participants (25.3%) indicated in the questionnaire that it was their main bank, whereas approximately equal numbers of the majority indicated that their main bank was one of the four largest banks in Sweden (Handelsbanken, Nordea, SEB, and Swedbank). Participants were invited by e-mail to answer a questionnaire that in pretests took about 10 min to complete. The e-mail explained the general purpose of the study, informed about the funders of the research (a governmental agency and Skandiabanken), and whom to contact to make enquiries. Participants were guaranteed confidentiality and were not prom- ised any monetary compensation. A reminder e-mail was sent to those who after three days had not replied. The invitations were sent to 4,968 potential participants of which 293 (5.9%) Fig. 1. The hypothesized direct effects of perceived quality of personal service on perceived competence, perceived benevolence, and perceived transparency, the direct effects of perceived transparency, perceived competence and perceived benevolence on trust, and the indirect effect of perceived quality of personal service on satisfaction through perceived competence, perceived benevolence, perceived transparency, and trust. (Solid lines represent the direct effects, broken lines the indirect effect.) 87A. Carlander et al. / Financial Services Review 27 (2018) 83-98 completed the questionnaire (28.0% women, age ranging from 21 to 79 years with a mean of 55.4 years). Sample descriptives obtained from the questionnaire are given in Table 1. A detailed description of the questionnaire is given in Carlander (2015). Briefly, partic- ipants were in the indicated order asked questions about frequency and types of bank contacts, quality of personal service, trust in and satisfaction with their main bank (defined as the bank with which they had the closest relation), trust in society and its institutions, Table 1 Sample descriptives (n � 293) Variable Value Sex (% women) 28.0 Mean age (years)a 55.4 Education (spouse) Elementary school (% degree) 6.1 (8.5) High-school (% degree) 23.9 (18.4) College (% degree) 69.6 (54.3) Employment (spouse) Employed (%) 60.4 (50.5) Self-employed/entrepreneur (%) 17.4 (11.3) Agriculturist (%) 0.3 (0.3) Student (%) 1.4 (0.7) Unemployed (%) 0.3 (1.4) Retired/house-wife, -man (%) 28.7 (20.8) Other type of occupation (%) 3.1 (1.0) Married/cohabiting (%) 81.6 Annual household income (SEK)b 200,000 or less (%) 1.4 201,000–350,000 (%) 7.8 351,000–500,000 (%) 12.6 501,000–650,000 (%) 19.8 651,000–800,000 (%) 15.0 801,000–950,000 (%) 13.0 More than 950,000 (%) 25.6 Disposal assets (cash, savings, investments) (SEK) 10,000 or less (%) 1.4 11,000–50,000 (%) 5.5 51,000–150,000 (%) 10.2 151,000–300,000 (%) 12.6 301,000–600,000 (%) 16.7 601,000–900,000 (%) 10.2 More than 901,000 (%) 34.8 Type of home House/garden apartment 63.1 Apartment/flat (%) 35.2 Other 1.4 Type of agreement Rented apartment (%) 14.7 Owned apartment (%) 21.8 Owned house (%) 60.8 Other 2.3 a National average in Sweden: 50.1 % women (Statistics Sweden, 2013), mean age 41.1 years (Statistics Sweden, 2012). Other national averages are given within brackets. b 1SEK was approximately 0.16 US$ and 0.11 EUR€ at the time of the study. 88 A. Carlander et al. / Financial Services Review 27 (2018) 83-98 financial literacy assessed through both self-reported knowledge and true/false responses to knowledge questions, and questions about socio-demographic variables. The model test uses only the questions about quality of personal service, trust, and satisfaction described below. We based our trust questions on those developed in a larger set of studies (van Raaij and van Esterik-Plasmeijer, 2017) aimed at measuring trust in the finance sector. We received from these authors (personal communication, November 2011) an earlier version of the questions that at that time had been used to measure trust in insurance companies. The questions we used asked participants to rate statements about their main bank. They had been translated, modified, and pretested in two pilot studies to obtain reliable indicators of the latent variables perceived competence, perceived benevolence, perceived transparency, and trust.1 Similarly phrased statements about perceived quality of personal service and satis- faction were constructed by us. The participants rated each statement on a seven-point Likert-type numeric scale ranging from 1 “totally disagree” to 7 “totally agree.” The statements (translated from Swedish) are given in Table 2. In the questionnaire the state- ments were presented in a counterbalanced order. 3.3. Results Firstly, we assess discriminant and convergent validity of the theoretical constructs (Hair et al., 2010). Secondly, we use covariance-based SEM (Bollen, 1989; Kaplan, 2009) to test the relationships between the latent constructs. We report goodness-of-fit of the model as well as estimates of standardized regression coefficients corresponding to total, direct and indirect effects as specified in the model. Because missing data are not permitted in the analyses, we conducted a preliminary test showing that the missing values were not systematic (Little’s MCAR test, c2 � 1376.47, df �1470, p � 0.960). Expectation maximization was, therefore, used for replacing the missing values. Table 2 reports means, standard deviations, skewness, kurtosis, and inter- correlations of the indicators. By means of the maximum likelihood method in IBM SPSS AMOS 21, we first tested a six-factor measurement model of the latent constructs satisfaction, trust, perceived compe- tence, perceived benevolence, perceived transparency, and perceived quality of personal service. This resulted in a marginally acceptable model fit: c2 � 377.84, df � 120, p � 0.001; normed fit index (NFI) � 0.92; comparative fit index (CFI) � 0.95; and root mean square error of approximation (RMSEA) � 0.09. The standardized indicator weights are given in Table 2. With one exception these are uniformly high. In contrast, a one-factor measurement model yielded an unacceptable fit, c2 � 691.56, df � 90, p � 0.001; NFI � 0.84; CFI � 0.85; and RMSEA � 0.15. Table 3 reports the correlations between constructs, composite reliability (CR), Cron- bach’s �s, average variance extracted (AVE), maximum shared squared variance (MSV), and average shared squared variance (ASV). As recommended by Hair et al., (2010) and Fornell and Larcker (1981), all CRs and �s are above 0.70. Furthermore, the recommended criterion for convergent validity (Hair et al., 2010) is fulfilled in that AVE is above 0.50 and CR is larger than AVE. However, the criterion for discriminant validity, that AVE should be larger than both MSV and ASV, is only approximately fulfilled. In the main diagonal of the 89A. Carlander et al. / Financial Services Review 27 (2018) 83-98 T ab le 2 M ea ns (M ), st an da rd de vi at io ns (S D ), st an da rd iz ed sk ew ne ss , st an da rd iz ed ku rt os is , st an da rd iz ed w ei gh t co ef fic ie nt s (w ) of ra tin gs of st at em en ts (i nd ic at or va ri ab le s) on sc al es ra ng in g fr om 1 to 7, an d pr od uc t m om en t co rr el at io ns be tw ee n in di ca to rs L at en t va ri ab le In di ca to rs M SD Sk ew ne ss K ur to si s w Sa tis fa ct io n 1. I am sa tis fie d w ith m y ba nk 5. 28 1. 36 � 5. 25 1. 11 0. 89 2. I am pl ea se d w ith m y ba nk 5. 20 1. 41 � 5. 35 0. 76 0. 95 3. I am de lig ht ed w ith m y ba nk 5. 31 1. 36 � 6. 61 2. 61 0. 96 0. 96 T ru st 4. I tr us t m y ba nk 5. 50 1. 32 � 6. 18 1. 35 0. 81 5. I fe el gr ea t co nfi de nc e in m y ba nk 5. 05 1. 40 � 5. 46 1. 26 0. 93 6. M y ba nk is tr us tw or th y 5. 30 1. 28 � 4. 51 0. 76 0. 87 Pe rc ei ve d co m pe te nc e 7. M y ba nk un de rs ta nd s co m pl ex fin an ci al pr ob le m s 4. 74 1. 26 � 2. 47 0. 13 0. 71 8. M y ba nk is m ak in g de lib er at e de ci si on s 4. 61 1. 12 � 0. 59 � 0. 79 0. 70 9. M y ba nk is fin an ci al ly co m pe te nt 5. 26 1. 12 � 4. 48 1. 52 0. 77 Pe rc ei ve d be ne vo le nc e 10 . M y ba nk is be ne vo le nt to w ar ds m e 4. 26 1. 60 � 1. 53 � 2. 34 0. 84 11 . M y ba nk m ak es de ci si on s th at be ne fit m e 3. 95 1. 32 � 1. 52 � 0. 96 0. 87 12 . M y ba nk m ak es de ci si on s th at ar e be ne fic ia l fo r m e 4. 19 1. 51 � 1. 51 � 2. 38 0. 89 Pe rc ei ve d tr an sp ar en cy 13 . M y ba nk di sc lo se s fe es an d co nd iti on s th or ou gh ly 5. 00 1. 54 � 3. 99 � 1. 30 0. 63 14 . M y ba nk pr ov id es tr ut hf ul in fo rm at io n w he n as ke d 5. 43 1. 24 � 4. 35 � 0. 62 0. 82 15 . M y ba nk pr ov id es al l th e in fo rm at io n I as k fo r 5. 42 1. 26 � 5. 68 2. 43 0. 86 Pe rc ei ve d qu al ity of pe rs on al se rv ic es 16 . T he st af f at m y ba nk pr ov id es m e w ith th e ri gh t ty pe s of se rv ic es 4. 82 1. 49 � 2. 05 � 2. 96 0. 80 17 . I al w ay s fe el w el co m ed by th e st af f at m y ba nk 5. 54 1. 46 � 7. 09 2. 24 0. 77 18 . I re ce iv e ad eq ua te he lp fr om th e st af f at m y ba nk 5. 34 1. 35 � 5. 39 0. 77 0. 88 (c on ti nu ed on ne xt pa ge ) 90 A. Carlander et al. / Financial Services Review 27 (2018) 83-98 T ab le 2 (C on ti nu ed ) 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 2. 0. 86 3. 0. 85 0. 90 4. 0. 72 0. 73 0. 75 5. 0. 75 0. 83 0. 82 0. 77 6. 0. 64 0. 70 0. 76 0. 70 0. 80 7. 0. 48 0. 50 0. 49 0. 50 0. 54 0. 51 8. 0. 48 0. 56 0. 54 0. 44 0. 62 0. 58 0. 50 9. 0. 47 0. 53 0. 55 0. 49 0. 60 0. 65 0. 56 0. 52 10 . 0. 57 0. 66 0. 64 0. 58 0. 68 0. 65 0. 41 0. 46 0. 45 11 . 0. 57 0. 66 0. 66 0. 55 0. 67 0. 62 0. 42 0. 56 0. 45 0. 73 12 . 0. 56 0. 66 0. 67 0. 53 0. 69 0. 69 0. 48 0. 52 0. 49 0. 73 0. 77 13 . 0. 51 0. 54 0. 53 0. 52 0. 57 0. 57 0. 32 0. 42 0. 35 0. 52 0. 48 0. 47 14 . 0. 45 0. 54 0. 56 0. 58 0. 65 0. 72 0. 47 0. 49 0. 51 0. 54 0. 56 0. 64 0. 49 15 . 0. 50 0. 59 0. 63 0. 54 0. 67 0. 72 0. 45 0. 47 0. 52 0. 54 0. 54 0. 58 0. 51 0. 73 16 . 0. 63 0. 76 0. 72 0. 55 0. 65 0. 59 0. 46 0. 51 0. 49 0. 55 0. 65 0. 64 0. 47 0. 54 0. 58 17 . 0. 55 0. 63 0. 68 0. 50 0. 63 0. 63 0. 41 0. 40 0. 52 0. 54 0. 47 0. 56 0. 41 0. 52 0. 56 0. 60 18 . 0. 57 0. 66 0. 70 0. 57 0. 69 0. 70 0. 52 0. 45 0. 56 0. 56 0. 56 0. 63 0. 48 0. 70 0. 74 0. 70 0. 72 A ll co rr el at io ns ar e st at is tic al ly si gn ifi ca nt at p � 0. 00 1. 91A. Carlander et al. / Financial Services Review 27 (2018) 83-98 correlation matrix, the square roots of AVE are given for each construct. As may be seen, these are lower than the correlations with the other constructs for perceived competence, perceived transparency, trust, and perceived quality of personal services. An overlap between indicators of these constructs is hence implied. Adding a common latent factor improved the model fit marginally, c2 � 284.01, df � 102, p � 0.001; NFI � 0.94; CFI � 0.96; and RMSEA � 0.08. Still suggesting that a common latent factor may account for the covariance between the indicators, the standardized indicator weights were reduced for all the constructs (see MacKenzie and Podsakoff, 2012). The proposed structural model (Fig. 1) tested by the maximum likelihood method resulted in a marginally acceptable fit, c2 � 443.05, df � 127, p � 0.001; NFI � 0.91; CFI � 0.93; and RMSEA � 0.09. Two alternative models were also tested. A poor fit was obtained for the model positing that satisfaction is directly determined by trust, perceived competence, perceived benevolence, perceived transparency, and perceived quality of personal service, c2 � 1489.23, df � 130, p � 0.001; NFI � 0.70; CFI � 0.72; and RMSEA � 0.19, as well as for the other model, positing direct effects on satisfaction of trust and perceived quality of personal service, and indirect effects of perceived competence, perceived benevolence, and perceived transparency through trust, c2 � 1127.58, df � 130, p � 0.001; NFI � 0.77; CFI � 0.79; and RMSEA � 0.16. Table 4 gives standardized regression coefficient estimates of total, direct, and indirect effects. Bootstrapping was used to obtain bias-corrected 95% confidence intervals (as recommended by Hayes and Scharkow, 2013) for a total of 2,000 replicates. As expected (Hypothesis 1), a significant indirect effect is observed of perceived quality of personal service on trust through perceived competence, perceived benevolence, and perceived trans- parency. Perceived competence, perceived benevolence, and perceived transparency also have significant direct effects on trust. Furthermore, according to Hypothesis 2 there is a significant indirect effect on satisfaction of perceived quality of personal service through perceived competence, perceived benevolence, perceived transparency, and trust. The direct effect of perceived personal service on satisfaction is positive although not significant. Perceived competence, perceived benevolence, and perceived transparency have all signif- icant indirect effects on satisfaction through trust. Table 3 Composite Reliability (CR), Cronbach’s �, Average Variance Extracted (AVE), Maximum Shared Squared Variance (MSV), and Average Shared Squared Variance (ASV), and correlations between constructs with square root of AVE in the min diagonal Construct CR � AVE MSV ASV Factor correlation matrix SAT TRUST COMP BEN TRANS PERS Satisfaction (SAT) 0.95 0.95 0.87 0.84 0.66 0.93 Trust (TRUST) 0.90 0.90 0.76 0.84 0.77 0.92 0.87 Competence (COMP) 0.77 0.77 0.53 0.76 0.63 0.76 0.87 0.73 Benevolence (BEN) 0.90 0.89 0.75 0.71 0.63 0.79 0.84 0.75 0.86 Transparency (TRANS) 0.82 0.79 0.60 0.80 0.67 0.73 0.89 0.79 0.79 0.78 Personal service (PERS) 0.87 0.86 0.68 0.79 0.70 0.85 0.86 0.80 0.80 0.89 0.82 92 A. Carlander et al. / Financial Services Review 27 (2018) 83-98 4. Discussion Our results obtained from the survey of customers of Swedish retail banks support Hypothesis 1 that that the quality of personal service increases the customers’ trust in their banks through perceived competence, perceived benevolence, and perceived transparency. We propose that the basis for customers’ perceptions of competence, benevolence, and transparency is personal contacts with the bank employees, but we are only able to infer this from the indirect effects in statistical model tests based on questionnaire data. Additional research, therefore, needs to verify our results by observations of actual interactions between bank customers and employees. In support of Hypothesis 2 we found that the quality of personal service influences satisfaction through trust and its determinants perceived competence, perceived benevolence, and perceived transparency. In a financial service context, previous studies of trust, including some aspects of the interaction between employee and customer (e.g., Shainesh, 2012), have not attempted to draw implications for satisfaction. However, the focus on a satisfied customer is important from a practical perspective because customer satisfaction indices are measured regularly and attract much attention. Assessing the determinants of satisfaction with specific banks has furthermore the virtue of identifying concrete measures aimed at increasing the level of customer satisfaction proximal to the management of the banks. Previous research has reported a direct effect on satisfaction of personal service (Crosby et al., 1990; Krishnan et al., 1999; Taylor and Baker, 1994). A direct effect was, however, only weakly supported by our results. A direct effect may be explained by personal contacts that elicits affective (Johnson and Grayson, 2005) or relational (Ponder, Bugg Holloway, and Table 4 Standardized estimates (�), 95% confidence intervals (CI), and p values for total, direct, and indirect effects Personal service Competence Benevolence Transparency Trust � CI� p � CI� p � CI� p � CI� p � CI� p Total effects Competence 0.83 0.10 0.001 Benevolence 0.83 0.12 0.001 Transparency 0.91 0.14 0.001 Trust 0.88 0.08 0.002 0.37 0.32 0.005 0.32 0.30 0.001 0.33 0.42 0.009 Satisfaction 0.84 0.16 0.002 0.27 0.26 0.003 0.23 0.28 0.001 0.24 0.26 0.005 0.73 .50 0.002 Direct effects Competence 0.83 0.10 0.001 Benevolence 0.83 0.12 0.001 Transparency 0.91 0.14 0.001 Trust 0.37 0.32 0.005 0.32 0.30 0.001 0.33 0.42 0.009 Satisfaction 0.20 0.50 0.173 0.73 .50 0.002 Indirect effects Competence Benevolence Transparency Trust 0.88 0.04 0.002 Satisfaction 0.64 0.21 0.002 0.27 0.13 0.003 0.23 0.28 0.001 0.24 0.26 0.005 93A. Carlander et al. / Financial Services Review 27 (2018) 83-98 Hansen, 2016) components, which are likely to be less salient when answering a question- naire. This also speaks to the need for additional research using, for instance, in-depth interviews to disentangle direct affective or relational effects on satisfaction. We note that the sample we recruited represents customers of all the large Swedish banks despite that is was obtained from the customer register of one of the smaller niche banks. This is probably because of the fact that individuals increasingly are customers of more than one bank and hence appear in several customer registers. For the purpose of our study, this creates an opportunity to generalize the results that would otherwise have not been possible. However, it also creates a potential bias since our questions pertained to the “main bank,” that is the bank with which the customers perceive having the closest relation. Likely for this reason the ratings were generally positive, thus resulting in negatively skewed distributions that is a common finding in studies of customer satisfaction (Moe, Netzer, and Schweidel, 2017). This poses a problem for the statistical analyses. Yet, we found no differences in results in supplementary tests of our structural model using distribution-free estimation methods. Another problem that the statistical analyses do not address is that the skewed (or J-shaped) distribution reflects that predominantly positive but also more negative than neutral customers were more likely to participate. Although sampling biases cannot be discounted because of the low response rate, there are other possible causes of the positive ratings. A choice-supportive bias (Mather and Johnson, 2000) may have caused participants to be overly positive towards the bank of which they are regular customers. Previous research has also found that bank customers judge their own financial advisor as more trustworthy than financial advisors in general (Sunikka et al., 2010). It is, furthermore, possible that our somewhat older and more than average affluent participants may have closer relations to their banks than other customers, thus rating trust and satisfaction higher. In addition, 70% of the participants stated that they have been a customer with their main bank for more than 10 years. Previous studies have reported that the longer a successful trusting relationship lasts, the stronger the trust is (Gulati, 1995; Lewicki and Bunker, 1995). A halo effect (Rosensweig, 2014) implies that satisfaction or trust with their own bank may have spilled over to participants’ evaluations of other aspects (Sunikka et al., 2010). If the ratings we obtained have a common variance component because of the halo effect, it may account for the low discriminant validity. Statistical support was, however, only partial for that a single construct explained some of the common variance of the indicators. Another not mutually exclusive explanation includes a common method of data collection resulting in correlations between error variances. Trust in and satisfaction with banks may be increased or decreased by, for example, media reporting about financial crises. This may also have potential spill-over effects that can affect similar financial products that are being discussed in the media (Shin, 2009) as well as similar companies and institutions (Dahlén and Lange, 2006). It has been suggested that the 2008 financial crisis was sparked by a loss of trust (Sapienza and Zingales, 2012) that spread through the financial system like a pandemic outbreak. A serious outcome of such an outbreak is that it may create a self-fulfilling prophecy because banks are not fortified against a surge in liquidity outflow (Diamond and Dybvig, 1983). Considering that both good and bad news may spill over in an interconnected financial system, an issue for future research 94 A. Carlander et al. / Financial Services Review 27 (2018) 83-98 is to understand how under such circumstances loss of trust would be regained to avoid the severity of financial meltdowns. Additional research focusing on other customer segments is needed to generalize the results of the study. It is possible that a younger, less affluent sample would judge personal service to be less important considering the penetration of internet services, smartphone apps, and financial intermediaries, such as for example PayPal, that makes financial errands convenient without a long-term bank relationship (KPMG NunWood, 2016). Self-service technology is at least a complement that may serve as an important means of increasing product differentiation and thus satisfaction. Whether automation increases or decreases trust in the provider is another issue for future research to address. 5. Conclusions Perceived quality of personal service deserves further empirical examination because is likely to be an important factor for both trust in and satisfaction with retail banks. Compe- tence, benevolence, and transparency are possible channels through which personal service influences trust and satisfaction. Improving bank employees in these respects are, therefore, a potential means by which banks may increase their customers’ trust and satisfaction. Note 1 We also included statements to measure stability as a determinant of trust as van Raaij and van Esterik-Plasmeijer (2017) did. Stability did however not show the expected correlation with trust. Two other variables that we measured were “good citizenship” and “bank’s trust in customers.” “Good citizenship” was excluded because it had a high correlation with benevolence, whereas the other variable did not correlate with trust. Acknowledgment This research was supported by Grant 2011-03872 from the Swedish Agency for Inno- vation Systems (Vinnova) to the Centre for Finance, University of Gothenburg, Göteborg, Sweden. We thank Martin Hedesström for comments on the manuscript. References Anderson, E., & Weitz, B. (1992). The use of pledges to build and sustain commitment in distribution channels. Journal of Marketing Research, 29, 18–34. Berry, L. L. (2000). Cultivating service brand equity. Journal of the Academy of Marketing Science, 28, 128–137. Bloemer, J., de Ruyter, K., & Peeters, P. (1998). Investigating drivers of bank loyalty: The complex relationship between image, service quality and satisfaction. International Journal of Bank Marketing, 16, 276–286. 95A. Carlander et al. / Financial Services Review 27 (2018) 83-98 Bollen, K. A. (1989). Structural Equations With Latent Variables. New York, NY: Wiley. Carlander, A. (2015). Trust as a strategy for handling uncertainty in private savings choices (Doctoral disser- tation). Retrieved from GUPEA. (ISBN: 978-91-628-9464-1). Chakravarty, S., Feinberg, R., & Widdows, R. (1997). Reasons of their discontent. Bank Marketing, 29, 49–52. Colquitt, J. A., Lepine, J. A., Piccolo, R. F., Zapata, C. P., & Rich, B. L. (2012). Explaining the justice- performance relationship: Trust as exchange deepener or trust as uncertainty reducer? Journal of Applied Psychology, 97, 1–15. Crosby, L. A., Evans, K. R., & Cowles, D. (1990). Relationship quality in services selling: An interpersonal influence perspective. Journal of Marketing, 54, 68–81. Dahlén, M., & Lange, F. (2006). A disaster is contagious: How a brand in crisis affects other brands. Journal of Advertising Research, 46, 388–397. Devlin, J. F., & Yeung, M. (2003). Insights into customer motivations for switching to Internet banking. International Review of Retail, Distribution and Consumer Research, 13, 375–392. Dia, E. (2011). Uncertainty, trust, and the regulation of the banking industry. International Review of Economics, 58, 213–228. Diamond, D. W., & Dybvig, P. H. (1983). Bank runs, deposit insurance, and liquidity. Journal of Political Economy, 91, 401–419. Edelman. (2017). Edelman Trust Barometer (available at https://www.edelman.com/research/2017-trust- barometer-global-results). Ennew, C., Kharouf, H., & Sekhon, H. (2011). Trust in UK financial services: A longitudinal analysis. Journal of Financial Services Marketing, 16, 65–75. Fornell, C., & Larcker, D. F. (1981). Evaluating structural equation models with unobservable variables and measurement error. Journal of Marketing Research, 18, 39–50. Gallup. (2013). Confidence in Institutions (available at http://www.gallup.com/poll/1597/confidince-institutions. aspx#1). Gambetta, D. (1988). Can we trust trust? In D. Gambetta (Ed.), Trust: Making and Breaking Cooperative Relations (pp.725–737). Oxford: University of Oxford. Geyskens, I., Steenkamp, J. E. B. M., & Kumar, N. (1999). A meta-analysis of satisfaction in marketing channel relationships. Journal of Marketing Research, 36, 223–238. Gough, O., & Niza, C. (2011). Retirement saving choices: Review of the literature and policy implications. Journal of Population Ageing, 4, 97–117. Gulati, R. (1995). Does familiarity breed trust? The implications of repeated ties for contractual choice in alliances. Academy of Management Journal, 38, 85–112. Hair, J. F. J., Black, W. C., Babin, B. J., Anderson, R. E., & Tatham, R. L. (2010). Multivariate Data Analysis. Upper Saddle River, NJ: Prentice-Hall. Hayes, A. F., & Scharkow, M. (2013). The relative trustworthiness of inferential tests of the indirect effect in statistical mediation analysis: Does method really matter? Psychological Science, 24, 1918–1927. Hermansson, C. (2015). Understanding the relationships between bank-customer relations, financial advisory services and saving behaviour (Doctoral dissertation). Royal Institute of Technology, Stockholm, Sweden. ISBN: 978-91-87111-02-0. Hong, I. B., & Cho, H. (2011). The impact of consumer trust on attitudinal loyalty and purchase intentions in B2C e-marketplaces: Intermediary trust vs. seller trust. International Journal of Information Management, 31, 469–479. Howcroft, B., Hewer, P., & Durkin, M. (2003). Banker-customer interactions in financial services. Journal of Marketing Management, 19, 1001–1020. Johnson, D., & Grayson, K. (2005). Cognitive and affective trust in service relationships. Journal of Business Research, 58, 500–507. Jones, G. R., & George, J. M. (1998). The experience and evolution of trust: Implications for cooperation and teamwork. Academy of Management Review, 23, 531–546. Kantsperger, R., & Kunz, W. (2010). Consumer trust in service companies: A multiple mediating analysis. Managing Service Quality, 20, 4–25. 96 A. Carlander et al. / Financial Services Review 27 (2018) 83-98 Kaplan, D. (2009). Structural Equation Modeling: Foundations and Extensions. Thousand Oaks, CA: SAGE. King-Casas, B., Tomlin, D., Anen, C., Camerer, C. F., Quartz, S. R., & Montague, P. R. (2005). Getting to know you: Reputation and trust in a two-person economic exchange. Science, 308, 78–83. Krishnan, M. S., Ramaswamy, V., Meyer, C. M., & Damien, P. (1999). Customer satisfaction for financial services: The role of products, services, and information technology. Management Science, 45, 1194–1209. KPMG Nunwood. (2016). Banking the Customer Experience Dividend (available at https://www.nunwood.com/ excellence-centre/publications/white-papers/). Levy, S., & Hino, H. (2016). Emotional brand attachment: a factor in customer-bank relationships. International Journal of Bank Marketing, 34, 136–150. Lewicki, R. J., & Bunker, B. B. (1995). Trust in relationships: A model of development and decline. In B. B. Bunker, J. Z. Rubin, & Associates (Eds.), Conflict, Cooperation and Justice: Essays Inspired by the Work of Morton Deutsch (pp. 133–173). San Francisco, CA: Jossey-Bass. Lusardi A., & Mitchell, O. S. (2007). Financial literacy and retirement preparedness: Evidence and implications for financial education. Business Economics, 42, 35–44. MacKenzie, S. B., & Podsakoff, P. M. (2012). Common method bias in marketing: Causes, mechanisms, and procedural remedies. Journal of Retailing, 88, 542–555. Malhotra, D., & Lumineau, F. (2011). Trust and collaboration in the aftermath of conflict: the effects of contract structure. Journal of the Academy of Management, 54, 981–998. Mather, M., & Johnson, M. K. (2000). Choice-supportive source monitoring: Do our decisions seem better to us as we age. Psychology and Aging, 15, 596–606. Moe, W. W., Netzer, O., & Schweidel, D. A. (2017). Social media analytics. In B. Wierenga & R. van der Lans (Eds.), Handbooks of Marketing Decision Models (pp. 483–504). New York, NY: Springer Publishing. Moin, S. M. A., Devlin, J., & McKechnie, S. (2015). Trust in financial services: Impact of institutional trust on trusting belief. Journal of Financial Services Marketing, 20, 91–106. Plitt, M., Savjani, R. R., & Eagleman, D. M. (2015). Are corporations people too? The neural correlates of moral judgments about companies and individuals. Social Neuroscience, 10, 1–13. Ponder, N., Bugg Holloway, B., & Hansen, J. D. (2016). The mediating effect of customers’ intimacy perceptions on the trust-commitment relationship. Journal of Services Marketing, 30, 75–87. Rawlins, B. (2009). Give the emperor a mirror: Toward developing a stakeholder measurement of organizational transparency. Journal of Public Relations Research, 21, 71–99. Rosenzweig, P. (2014). The Halo Effect: How Managers Let Themselves be Deceived. New York, NY: Simon & Schuster. Sapienza, P., & Zingales, L. (2012). A trust crisis. International Review of Finance, 12, 123–131. Sekhon, H., Ennew, C., Kharouf, H., & Devlin, J. (2014). Trustworthiness and trust: influences and implications. Journal of Marketing Management, 30, 409–430. Selnes, F. (1998). Antecedents and consequences of trust and satisfaction in buyer-seller relationships. European Journal of Marketing, 32, 305–322. Shainesh, G. (2012). Effects of trustworthiness and trust on loyalty and intentions: Validating a parsimonious model in banking. International Journal of Bank Marketing, 30, 267–279. (available at http://dx.doi.org/ 10.1108/02652321211236905). Shin, H. S. (2009). Reflections on Northern Rock: The bank run that heralded the global financial crisis. Journal of Economic Perspectives, 23, 101–119. (available at http://dx.doi.org/10.1257/jep.23.1.101). Siegrist, M., & Cvetkovich, G. (2000). Perceptions of hazards: The role of social trust and knowledge. Risk Analysis, 20, 713–720. Sirdeshmukh, D., Singh, J., & Sabol, B. (2002). Consumer trust, value, and loyalty in relational exchanges. Journal of Marketing, 66, 15–37. Sunikka, A., Peura-Kapanen, L., & Raijas, A. (2010). Empirical investigation into the multi-faceted trust in the wealth management context. International Journal of Bank Marketing, 28, 65–81. Taylor, S. A., & Baker, T. L. (1994). An assessment of the relationship between service quality and customer satisfaction in the formation of consumers’ purchase intentions. Journal of Retailing, 70, 163–178. 97A. Carlander et al. / Financial Services Review 27 (2018) 83-98 Twenge, J. M., Campbell, W. K., & Carter, N. T. (2014). Declines in trust in others and confidence in institutions among American adults and late adolescents, 1972–2012. Psychological Science. (Advance online publication). van Esterik-Plasmeijer, P. W. J., & van Raaij, F. (2017). Banking system trust, bank trust, and bank loyalty. International Journal of Bank Marketing, 35, 97–111. Vargo, S. L., & Lusch, R. F. (2004). Evolving to a new dominant logic for marketing, Journal of Marketing, 68, 1–17. Yim, C. K., Chan, K. W., & Lam, S. S. K. (2012). Do customers and employees enjoy service participation? Synergistic effects of self- and other-efficacy. Journal of Marketing, 76, 121–140. 98 A. Carlander et al. / Financial Services Review 27 (2018) 83-98