1
Trust in technology and institution under PSD2
regulation: Will consumers accept third parties
interface to use their daily financial services?
Ester Zijlmans Student number: 11153466
University of Amsterdam Faculty of Science
Thesis Master Information Studies: Business Information Studies Final version: 23-09-2017
Supervisor: Dick Heinhuis Examiner: Tom van Engers
Abstract. Understanding why people adopt or reject a new technology has been proven one of the most challenging issues in the information system research (Swanson, 1988). Most predominant theories explaining the consumer behaviour to use a specific technology focuses on technology acceptance models like TAM and UTAUT. Under PSD2 regulation, it becomes possible for consumers to make use of third parties interfaces to manage their finance. Therefore consumers have to trust this third party with their personal financial information. Institutional Trust is considered to be an important factor to predict and explain whether consumers will share their personal financial information with third parties. This study investigates the role of Institutional Trust, Technology Acceptance and Perceived Risk to explain the Behaviour Intent of consumers. Furthermore, the role of Consumer Attitude functions and how this is related to Institutional Trust and Behaviour Intent will be examined. A new research model is conducted and tested through an online field study of 136 participants using structural equation modelling. Results indicate a significant indirect effect of Institutional Trust on Behaviour Intent by lowering Perceived Risk and increase the Technology Acceptance. The findings yield implications for both researchers and financial institutes.
Keywords. Institutional Trust, Technology Acceptance, Perceived Risk, Behaviour Intent, Consumer Attitude, PSD2, PLS-SEM
1. Introduction
1.1. Problem indication
In 2015 the European Union (EU) adopted a new directive on Payment Services (PSD2) to “improve the existing rules and take new digital payment services into account.” (European Commission, 2015a) The EU Single Market allows people, services goods and capital to move freely in an economy producing around 15 trillion euro annually. Therefore it offers new opportunities for European Business and enhances competition leading to lower prices for over
2 500 million consumers. (European Commission, 2017). PSD2 main goal is to make it easier, faster, and less expensive for consumers to pay for goods and services, by “promoting innovation enhancing payment security, and standardizing payment systems across Europe.” (Sandrock and Firnges, n.d.)
The first PSD directive, implemented in 2007 establishes a “modern and harmonised legal framework” necessary for the creation of an integrated payments market to make sure that quick, efficient and secure payments can be made throughout the European Union. One of the main objectives is to “remove legal barriers to entry into new markets within the EU/EEA.” (European Commission, 2007) and provide more transparency and information for consumers. (European Commission, 2015b) Since the adoption of Directive 2007/64/EC (PSD) new types of payment services have been developed especially in the area of digital payments. An example of such a service is the account information service, which “provide the payment service user aggregated online information on one or more payment accounts” and therefore enable the payment service user an “overall view of its financial situation immediately at any given moment.” (European Commission, 2007)
PSD2 will be active in 2018 and its primary goals are: I) make it easier and safer to use internet payment services; II) better protect consumers against fraud, abuse, and payment problems; III) promote innovative mobile and internet payment services; IV) strengthen consumer rights; V) strengthen the role of the European Banking Authority (EBA) to coordinate supervisory authorities and draft technical standards. (European Commission, 2015)
Under PSD2 legislation, bank customers are able to use third-party providers to manage their finances. One of the rules of the legislation is known as Access to Accounts (XS2A). This rule obliges banks to provide other companies, known as third-party providers (TTP), access to their customer’s accounts through open APIs (application program interface). This will enable third parties to build financial services on top of bank’s data and infrastructure. (Helstrom, n.d.)
PSD2 therefore introduces two new types of licence for TPPs: a licence for Payment Initiation Service Providers (PISPs) and a licence for Account Information Service Providers (AISPs). (Moinian, n.d.) Payments initiation service enables banks to initiate a payment order at the request of a Payment Service User (PSU) in this case the consumer. Therefore Payment Initiation information of consumers is required. The second one, Account Information Services, allow consumers and businesses to have an aggregated view on their financial situation, by enabling consumers to consolidate the different current accounts and transactional information, which requires access to the Account Information of the consumer. (European Commission, 2015b) Consumers need to give permission to their banks to share their financial information and determine which information is shared with TPP’s or other banks. Facilitating access to banks accounts, banks are required to deliver three types of sources: I) Account Information Services (AIS) which allowed third parties to request account information; II) Payment Initiation Services (PIS) to allow third parties to initiate payments and III) Confirmation on the Availability of Funds to allow third parties to check the customer's balance before a payment transaction. (Hoedt, 2016) The first two are depicted in figure 1.
3
Figure 1. AIS and PISP
Confirmation on the Availability of Funds to allow third parties to check the customer's balance before a payment transaction enables the AISP to offer consumers new services based on their financial information like a complete overview of their financial situation. Therefore third parties can use the API’s provided by AS PSP’s (fig. 2).
Figure 2. Overview of API provider by AS PSP’s Source: https://marcabraham.com/tag/fintech/
4 Allowing access to third parties potentially has profound consequences to banks. It may reduce “their ability to use current account relationships as gateway products for sale of other products and services” and customers then have the opportunity to choose providers who offer better service and quality. (Temedos, n.d.)
Once PSD2 will be implemented in 2018, it will become possible for multiple organizations like FinTech and Banks to become a PISP and AISP. According to Sandrock and Firnges (2016) over 68% of bankers fear that PSD2 will cause them to lose control over their clients interface, however, many of them remain unsure how to prepare to the new directive and adopt a defensive “wait and see” strategy. Previous research of consultancy companies shows that 88% of the people already use third party digital payments services, such as PayPal and that 85% of them rate the security of this payment services as ‘high’ or ‘very high’. (Sandrock and Firnges, 2016) Beside, the same research suggest that 82% of the respondents agree or even strongly agree that companies such as PayPal or Amazon can handle transfers as reliably and safely as their banks. The report concludes that third-party payment services therefore already have earned consumer trust. The level to which third-party providers can “capitalize on this high level of trust in their payment service” can be seen as the most important factor to determine the competitive threat of third parties.
1.2. Problem statement and practical relevance
Once PSD2 will be implemented in 2018, it will become possible for consumers to make use of third-party providers to manage their finances. Personal financial information, including detailed account information, obtained financial services and transaction-related data can be shared with third parties. However, consumers have to give permission first to share their personal information. Therefore, consumers have to trust a third party with their personal financial information. Hence, the question whether consumers will trust the third parties with their personal information and how trust is related to Behaviour Intent arises. From a marketing perspective, knowing the impact of trust on the Behaviour Intention of consumers, and how this trust is build is of strategic relevance for banks in order to proactively adjust their digital strategy concerning PSD2.
Previous research mainly investigated Trust within Technology to explain the Behaviour Intent of consumers regarding technology innovations. Most studies adopt one or more constructs of dominant theoretical approaches such as the technology acceptance model (Davis, 1989), the unified theory of acceptance and use of technology (Venkatesh, Morris and Davis, 2003) or the theory of planned behaviour (Taylor and Todd, 1995) in order to predict the consumer Behaviour Intent. Nowadays, technology seems to be more and more accepted and is therefore less important to predict consumer’s Behaviour Intent. Trust within the institution on the other hand has been barely investigated within the past. Institutional Trust makes consumers “comfortable sharing personal information, making purchases, and acting on Web vendor advice—behaviours essential to widespread adoption of e-commerce” (McKnight et al. 2002) and according to Giddens (1990) “trust in the expert system”, or institutional trust seems more important than interpersonal Trust to predict the consumer’s Behaviour Intent. Several studies confirmed a negative relation between Perceived Risk (PR) and BI (Kerviler et al. 2016,
5 Mingxing et al. 2014). Hence, the impact of PR on the BI of consumers to share their personal information and how this is related with Institutional Trust and Trust in Technology seems to be important as well.
Finally, the consumer attitude functions have been demonstrated as important for explaining consumer behaviour (Grewal et al. 2004). Beside, from the institution point of view, the consumer’s attitude functions are also the ones who can be influenced by the institute.
Summarizing, this research will give insight in the impact of Trust on the Behaviour Intent of consumers. It provides insight in the role of Institutional Trust and Trust in Technology on Behaviour Intent and how the institute can change this by influencing consumer’s attitudes. More specifically, it will give financial institutes insight in whether consumers are willingness to share their personal financial information with third parties and therefore the competitive role of third parties underlying the PSD2 regulation.
1.3. Research Questions
The description above leads to the topic of the influence of Technology Acceptance and Trust in Institutions on Perceived Risk and therefore on the Behaviour Intent of consumers to share their personal financial information with third parties under the PSD2 Regulation. To identify this influence, several theoretical models will be investigated to find out which constructs lead to Behavioural Intent within the context of financial services.
Hence, the following main research question derives:
What is the (mediated) impact/influence of Institutional Trust on Perceived Risk and Trust in Technology (TA) to predict the Behaviour Intent of consumers to share their personal financial information with third parties?
This study will answer the main research question as follows. First, a research model based on previous literature is introduced to define the main constructs, which influence the BI of consumers. Second, hypothesis based on the structural model will be formulated which together form the measurement model. The final measurement model will be tested conducting a survey based on the main constructs and their proposed relationships.
1.4. Research Gap and contribution (Theoretical relevance)
This study contributes to the academic research in three ways. First, previous research has been done on adopting technology in the field of e-shopping, e-commerce, e-finance, e-banking and mobile banking. (Cheng et al. 2006, Gu et al. 2009, Lee, 2009, Lin 2011, Zhou 2012). Most of these studies have drawn on widely accepted theoretical models like the Technology Acceptance Model (Davis, 1989), the Unified Theory of Acceptance and Use of Technology (Venkatesh, Morris and Davis, 2003) or the Theory of Planned Behaviour (Taylor and Todd, 1995), on exploring the adoption of new financial technologies. Almost no prior research has investigated the influence of Institutional Trust on Behaviour Intent within the field of finance and e-banking. This paper contributes by proposing and validating measures of Institutional Trust as
6 one of the strongest influence on the BI of consumers to share their personal financial information.
Second, since PSD2 regulation is not implemented yet, this study can be considered as exploratory to predict weather Institutional Trust will influence the BI of consumers to share their personal financial information. This study contributes by taking a case study, which hasn’t been researched within the academic literature before.
Third, a new research model will be developed, based on several existing models which have not been combined before to explain the BI of consumers to adopt new financial technologies. The impact of Institutional Trust on BI by lowering Perceived Risk is barely discussed in academic literature so far. Previous research did found a significant and positive influence of ‘Institutional trust‘ on customer attitude within the context of 'financial innovation’ (Lee, 2011). However, the proposed research model lacks the mediated effect of Institutional Trust on BI by decreasing Perceived Risk. Therefore, this paper contributes to this research gap.
1.5. Methodology 1.5.1. Thesis outline
This thesis can be read as follows: The second chapter introduces the main concepts based on prior literature of Institution Trust, Technology Acceptance, Behaviour Intention, Perceived Risk and Consumer Attitude. The main purpose of this chapter is to provide an understanding of the underlying constructs of the main concepts. The third chapter consist of the theory and development of the hypotheses derived from the literature review. The main purpose of this chapter is to provide a framework, which can be tested in the empirical research part. Next, the measurement, data collection and data sample will be explained in chapter four. Chapter five discusses the outcome of the research model, the hypothesis and the validity and liability of the results. Furthermore, the conclusion and recommendations for banks will be outlined.
Finally, chapter six describes the limitations of this research and outline future research directions.
1.5.2. Research design and methodology
In order to answer the research question a literature review is conducted to identify the current state of knowledge concerning Technology Acceptance, Institutional Trust, Behaviour Intent, Perceived Risk and Consumer Attitude in the field of e-finance, mobile banking, internet banking and financial innovations. The literature review results in a new research model and the respective hypotheses. Additionally, the literature review is used to define and propose questions in order to measure the different constructs. A quantitative questionnaire will be spread through multiple communication channels in order to collect a significant amount of respondents. Quantitative data analysis methodology will be conducted to test the hypotheses from the research model. Based on these results supported by literature research the outcome will be examined and explained in order to answer the research question.
7 analysts. SEM has “the ability to incorporate several regression equations in a single analysis and can also be used in testing the direct and indirect relationships simultaneously.“ (Abroud et al. 20013) The confirmatory factor analysis (CFA) will be used to validate the measurement scales for each construct. Composite reliability (CR) and AVE measures will measure the convergent validity of the constructs. (Hair et al. 2006)
2. Theoretical research and foundation
In chapter one, several topics are identified which seems important to predict and explain the Behaviour Intent of consumers to share their personal financial information. This chapter provide a literature research to define and elaborate this main concepts and their underlying constructs to obtain a broader understanding of the research object. The findings of this research serve as the theoretical foundation for developing the structural model and formulating hypotheses to answer the main research question.
2.1. Trust
According the latest report of the Edelman Trust Barometer (2016), only 51% have trust within the Financial Industry, which is therefore still the least trusted industry. Previous literature shows that trust can be seen as a multidimensional concept and that “each trust type has different implications for consumer behaviour and for how trust is built” (McKnight et al. 2002).
For example, Mayer (2008) argues that, in the past financial markets used to rely heavily on interpersonal trust, instead of Institutional Trust. Giddens (1990), on the other hand, connects interpersonal and system trust in the context of a broad historical or “evolutionary” progression. He argues that “the development towards (late) modernity shows how virtually all sectors of contemporary organized social life – from flying in planes over visiting a doctor to drinking clean tap water – fundamentally rely on trust in the respective expert systems, where in pre-modern or early pre-modern times, trust in the individual medic was typically required.” According to Kroeger (2015), the financial sector provides an obvious example for this development as mentioned before since financial markets are “regarded as essentially technical phenomena, to be anticipated and engaged with on the basis of rational calculation and largely technical expertise.” Hence, the concept of Trust in Technology will be investigated to incorporate within the proposed research model.
2.1.1. Trust in Technology
Trust within Technology has been widely discussed within previous research (Fishbein & Ajzen, 1977; Davis, 1989; Venkatesh et al., 2003) and seems to be closely related with the concept of technology acceptance- or technology adoption models. An overview of most dominant theories in the field of technology acceptance models can be found in Appendix I.
Most predominant theories regarding Trust in Technology explain the user acceptance and Behaviour Intent regarding (new) technology-using concepts like Attitude Towards a Behaviour (Ajzen, 1991) and Perceived Ease of Use (Davis, 1989). The UTAUT2 model explains 40 to
8 52% of the Behaviour Intent and is mentioned improved to be more oriented towards a consumer context (Venkatesh et al. 2012). This model was tested and applied to several technologies (Im, Hong & Kang, 2011) and used in many studies. Like earlier acceptance and adoption models, it aims to explain user intentions to use an Information System (IS) and further the usage behaviour (Alshehri, Drew, & AlGhamdi, 2012). Hence, I accept that UTAUT2 is best theoretical based to measure the TA of consumers. Therefore, the constructs that fit within the scope of this research are adopted.
2.1.2. Institutional Trust
A number of empirical studies have investigated the role of trust in the specific context of e-commerce, focusing on different aspects of this multi-dimensional construct. (Grabner-krauter and Kaluscha, 2003) Previous Literature distinct Trust in the Institute from Interpersonal Trust. Giddens (1990) argue that “trust in the expert system”, or institutional trust seems more important than interpersonal Trust to predict the consumers Behaviour Intent to share their personal financial Information with third parties. Luhmann's (1979) classical definition of system trust said that, in system trust the trustee “basically assumes that a system is functioning and places his trust in that function, not in people”. Institutional trust can exist at the organizational (Costigan et al., 1998; Rousseau et al., 1998), interorganizational (Zucker, 1986) or societal level (Bachmann, 2001). McKnight et al. (2002) defined Institution-based trust as “the sociological dimension of trust” and refers to an “individual’s perceptions of the institutional environment”—in this case, the party providing access to bank accounts and enabling transactions.
According to Keen et al. (2009) “perceptions of the structural characteristics of digital transactions, such as safety and security, can influence trusting beliefs and trusting intentions towards a specific vendor”, in this case the institution deliver the interface for handling financial digital transactions. Individual perceptions therefore can be influenced by the institute to change the consumer's BI. Institutional Trust makes consumers “comfortable sharing personal information, making purchases, and acting on Web vendor advice—behaviours essential to widespread adoption of e-commerce.” (McKnight et al. 2002). Therefore, to predict the Behaviour Intention (BI) of consumers to share their personal financial information, not only trust in the underlying technology is important, trust in the institution offering the technology seems even more important. Institutional Trust is adopted as the strongest influence for predicting the consumer's BI.
2.2. Consumer Attitude
To explain how Institutional Trust can influence the BI of consumers, individual perceptions and attitudes are the constructs, which can be influenced by the institute. During the past decades, much research is done on the concept of attitude and attitude formations within the field of consumer behaviour. (Chen and Wells, 1999; Grewal; et al. 2004; Stuart et al., 1987) Most literature makes a clear distinction between functional (Katz, 1960) and constructive theories (Reed et al. 2002). The functional approach to attitudes addresses the motivational bases of people’s attitudes and is originally conceptualized by Katz (1960) and Smith et al. (1956). An overview of most important functional and constructive attitude theories and their constructs are therefore summarized in Appendix II.
9 The theory of reasoned action (TRA; Fishbein & Ajzen, 1975) and the theory of planned behaviour (TPB; Ajzen, 1991) are well-researched intention models that have proven successful in predicting and explaining behaviour across a wide variety of domains. (Yousafzai et al. 2010) Therefore, to explain the attitude function within the context of this research the relevant constructs of the theory of reasoned action from Fishbein & Ajzen (1975) are adopted.
2.3. Perceived Risk
Several studies confirmed a negative relation between Perceived Risk and BI (Kerviler et al. 2016, Mingxing et al. 2014). According to Featherman and Pavlou (2003) “Privacy and financial risks are linked to the potential monetary and psychological losses due to a loss of control over personal information” and “transaction errors or fraudulent uses of banking information.” (Lee, 2009) Therefore to identify what the PR is towards the BI of consumers to share their personal information is important.
The concept of Perceived Risk is originally introduced by Bauer (1960) in the Marketing Literature, stating that “consumer behaviour involves risk in the sense that any action of a consumer will produce consequences which he cannot anticipate with anything approximating certainty, and some of which at least are likely to be unpleasant” and later by Cox and Rich (1964) who defined Perceived Risk as “the nature and amount of uncertainty perceived by a consumer in contemplating a particular purchase decision.” Since then, the construct of Perceived Risk has become a common construct used by researcher in consumer behaviour (Hoover et al. 1978) and online consumer behaviour (Archer and Yuan, 2000; Ha, 2002). Previous research describes the construct of Trust as having a negative relationship with Perceived Risk. For example Morgan and Hunt (1994) and Gao et al. (2002) stated that “when a trust-based relationship between a buyer and a seller was developed, the buyer’s trust was likely to be higher and therefore the buyer’s uncertainty or perceived risk reduced”, possibly suggesting that trust may moderate perceived risk. Hence, the question whether Institutional Trust may moderate Perceived Risk seems to be justified and is therefore incorporate within the model.
3. Research model and hypothesis
Summarizing the theoretical research, four concepts are identified to predict the BI of consumers to share Personal Financial Information: Institutional Trust, Perceived Risk, Technology Acceptance and Consumer Attitude to answer the main research:
What is the (moderated) impact/influence of Institutional Trust (IT) on Perceived Risk (PR) and Trust in Technology (TA) to predict the Behaviour Intent of consumers to share their personal financial information with third parties?
A structural model of the defined constructs and their influence on BI of consumers derived from the research question is proposed in Figure 3.
10
Figure 3. Structural research model
In order to test the structural model, hypotheses will be formulated and the final measurement model introduced.
3.1. Hypothesis development Trust in Technology (Technology Acceptance)
To measure the concept of Technology Acceptance, two constructs are identified from previous models that fit within the scope of this study: Performance Expectancy and Hedonic Motivation. (Venkatesh et al. 2012). ‘Performance Expectancy’ (PE) is defined as the degree to which using a technology will provide benefits to consumers in performing certain activities. (Venkatesh et al. 2012). Hedonic Motivation is the degree of fun and pleasure that is perceived by an individual or consumer of using a specific technology. Since additional financial services provided by third parties include design and user experience as well, this construct is adaptedto the proposed research model to measure the TA of consumers. Thus, I hypothesis that:
H1. Technology Acceptance positively influences the Behaviour Intent. H1.a Hedonic motivation positively influences the Behaviour Intent. H1.b Performance Expectancy) positively influence the Behaviour Intent.
3.2. Hypothesis development Consumer Attitude function
Two constructs are adopted from the literature to explain the Consumer Attitude towards Behaviour Intent: “Attitude Towards Behaviour (ATB)” and the “Subjective Norms (SN)” (Fishbein & Ajzen, 1975). Attitude toward behaviour is defined as an individual’s positive or negative feelings (evaluative affect) toward performing the target behaviour. The Theory of Reasoned Action (TRA) posits that “behavioural intention is a function of two determinants: a personal factor termed attitude toward behaviour, and a person's perception of social pressures
11 termed subjective norm” (Fishbein & Ajzen, 1975). I assume that a higher level of attitude towards a behaviour and a better perception of a behaviour, the subjective norm, have a positive influence on Behaviour Intent. I therefore hypothesis that:
H.2 Consumer attitude has a positive influence on Behaviour Intent.
H2.a Attitude towards behaviour has a positive influence on Behaviour Intent.
H.2.b Subjective Norm towards behaviour has a positive influence on Behaviour Intent.
3.3. Hypothesis development Trust in Institutions
To measure the concept Trust within the banking industry, three constructs are adopted from previous Literature: Reputation, Satisfaction and Communication. (Johnson and Grayson, 2005; Kim et al., 2009) Organizational reputation within the context of this research is defined as “a collective conception formed among a group of people toward the organization” (Yang and Grunig, 2005). Within the financial industry, reputation and organizational trust are identified to be “important influencers of the type of relationships between an organization and its customers” (Mukherjee and Nath, 2003; Kwon and Suh, 2004). Hence, I assume that the organizational reputation is an important factor to measure organizational or institutional Trust.
Customer satisfaction is closely linked to past service experiences or evaluations of a product or service attributes and determined by the discrepancy between the customer’s expectations and what they actually feel about the experience (Oliver, 1981). According to Hart and Johnson (1999), “trust mediates the satisfaction-loyalty relationship.” It can strengthens organizational trust through the transferral of positive past experiences on to future exchanges, and via reducing uncertainty about the long term (Gill, et al., 2006). Therefore, Satisfaction is adopted as the second construct to measure Institutional Trust.
The last construct, Communication can be defined as “the formal as well as informal sharing of meaningful and timely information” (Anderson and Narus, 1990). Mukherjee and Nath (2003) confirmed that communication plays a significant positive role on trust within the online financial sector and found that more open and active communication enhance organizational trust through the development of innovations. Morgan and Hunt (2004) stated that “perception of intense and high-quality communication are argued to have a positive impact on organizational trust, because both parties assume that future communication will continue to be timely and reliable.” This is supported by Ball et al. (2004) who argue that “communication has an significant impact on customer trust.” It therefore seems to be justified that Communication may enhance Institutional Trust and is therefore adopted as the third construct for Institutional Trust.
All three constructs have been identified as important factors to measure organizational or Institutional Trust within the financial sector. Institutional Trust makes consumers comfortable sharing personal information, making purchases, and acting on Web vendor advice—behaviours (McKnight et al. 2002) and helps to reduce both system-dependent and transaction-specific uncertainty or risk (Zucker, 1986). I therefore assume that there is a positive relation between Institutional Trust and Perceived Risk. Hence, I hypothesis that:
12 H3. Institutional Trust (reputation, satisfaction and communication) has a positive influence on Perceived Risk.
H.3a. Reputation has a positive influence on Financial Risk. H.3.b Satisfaction has a positive influence on Financial Risk. H3.c Communication has a positive influence on Financial Risk. H.3.d. Reputation has a positive influence on Performance Risk. H.3.e Satisfaction has a positive influence on Performance Risk. H.3.f Communication has a positive influence on Performance Risk. H.3.g. Reputation has a positive influence on Privacy Risk.
H.3.h Satisfaction has a positive influence on Privacy Risk. H.3.i Communication has a positive influence on Privacy Risk.
In addition, a higher level of Institution Trust implies a higher level of BI as well (Lee, 2001) Hence, I hypothesis that:
H4. Institutional Trust (reputation, satisfaction and communication) has a positive influence on Behaviour Intent.
H.4.a. Reputation has a positive influence on BI. H.4.b Satisfaction has a positive influence on BI. H.4.c Communication has a positive influence on BI.
Finally, the constructs underlying the consumer attitude functions has been demonstrated as important for explaining consumer behaviour (Grewal et al. 2004). From the institution point of view, the institute can influence the consumer attitude functions. It therefore seems to be justified that Institutional Trust positively influence the consumer's Attitude toward that institute. Hence I hypothesis that:
H5. Institutional Trust (reputation, satisfaction and communication) has a positive influence on consumer attitude.
H.5.a. Reputation has a positive influence on Consumer Attitude. H.5.b Satisfaction has a positive influence on Consumer Attitude. H.5.c Communication has a positive influence on Consumer Attitude. H.5.d. Reputation has a positive influence on Consumer Attitude. H.5.e Satisfaction has a positive influence on Consumer Attitude. H.5.f Communication has a positive influence on Consumer Attitude.
To explain the moderated influence of Institutional Trust on Perceived Risk and Technology Acceptance the following hypothesis is formulated:
H6. Institutional Trust has an indirect influence on Technology Acceptance by lowering Perceived
Risk
H.6.a Reputation indirect positively influences Performance Expectancy. H.6.b Satisfaction indirect positively influences Performance Expectancy.
13 H.6.c Communication indirect positively influences Performance Expectancy.
3.4. Hypotheses development Perceived Risk
Within the context of consumer behaviour, two types of risk have received strong attention and seem to be predominated: Financial Risk and Performance Risk. (Bettman, 1973; Grewal et al. 1994) Both constructs are adopted to explain the influence of Institutional Trust on Perceived Risk. Performance Risk is defined by Shimp and Bearden (1982) as “the uncertainty and the consequence of a product not functioning at some expected level.” Financial Risk can be explained as the uncertainty and the monetary loss one perceives to be incurring if a product does not function at some expected level (Grewal et al., 1994).
According to Doolin et al. (2005) Perceived Risk within the context of online behaviour specific relates to privacy, and a security risk of personal information. Privacy, of security risk therefore will be added as third construct to explain the concept of Perceived Risk and is defined as “the uncertainty associated with the negative consequences of using a particular product or service that involves any potential losses resulting from disclosing personal information” (Zhou, 2012) First, Prior literature (Luo et al. 2010) has been identified that PR negatively influence the Performance Expectancy and therefore Technology Acceptance. Hence, the relation between PR and TA will be researched. Second, previous research has shown that especially within the online and financial environment Perceived Risk influence the Behaviour Intent (Lee, 2009). Thus, I hypothesis that:
H7 Perceived Risk has a negative influence on Behaviour Intent.
H7.a Perceived Risk has a negative influence on Technology Acceptance. H.7.b Perceived Performance risk has a negative influence on Behaviour Intent.
H.7.c. Perceived Performance risk has a negative influence on Performance Expectancy. H.7.d. Perceived Financial risk has a negative influence on Behaviour Intent.
H.7.e. Perceived Financial risk has a negative influence on Performance Expectancy. H.7.f. Perceived Privacy risk has a negative influence on Behaviour Intent.
H.7.g. Perceived Privacy risk has a negative influence on Performance Expectancy.
Summarizing, 23 hypotheses were identified in order to predict the influence of Institutional Trust on Behavioural Intent to share personal financial information by lowering perceived risk.
4. Research methodology
In order to test the proposed research model, a quantitative questionnaire is conducted. This research method was chosen for two of reasons. Newsted et al. (1999) state that surveys seem to be most widely used method in information systems research. Beside, due the limitation of time, online survey allowed collecting necessary data from wider public in an effective and time efficient manner.
14
4.1. Measurement
For developing the questionnaire, questions from prior research are adopted within the context of this study. The measurements for Behaviour Intention, Performance Expectancy and Hedonic Motivation are adopted from the original UTAUT2 model of Venkatesh et al. (2012). For Institutional Trust, three underlying constructs are adopted in order to explain and define this concept: Reputation, Satisfaction and Communication. The Institute, in this case is both defined as the bank as the third party. Wartick (2002) state that “Corporate reputation should be measured as stakeholder’s perceptions, not factual representation” and that reputation is often issue specific. In this case, reputation is measured from a consumer perspective. To measure the construct ‘reputation’, questions are adopted from a 20-item reputation quotient scale (Gardberg & Fombrun, 2002), which measures different aspects of an organizational reputation.
For Satisfaction the relevant questions from the Consumer Products Questionnaire (CPQ) (McNamara & Kirakowski, 2011) are adopted, since their psychometric questionnaire is specific created to measure user-satisfaction with electronic consumer products (ECPs). Questions regarding specific products focus on the digital products available with banks, both the App and Internet Banking functionality to fit within the scope of this study. Questions for measuring communication are adopted from Ball et al. (2004) who studied the relation between communication and customer trust.
The questions for measuring the consumer Attitude (Ajzen and Fishbein, 2000) are borrowed from Laforet and Li (2005). In line with the suggestions of the theory of reasoned action (Ajzen & Fishbein, 1980), subjective norms are included to predict behavioural intentions. Measurements for the Subjective Norm are adopted from Ogle et al. (2004) who measured motivation to comply as “In general, I want to do what people who are important for me think I should do” which captures the core meaning of this concept and normative beliefs as the expected behaviour of people who are important to a person.
To measure the construct Risk, three underlying constructs are identified who fit in this study: Performance Risk (PR), Financial Risk (FR) and Perceived Privacy Risk (PPR). Questions regarding PR are based on Rao and Goldsby (2009) definition that PR includes “any uncertainty across the supply chain network that leads to failures in normal business operations for the supplier and/or the buyer, and the deployment of their resources/capabilities.” Measurements for FR are adopted from Featherman & Plavou (2003) who defines FR as “the loss of money due to an error or fraud.” PPR risk is the uncertainty associated with the negative consequences of using a particular product or service that involves any potential losses resulting from disclosing personal information (Zhou, 2012) Questions to measure PPR are adopted from Dinev and Hart (2006) who focused in their research on privacy issues related to the internet. An overview of the questions can be found within Appendix III. Data collection
To answer the main research question, 23 hypotheses are formulated based on the structural model, together forming the proposed research model as depicted within Fig. 4.
15 Figure 4. Proposed measurement model
In order to test the measurement model and questions, a pilot questionnaire was conducted and tested by seven participants. The participants were questioned regarding the consistency, understand ability and structure of the questionnaire. Some additional information regarding the legislation has been added after the pilot to increase the understand ability of the questionnaire. Some measurement was improved after the feedback obtained from the pilot questionnaire and simplified. The final Questionnaire can be found in Appendix IV. The questionnaire is conducted using https://www.enquetesmaken.com and, spread to several digital channels like Facebook, Linked-in, e-mail and intranet in order to obtain a representative sample. The incomplete response and pilot responses were omitted from the sample. A sample of 118 respondents was validated to use for the test and analysis of the proposed measurement model.
4.2. Data sample
Some additional questions were added to the 37 measurement items in order to obtain an understanding of the data sample including age, gender, and awareness of this legislation, highest degree and use of mobile banking applications. These results are depicted in Table 1. The average age of the sample is 39 years and most respondents are aged in the category 30-40 years (43%), followed by 20-30 years (23%). More than half of the sample is female (54%) and almost two-third (64%) was already aware of this legislation. From all respondents, more than
16 90% is highly educated and obtained a HBO degree or higher and the majority (40%) obtained a University Master Degree.
Table 1. Sample demographics
Measure Item Frequency Percentage
Age <20 20-30 30-40 40-50 >50 1 27 49 23 18 1% 23% 42% 19% 15% Gender Male Female 54 64 46% 54% Familiar with legislation Yes No 76 42 64% 36% Highest Degree High school
HBO Bachelor of Applied Science
University Bachelor Degree University Master Degree PH.D. Other 6 46 13 47 1 5 5% 39% 11% 40% 1% 4% Familiar with MB Yes
No 118 0 100% 0%
4.3. Data analysis and Results
After describing the characteristics of the respondents, the next step is analysing the collected questionnaire using the partial least square-structural equation modelling (PLS-SEM) approach which has been used by many researchers in recent years (Hair et al. 2011; Toe et al. 2015) In PLS-SEM, the sample size requirement must be at least ten times the largest number of structural paths directed at a particular latent construct in the structural model (Hair et al., 2011). As shown in Fig. 4, there are four paths directed at Behaviour Intent. Therefore the sample size must be larger than 40 respondents. The sample size of 136 respondents satisfies the minimum size requirements for PLS-SEM.
The partial least square-structural equation modelling (PLS-SEM) approach will be used to analyse the data. Two different types of SEM, covariance-based structural equation modelling (CB-SEM) and PLS-SEM can be used to estimate the parameters in hierarchical latent variable models (Wetzels et al., 2009). In the context of PLS-SEM models, hierarchical latent variable models have shown an increasing popularity in recent years (Edwards, 2001; Jarvis et al., 2003) where both higher-order as first-order constructs are defined. A higher (or second)-order construct is a general concept that is either represented (reflective) or constituted (formative) by its dimensions (lower (or first)-order constructs) (Becker et al. 2012). The high-order constructs within this study (Institutional Trust, Perceived Risk, Attitude and Technology Acceptance) are formed by a combination of several specific (latent) dimensions into a general concept and therefore formative (Edwards, 2001; Wetzels et al., 2009). The lower-order constructs are reflective measured constructs that can be distinguished from each other but also are correlated.
17 Before conducting PLS-SEM, the reliability and validity of the measurement model has to be ensured. To test the reliability of the model, the indicator reliability and internal consistency reliability are assessed using the outer loadings of the model, the Composite Reliability (CR) and the Cronbach’s Alpha (CA). According to Henseler et al. (2009) and Hulland (1999) outer loadings should be higher than 0,70. Two items of Institutional Trust Reputation are dismissed since both outer loadings are below 0,70 (IT_rep_Q3=0,48 and IT_rep_Q6=0,179). Two Item of Institutional Trust Satisfaction are below 0,7 as well (IT_sat_Q10=0,266 and IT_sat_Q9=0,65) and therefore dropped. Two items of Attitude are dismissed as well for the same reason: TRA_a_Q23 and TRA_a_Q26 (0,36 and 0,29). The last item dismissed to confirm indicator reliability is an item of Perceived Privacy Risk, PR_pri_Q37 with an outer loading of 0,043.
Secondly, before conducting PLS-SEM, one-dimensionality of each construct needs to be checked. A construct is one-dimensional when CA is greater than 0.7 (Tenenhaus et al., 2005; Straub, 1989) and have a CR value above 0.6 (Henseler et al., 2009). Table 2shows that the Cronbach’s alpha ranged from 0.71 to 0.95 and the CR values range from 0.83 to 0.98. All constructs are therefore statistically reliable. Validity testing identifies the convergent validity of the constructs using the average variance extracted (AVE) and the Fornell-Larcker criterion (1981). The AVE value should be above 0.5 to identify if the latent variable explains the variance of its indicator more than half. All constructs within table 2 meet this criterion.
Table 2. Item loadings, mean, standard deviation, CR, CA and AVE
Measure Item Mean St. dev. Outer
loading s CR CA AVE Institutional Trust reputation IT_rep_Q1 IT_rep_Q2 IT_rep_Q4 IT_rep_Q5 IT_rep_Q7 4,17 3,99 4,03 4,41 4,41 1,40 1,72 1,57 1,48 1,77 0,72 0,83 0,92 0,82 0,89 0,92 0,8 9 0,7 Institutional Trust Satisfaction IT_sat_Q8 IT_sat_Q11 5,75 5,26 0,94 1,26 0,87 0,91 0,89 0,7 5 0,8 Institutional Trust Communicati on IT_com_Q12 IT_com_Q13 IT_com_Q14 4,14 4,1 3,94 1,51 1,45 1,58 0,66 0,83 0,87 0,83 0,7 2 0,63 Technology Acceptance PE TA_pe_Q15 TA_pe_Q16 TA_pe_Q17 4,38 4,42 4,82 1,62 1,38 1,24 0,93 0,95 0,78 0,92 0,8 7 0,79 Technology Acceptance BI TA_bi_Q18 TA_bi_Q19 TA_bi_Q20 3,26 2,7 3,18 1,70 1,50 1,63 0,88 0,93 0,93 0,94 0,9 0,84 Technology Acceptance HM TA_hm_Q21 TA_hm_Q22 4,28 4,41 1,60 1,57 0,97 0,98 0,98 0,9 5 0,95 TRA Attitude TRA_a_Q24
TRA_a_Q25 3,39 3,61 1,76 1,75 0,89 0,89 0,89 0,7 4 0,8 TRA Subjective TRA_sn_Q27 TRA_sn_Q28 3,39 3,33 1,52 1,59 0,91 0,83 0,86 0,6 9 0,76
18 Norm Perceived Risk Financial Risk PR_fr_Q32 PR_fr_Q33 PR_fr_Q34 5,17 3,76 4,21 1,22 1,48 1,51 0,79 0,88 0,86 0,88 0,7 9 0,71 Perceived Risk Privacy Risk PR_pri_Q35 PR_pri_Q36 5,51 5,29 1,21 1,38 0,86 0,82 0,83 0,7 1 0,63 Perceived Risk Performance Risk PR_per_Q30 PR_per_Q31 3,64 4,17 1,39 1,33 0,97 0,75 0,86 0,7 3 0,75
The discriminant validity is assessed using the Fornell-Larcker criterion stated that the square root of AVE on each construct is greater than the correlations of the construct with other constructs. The results can be found in Appendix V and show that all constructs meet this condition. The items and constructs are used to assess the structural model.
After validating the measurement model, the structural model is assessed using Partial Least Squares - Structural Equation Modelling (PLS-SEM). The strength of the relationship between the first ordered constructs and second-order constructs are evaluated and can be found in Appendix VI. All first-order constructs except Privacy Risk, Performance Risk and Subjective Norm were found to have a significant path coefficient according to Mathieson et al. (2001). Formative constructs are allowed to contain non-significant indicators. Therefore, to ensure the content domain of perceived risk, all first-order risk, including the non-significant constructs, are kept.
The variance-based type of SEM is used within this study since the research model has not been tested before and therefore rather exploratory (Martins et al. 2014) and the requirements for PLS-SEM are less stricter concerning the sample size of residual distributions than covariance based (Luo et al. 2010).
To answer the research questions, both the structural model on a second-order level, as the underlying relations between the separated constructs (first-order-level) are investigated. For modelling hierarchical latent variables in PLS-SEM the sequential latent variable score method, or two-stage approach has been used (Ringle et al., 2012; Wetzel’s et al., 2009) where latent variables scores for lower-order latent variables are obtained from SPSS. It estimates the construct scores of the first-order constructs in a first-stage model, in this case done within SPSS, without the second-order construct present, and subsequently uses these first-stage construct scores as indicators for the higher-order latent variable in a separate second-stage analysis, in this case using SMART-PLS software (Ringle et al., 2012). On a second-order construct level, eight hypotheses will be tested. Fig. 5 provides a summary of the outcomes of the PLS-SEM statistical method.
19
Figure 5. Summary PLS SEM
PLS-SEM method is performed using a bootstrapping of 500 subsamples to identify the significance levels. Using the R2 criteria of Chin (1998), an r2-value must be larger than 0.67 to be substantial explained by the exogenous constructs. Values around .333 are average, and values of .190 and lower weak. The variance of Behaviour Intent is explained for 0.675 and the variance of TRA Attitude for 0.33 and therefore considered both as substantial. Technology Acceptance and Perceived Risk are explained for 0.233 and 0.087 and therefore weak according to Chin (1998). PLS-SEM use standardized outputs for path coefficients. A P-value lower than 0.05 is used to be statistically significant within this study. Three out of four paths towards BI are considered as significant with a P-value lower than 0.001, TRA (0.235; p<0.000), Perceived Risk (-0.260; p<0.000) and TA (0.484; p<0.000). Three out of four hypotheses towards BI are confirmed since the influence of Institutional Trust on Behaviour Intent is not significant. Hence, H1, H2 and H7 are confirmed. Institutional Trust is confirmed to be statistically significant towards Perceived Risk (-.294; p<0.001) and TRA (0.574;p<0.000). Therefore H3 and H5 are confirmed. Perceived Risk show a statistical significant effect as well towards Technology Acceptance (-0.483;p<0.000), hence H7a is confirmed. Institutional Trust shows a positive significant indirect effect on BI as well (0.283;p<0.003) by influencing Perceived Risk and Therefore TA, which confirm H6. An overview of these results is depicted in table 4 and Figure 6.
Table 4. Hypotheses and Results
Hypotheses
Path coefficient
s P-value Results Indirect effects P-value H.1 Technology acceptance --> Behaviour Intent 0.484 0.000 Supported
20 H.3 Institutional Trust --> Perceived Risk -0,294 0.001 Supported -0,245
H.4 Institutional Trust --> Behaviour Intent -0,093 0.55 0.283 0.003
H.5 Institutional Trust --> TRA 0.574 0.000 Supported
H6 Institutional Trust --> Indirect effect TA < PR Supported 0.142 0.008 H.7 Perceived Risk --> Behaviour Intent -0,26 0.000 Supported -0,245 0.001 H.7 a Perceived Risk --> Technology Acceptance -0,483 0.000 Supported
Figure 6. Structural Model PLS-SEM path coefficient and r-squares
Besides testing the hypotheses on a second-order constructs structural model, the relations between the lower-order constructs are assessed as well within PLS-SEM and can be found within Appendix VII.
5. Conclusion and Discussion
Current literature models have not found to be satisfied to research whether consumers will share their financial information with third parties and the role of Institutional Trust. Therefore this study provides a new integrated model in order to answer this question. Using constructs of Technology Acceptance, Perceived Risk, Institutional Trust, Theory of Reasoned Action, an unique model has been created. This chapter discusses the main findings and outcome of the statistical methodology PLS-SEM.
21
5.1. Conclusions
The main research question ‘What is the (moderated) impact/influence of Institutional Trust (IT)
on Perceived Risk (PR) and Trust in Technology (TA) to predict the Behaviour Intent of consumers to share their personal financial information with third parties?’ can be answered as
follows:
Understanding why people adopt or reject a new technology has been proven one of the most challenging issues in the information system research (Swanson, 1988). Previous research has been mainly focused on Technology Acceptance Models to explain the Behaviour Intent of new technologies. The original UTAUT model for example explains 40 to 52% of the Behaviour Intent (Venkatesh, 2012). Adding Institutional Trust as a moderated construct to predict the BI of consumer’s results in 67% explanation of the BI of consumers. The proposed research model therefore is capable of explaining a high proportion of variance in consumer behaviour within the context of this research. This suggests that Institutional Trust can be seen as an important predictor to explain why people accept or reject new technologies.
However, the direct effect of Institutional Trust on both first order levels (effect of the separate constructs) as second order (all three constructs together) on BI is rather low and not significant. A direct effect of Institutional Trust therefore cannot be showed. On the other hand, the results do show a strong indirect or moderated effect of Institutional Trust on Behaviour Intent influencing the other constructs within the proposed research model: Perceived Risk, Technology Acceptance and the Consumer Attitude.
Perceived Risk has been shown to be an important indicator to explain or predict consumer’s behaviour within previous research (Bhatnagar et al. 2000; Cox 1964; Featherman et al. 2003). On a higher or second-order level, a direct negative significant effect of Perceived Risk on Behaviour Intent has been found. This suggest that how higher the Perceived Risk, the lower the Behaviour Intent of consumers. Perceived Risk strongly direct negatively affects the Technology Acceptance as well, which implies an indirect negative effect on BI by lowering the Technology Acceptance. The results of this study show both for the second-order construct TA as for first order constructs (PE and HM) a positive significant effect towards BI where the effect of PE can be considered as strong while the effect of HM is rather low. These results are consistent with those of previous research and suggest that TA still can be seen as a strong predictor of BI.
Thus, Institutional Trust directly positively influences Perceived Risk and indirectly increases the Technology Acceptance. Therefore it can be inference from present and current results that this indicate that although Institutional Trust does not influence directly the Behaviour Intent, the results still emphasize that Institutional Trust is an important factor for BI because it reduces Perceived Risk and increase Technology Acceptance. Within the context of this research, this implies that trust within the institute offering a new technology and not only
trust within the technology offered by the institute is important to predict the BI of consumers.
This suggest that since Institutional Trust decrease Perceived Risk, which increase Technology Acceptance and therefore the BI of consumers, a higher level of Institutional Trust will lead to a
22 higher BI. Beside, Institutional Trust explains 23% of the consumer attitude within the proposed research model. A strong significant effect of Institutional Trust on the Consumer Attitude has been found. The consumer Attitude positively influences the Behaviour Intent as well. Within the context of this research, this implies that Institutional Trust, including Reputation, Satisfaction and Communication affect the consumer's perception which lead to a higher Behaviour Intent. Earning consumers trust as an institute therefore may enhance consumers to use their technologies.
The indirect effect of Institutional Trust on BI has been researched before within the context of digital purchase environments. Pennington et al. (2003/04) found that “system trust plays an important role in the belief system by directly affecting trust in vendor and indirectly affecting attitudes and intentions”. Pavlou (2003) made a similar observation within his research where the relationship between trust and intention to transact was found to be as low as 0.18 (significant where p < 0.05). According to Ajzen & Fishbein (1980) favourable trust does not directly lead to positive Behaviour Intention, but instead first affects the attitude, which then brings about changes in intentions and behaviour. Those findings support the current findings of this research.
The results of this research are generalizable in three ways: Institutional Trust within this research has been shown to be a strong- moderated predictor for Behaviour Intent. Different sectors can use the research framework to find out whether Institutional Trust influences the Behaviour Intention to use or accept a new product or innovation. Second, from a marketing related perspective, Perceived Risk has been identified as an important factor decreasing the Technology Acceptance and therefore the Behaviour Intent. Perceived Risk can be reduced earning consumers Trust. Communication and Reputation are the main influencers, which affect Consumers Trust. Companies in other privacy and integrity related sectors could build on their Reputation and Communication strategies to earn this Trust and therefore influence the Behaviour Intent of consumers. Third, Institutional Trust positively influences the consumer attitude functions underlying the Theory of Reasoned Action. Other sectors can use these findings to research how they can enhance Institutional Trust by influencing the consumer Attitude.
5.2. Theoretical implications
This research contributes to the current state of the art in academic research in three ways. Prior research has lack to identified and research the role of Institutional Trust within Behaviour Intent in the field of e-commerce and e-banking. Technology Acceptance has been shown in previous research to be one of the strongest influencer for explaining the BI. However, this research suggest that adding Institutional Trust show a strong mediated effect on reducing Perceived Risk and therefore increasing the Technology Acceptance and the Behaviour Intent. Thus, Institutional Trust, how this trust is built and the impact on BI for adopting new technologies should therefore not longer be neglected.
Second, since PSD2 regulation is not implemented yet, this study can be used for future research towards the intention of sharing personal financial information. Both the role of
23 Technology Acceptance as Institutional Trust and how those are related to Perceived Risk should be investigated. Third, a new unique model is created elaborating on previous existing models to explain the Behaviour Intent of consumers. Since the research model explains 67% of variance of the Behaviour Intent of consumers to share their personal financial information, this model has shown to explain a high proportion of the variance of Behaviour Intent. Therefore, this research model contributes to the research areas of technology innovation and consumer behaviour.
5.3. Managerial implications
The results of this research provide several implications for business management. Knowing the BI of consumers to share their financial information and which factors are of importance provide insight for Business Management within the financial sector. First, the results show that Institutional Trust strongly influence the Consumer Attitude. Both Communication and Reputation increase the consumer Attitude regarding a specific behaviour. Therefore, communication strategies can be used to inform consumers regarding the legislation to increase Institutional Trust. Almost half of the respondents answer the question whether they trust banks with ‘agree’ or ‘strongly agree’, which implies that in general most people do trust Banks. More than half of the people think that Banks are well managed and ‘somewhat agree’ to ‘strongly agree’ with the question that they ‘usually believe what banks are saying’. Therefore it can be assumed that Banks in general have a good reputation and already earned consumers trust.
Second, respondents who were already aware of this legislation and their consequences had a stronger BI to share their personal financial information than people who weren’t aware (Appendix VIII). Hence, making consumers aware of this new legislation and their consequences increase a higher intention to share their information. Third, more than half of the respondent think that additional services, like managing all financial products in one environment including personal advice will be useful in their daily live. Banks should therefore pay attention to experiment with offer such services. However, almost 75% of the respondent to believe that sharing financial information led to possible fraud. Banks therefore should invest in reducing potential fraud related risk and prove to customers that they are aware of these possible risks and that it is trustworthy to share their personal financial information.
6. Limitations, Future Directions and discussion
Several limitations can be considered for this research. Due the limit of time, the questionnaire was conducted within my own professional and social network and consists of mainly Dutch high-educated people. Therefore, it would be interesting to determine whether the same results will be achieved when other countries of the European Union will be included as well. Since the PSD2 legislation affect the whole population within the European Union, future research should be focused on other countries and lower educational levels as well. The presented research model can be used to identify if the results can be generalized within other countries within the scope of this research.
Second, the data-analysing tool PLS-SEM used for conducting the first- and second order SEM analyses has proven some limitations. Since a student licence was used, not all features
24 were accessible including exporting and importing the second order constructs loadings. Hence, the latent variables score for lower-order constructs are obtained from SPSS. Since SPSS is a different statistic tool, some of the scores can be differ from PLS-SEM.
Third, it was rather complex to explain the legislation in a short amount of time, which results in the fact that many respondents quit the questionnaire unfinished. In the future, after PSD2 is implemented, more people will understand the legislation and their consequences, which make it easier to finish the questionnaire and lead to a bigger sample size. Some questions had to be omitted since their outer loadings were below 0.7. Hence, some constructs consist of only two items. Therefore their indicator reliability can be considered as weak. A recommendation for future research therefore will be to extend some of the incorporated constructs and add some additional questions to increase the indicator reliability.
Not all constructs on a first order level have been found significant influencing the other constructs. On a first-order level, only Attitude has been found significant influencing the BI of consumers explaining the attitude function. A possible explanation is given by Chang (1998) who stated that the Subjective Norm directly influences the Attitude and therefore indirectly influence the BI. Future research may incorporate other attitude constructs to explain the consumer attitude function.
For Perceived Risk, only one first order construct has been found significant directly influencing the BI. Both Performance and Privacy Risk show a weak non-significant effect. Privacy Risk did show a direct significant negative effect on Technology Acceptance, which implies an indirect effect on the Behaviour Intent of consumers. These phenomena can be explained by the little importance that most of the participants of the questionnaire are familiar with online banking and therefore projecting previous experience on possible future performance and privacy risk. For future research, incorporate other risk constructs therefore is recommended. This study is rather explorative and predictive since PSD2 is not implemented yet. Therefore it is difficult for the respondents to oversee possible consequences of this legislation. Future research can use the proposed research model after PSD2 is implemented to test whether the same results will be achieved.
25 7. References
Abroud, A., Choong, Y.V., Muthaiyah, S., & Fie, D.V.G. (2015). Adopting e-finance: decomposing the technology acceptance model for investors. Service Business, 9(1), 161-182 DOI: 10.1007/s11628-013-0214-x Ajzen, I. (1991). The theory of planned behavior. Organizational Behavior and Human Decision Processes, 50,
179-211
Ajzen, I. & Fishbein, M. (2000). Attitudes and the attitude–behaviour relations: reasoned and automatic processes. In Stroebe, W. and Hewstone, M. (eds), European Review of Social Psychology, Vol. 11. Chichester: Wiley, pp. 1–33.
Alshehri, M., Drew, S., & AlGhamdi, R. (2012). Analysis of citizens acceptance for e-government services: applying the UTAUT model. Presented at the IADIS International Conference Theory and Practice in Modern Computing and Internet Applications and Research
Anderson, J.C. & Narus, J.A. (1990). A model of distributor firm and manufacturer firm working partnerships'. Journal of Marketing, , 54, 42-58
Archer, N. & Yuan, Y. (2000). Managing business-to-business relationships throughout the e-commerce procurement life cycle. Internet Research, 10(5). 385-95
Armstrong. A. (2012). Restoring trust in banking. National Institute Economic Review, 22, R4–R10
Ba, S. (2001). Establishing online trust through a community responsibility system. Decision Support Systems, 31(3), 323-336
Ball, D., Coelho, P.S., & Machás, A. (2004). The role of communication and trust in explaining customer loyalty: An extension to the ECSI model. European Journal of Marketing, 38 (9/10), 1272-1293, doi: 10.1108/03090560410548979
Barber, B. (1983). The Logic and Limits of Trust. New Brunswick: Rutgers University Press.
Bachmann, R. (2001). ‘Trust, power and control in trans-organizational relations. Organization Studies, 22, 337–65. Bauer, R. (1960). Consumer behavior as risk taking. paper presented at the Dynamic Marketing for a Changing World,
Proceedings of the 43rd Conference of the American Marketing Association, Chicago, IL, pp. 389-98 Becker, J. M., Klein, K., & Wetzels, M. (2012). Hierarchical Latent Variable Models in PLS-SEM: Guidelines for
Using Reflective-Formative Type Models. Longe Range Planning, 45(5-6), 359-394)
Bettman, J.R. (1973). Perceived Risks and Its Components: A Model and Empirical Test. Journal of Marketing Research, 184–190.
Bettman, J.R., Luce, M.F. & Payne, J.W. (1998). Constructive consumer choice processes. Journal of Consumer Research, 25, 187–217.
Bhatnagar, A., Misra, S., & Rao, H.R. (2000) On risk, convenience, and internet shopping behavior. Communications of the ACM, 43 (11), 98–114
Chang M.K. (1998). Predicting unethical behavior: a comparison of the theory of reasoned action and the theory of planned behavior. Journal of Business Ethics, 17, 1825-1834
Chen, Q. & Wells, W.D. (1999). Attitude toward the site. Journal of Advertising Research, 39, pp. 27–37. Cheng T, Lam D.Y.C., & Yeung A.C.L. (2006) Adoption of Internet banking: an empirical study in Hong Kong.
Decision Support Systeen 42(3):1558–1572
Chin, W. W. (1998). Issues and opinion on structural equation modeling. MIS Quarterly, 22 (1), vii-xvi
Chin, W.W. (1998). The partial least squares approach to structural equation modeling. Modern Methods for Business Research, Marcoulides, G.A. (ed.), Lawrence Erlbaum Associates, Mahwah, NJ, I295–1336.
Costigan, R. D., Ilter, S. S. & Jason, B. J. (1998). A multi-dimensional study of trust in organizations. Journal of Management Issues, 10, 303–17.
Coulter, K.S. & Coulter, R.A. (2003). The effects of industry knowledge on the development of trust in service relationships. International Journal of Research in Marketing, 20(1), 31-43.
Cox, D.F., & Rich, S.J. (1964). Perceived Risk and Consumer Decision Making. Journal of Marketing Research,1(November), 32–39
Davis, F.D. (1989). Perceived usefulness, perceived ease of use, and user acceptance of information technology. MIS Quarterly, 13 (3), 319–339
D’Alessandro, S., Girardi, A., & Tiangsoongnern, L. (2012). Perceived risk and trust as antecedents of online
purchasing behavior in the USA gemstone industry. Asia Pacific Journal of Marketing and Logistics, 24(3), 433-460
