The future of retail banking:
Will FinTech companies replace traditional banks?
A consumers’ perspective
Abstract:
The competitiveness of emerging FinTech companies will increase as a result of a digitalizing world, technological innovations, and future legislative changes. As of 2018, these FinTech companies are allowed to ask consumers to share their bank account data with them. The purpose of this paper was to find out if there is a relationship between customers’ Technology Readiness, measured in terms of Optimism, Innovativeness, Discomfort, and Insecurity, and customers’ intention to share their data. Additionally, the relationship between Customers’ intention to share data and Customers’ propensity to switch was investigated. Moreover, the effect of Relationship Quality, measured in terms of Satisfaction, Trust, and Commitment, on this relationship was examined. By administering an online survey among 240 Dutch
customers of Dutch banks, enough evidence was found to state that Customers’ Technology Readiness positively influences Customers’ intention to share data with FinTech companies, and a stronger intention to share data will increase a Customer’s propensity to switch. Additionally, the results revealed that Relationship Quality moderate this relationship negatively. In general, Customers’ intention to share financial data with FinTech companies was weak. This study offers insights into consumers’ behaviour towards emerging FinTech companies, which has not been investigated before. Finally, this paper concludes with the contributions and implications of the findings of this study, and suggests further research directions.
Key words: Technology readiness, Relationship quality, Customers’ intention to share data, Customers’ propensity to switch, FinTech companies
DDM - Advanced International Business Management and Marketing
Newcastle University Business School
University of Groningen
Thesis
Student: Dirk Gerco Zeegers
Student number: B6023961 / S2361272
Supervisors: dr. E. Alamanos and prof. dr. A.R. Muller Submission date: 5th of December, 2016
Acknowledgements
I would like to thank my supervisors Eleftherios Alamanos and Alan Muller for giving me the needed guidance and support during the writing process of my dissertation, and providing me with concise feedback when I asked for it. This definitely enabled me to increase the quality of this paper.
Writing this thesis has opened my eyes and even changed my plans for my future career. Moreover, I am glad to announce that I will be doing an internship at the FinTech and Innovation department of one of the largest banks in the Netherlands from the 12th of
Table of Contents
1. Introduction... 5
2. Literature review... 8
2.1 The financial services sector... 8
2.2 Theoretical background... ... 10
2.2.1 Customer’s intention to share data with FinTech companies... 10
2.2.2 Technology Readiness... 11
2.2.3 Customer’s propensity to switch to FinTech companies... 14
2.2.4 The moderating role of Relationship quality... 15
2.2.5 Conceptual model... 17 2.3 Conclusion... 18 3. Methodology... 19 3.1 Data collection... 19 3.2 Sample... 19 3.3 Measurement instrument... 20 3.4 Control variables... 21 3.5 Data analysis... 22 3.6 Ethical considerations... 22 3.7 Conclusion... 23 4. Findings... 24 4.1 Profile of respondents... 24 4.2 Descriptive results... 24 4.3 Reliability analyses... 25
4.3.1 Customer’s intention to share data with FinTech companies... 25
4.3.2 Technology Readiness... 26
4.3.3 Customer’s propensity to switch to FinTech companies... 27
4.3.4 Relationship quality... 27
4.4 Validity analyses... 29
4.5 Hypotheses testing... 31
4.5.1 Customers’ intention to share data with FinTech companies... 31
4.5.2 Customers’ propensity to switch to FinTech companies... 34
4.5.3 Testing the model... 35
4.5.4 The moderating role of Satisfaction, Trust, and Commitment.... 36
4.6 Overview of findings... 39
5. Discussion... 40
5.1 Customer’s intention to share data with FinTech companies... 40
5.2 Customer’s propensity to switch to FinTech companies... 42
5.3 The moderating role of Satisfaction, trust, and Commitment... 44
5.4 Conclusion... 44
6. Conclusion... 45
6.1 Contributions... 45
6.2 Managerial implications... 46
6.3 Limitations and future research... 47
References... 48
Appendices... 58
Appendix A: Survey... 59
Appendix B: Descriptive statistics... 62
Appendix C: Reliability and Validity analyses... 64
Appendix D: Independent samples t-test... 67
Appendix E: Structural equation modeling... 68
Appendix F: Minutes of conversations with supervisors... 69
1. Introduction
In recent times, the world has been moving to a more digital globalized world, one that is more interconnected than ever. The world is packed with new technological innovations that will change the way business is done. “Banking has historically been one of the business sectors most resistant to disruption by technology” (McKinsey & Company, 2015, p.1). However, a wave of digital disruption can also be seen in the financial services sector now. Financial technology (FinTech) based companies, e.g. the well-known Dutch company Adyen and the German Kreditech, are reshaping the sector by facilitating transactions and offering loans to individuals (PwC, 2014). The competitiveness of these FinTech companies could grow as a result of new European legislation called Payment Services Directive 2 (PSD2), which will be introduced in 2018. PSD2 require traditional banks to share customer’s bank account data with these third parties when a customer has given permission to do so. The aim of PSD2 is to improve payment efficiency and increase competition in the financial services sector through new market entrants like FinTech companies (Accenture, 2016). In contrary to the often requested contact information, e.g. e-mail and telephone number, by other online companies like web shops or social media websites, consumers will have to share their personal bank account data, which is considerably more sensitive information, if they want to use the offered services of a FinTech company. However, this raises the question: are
consumers willing to share their financial data with these FinTech companies?
The amount of possibilities that enable customers to share data has increased over the past decade, and so has the amount of research related to this topic. Previous research about customer’s intention to share data with other companies have shown influencing factors like privacy concerns, monetary incentives, the kind of requested customer data, and feelings of insecurity (Westwin, 2003; Cvrcek et al., 2006; Krasnova, Hildebrand and Guenther, 2009; Van Slyke et al., 2006; Krasnova et al. 2010; Phelps, Nowak and Ferrel, 2000). Furthermore, based on the theory of planned behaviour there is a positive relationship between individuals’ intention and their actual behaviour (Ajzen, 1991). With regards to customer-specific factors, previous research has tried to understand customers’ technology-related behaviour by
applying the technology acceptance model of Davis et al. (1989). However, the more recent Technology Readiness model of Parasuraman (2000) has been used far less frequently. According to Parasuraman, technology readiness can be defined as “people’s propensity to embrace and use new technologies for accomplishing goals in home life and at work” (2000, p. 308). This propensity to use new technologies is determined by customer’s level of
amount of customers with a high Technology Readiness. These people have weak feelings of Discomfort and Insecurity towards new technologies, therefore, are probably more willing to share their financial information with a FinTech company, which unlike banks do not offer the possibility for a face-to-face conversation. Therefore, the model of Parasuraman (2000) was applied to analyse both the effects of a customer’s Technology Readiness inhibitors and contributors on Customers’ intention to share data with a FinTech company. According to Edelman (2010), customers rely heavily on digital channel interactions nowadays, but their expectations are not fully satisfied. Moreover, the customers of a bank become more open to competitive advancements and switch more quickly to another bank (Brun et al., 2014). Previous research has shown that successful relationship marketing results in increased levels of customer retention (Roberts et al., 2003). The strength of a relationship between a
customer and a company can be measured with a key indicator called Relationship Quality (Garbarino and Johnson, 1999). Relationship quality consists of the dimensions: Satisfaction, Trust and Commitment (Crosby et al., 1990; De Wulf et al., 2001).
A survey completed by 560 global banking executives has shown that 70% of them think it is very important to form an idea of what the banking market in 2020 looks like and to understand how FinTech companies could impact the banking system (PwC, 2014). Retail banking, also known as consumer banking, has been identified as the most likely financial sector to be disrupted by FinTech companies over the next five years (PwC, 2016). However, research on this topic is very limited. Strikingly, little empirical research has been done on the impact of Technology Readiness on customer’s intention to share data within the financial services sector, whereas Parasuraman (2000) calls for more studies to evaluate the
generalizability of his Technology Readiness model. It has been proven that Technology Readiness positively influences customers’ online behaviour, but empirical findings are scarce (Zeithaml et al., 2002) and contradictory (Liljander et al., 2006). In order to fill this gap in literature and to make a contribution to the limited body of knowledge about the Technology Readiness model and customer’s intention to share data with a FinTech companies, this study tries to answer the research question ‘What influences customers’ intention to share financial data with FinTech companies and their propensity to switch to these companies?’ To answer this research question, this paper examines the influence of customer-specific factors on intentions to share personal bank account data with FinTech companies, and the effect of a customer’s intention to share data and relationship quality on a customer’s propensity to switch to these companies.
possibility of FinTech companies to disrupt the retail banking sector depends on the customer’s intention to share data with these companies in the future. Additionally, if customers do want to share their data with these companies, it will be relevant for banks to know how they could prevent their customers from switching. Therefore, this study
incorporated relationship quality as moderator, which could contribute to the body of
knowledge in the area of relationship marketing. An online survey has been distributed among Dutch customers of Dutch banks to gather information about these issues.
2. Literature review
This literature review is divided into two sections. The first section will review the financial services sector and the current and expected developments in this sector to create a better understanding of why technology and legislative changes could disrupt the banking sector. The second section will address the theoretical background of the key constructs of this study and the corresponding conceptual model.
2.1 The financial services sector
According to PwC “consumer banking is on the verge of disruption, much of which is led by the disaggregation of simple products and service offerings,” which gives a good impression of the current situation in the financial services sector (2016, p. 3). New entrants, like FinTech companies, are focussing on offering targeted innovative solutions, leaving traditional banks at a deficit, while developing only incremental improvements to their services (PwC, 2016; Accenture, 2015). These FinTech companies see opportunities in disaggregating the current services of traditional banks, focusing on one specific component and developing a highly effective solution to meet the current expectations of customers. They try to really understand and meet the evolving needs of customers and try to attract them by offering convenient product designs and services via non-traditional channels like social media (PwC, 2016). According to the report of PwC (2016), the most valuable services FinTech companies will offer are solutions for people who fail to get a loan because they have a bad or even no credit score, a platform for Peer-to-peer lending aimed at people who are unable to get a loan from traditional sources, and personal finance management tools which people could use to manage their income and expenses. So, it is not surprising that almost 80% of the participants of the 2016 PwC global FinTech survey, who were mainly executives of traditional banks from 46 countries, think some parts of their business are at risk due to these FinTech companies.
Moreover, the financial crisis had a negative influence on customers’ trust in traditional banks, which could be beneficial for FinTech companies (McKinsey & Company, 2015; Simon, 2009).
from forming a coherent view of their individual customers. However, this study was performed in 2011, and a lot has changed since then. Nowadays, banks have adopted new solutions to conquer their problems and improve their operations. For example, they expose their payment systems through applications programming interfaces (APIs), a quick way to exchange information between two different systems, to third parties. These APIs enable and motivate third parties to develop new services, which a banks could provide to her customers via her online channels like a mobile banking application (PwC, 2016).
Unfortunately, the competitiveness of FinTech companies will presumably rise from the beginning of 2018, as a result of the introduction of a new law called Payment Services Directive 2 (PSD2) (McKinsey & Company, 2015). “The PSD2 is a data and technology-driven directive which aims to drive increased competition, innovation, and transparency across the European payments market while also enhancing the security of Internet payments and account access” (Accenture, 2016, p. 8). As a result of this, FinTech companies could use the authentication details provided by the bank’s customers to connect with their bank’s system and data. They could now act as an aggregator of data relating to an individual’s accounts held across one or many different banks of other financial institutions and use this data to provide the customer details about his or her combined balance across all accounts, as shown in figure 1, or give tailored analyses of spending patterns (Accenture, 2016). From 2018, these new developments in legislation will require banks to provide customer’s account information to third parties, such as FinTech companies, granted the customer gives his or her permission (Accenture, 2016). Furthermore, banks are required to have security systems for this information exchange with third parties and are required to implement secure customer authentication for electronic payments and online account access. This strong customer authentication is defined as “the use of at least two independent factors from two different categories as part of the authentication process across the categories of ‘knowledge’ (e.g. a password), ‘possession’ (e.g. a personal device) and ‘inherence’ (e.g. a fingerprint)” (Accenture, 2015, p. 10). So, increased competition from FinTech companies seems
inevitable as they will try to persuade the customers of traditional banks to share their bank account information with them to be able to offer a bank’s customers a better or new service in exchange (Accenture, 2015).
Obviously, the consequences of this future change in legislation with regard to e.g. the competitiveness of banks and the switching behaviour of customers are still unknown. According to the report of Accenture (2015), PSD2 will lead to a greater use of FinTech companies’ services by customers. However, perhaps customers do not even want to share their bank account information and therefore do not wish to use the services of FinTech companies. Still, it is expected that if banks continue to operate in the same way their revenues in the area of “consumer finance, mortgages, lending to small- and medium-sized enterprises, retail payments and wealth management, which is 10% to 40% of banks’ total revenues, could be at risk by 2025” (McKinsey & Company, 2015, p. 5). So, today banks will have to compete with both other banks and FinTech companies to be the customer's first choice. The conceptual model, as shown in the theoretical background section, will be tested to increase our knowledge about customer’s intention to share their personal bank account data with these emerging FinTech companies and their propensity to switch to these companies.
2.2 Theoretical background
In this section of the literature review, the theoretical justification for the conceptual model will be described, and the following key constructs of this research will be defined
consecutively: Customers’ intention to share data with FinTech companies, Technology Readiness, Customers’ propensity to switch to FinTech companies, and Relationship quality. A comprehensive view of what is already known about these constructs from previous research in different areas e.g. marketing, CRM, psychology and financial services will be outlined, and the hypotheses according to each construct will be mentioned. Finally, the conceptual model will illustrate the suspected positive or negative relationships between these constructs.
2.2.1 Customers’ intention to share data with FinTech companies
people would enable access to their location data for non-commercial use for 20 British pounds. Krasnova, Günther, and Hildebrand (2009) conclude that people would pay between 14 and 17 Euro a year for more privacy of their online social network data. Furthermore, information privacy concerns have been found to negatively influence a customer’s willingness to conduct online transactions and a critical barrier for people to disclose
information on social network sites (Van Slyke et al., 2006; Krasnova et al., 2010). Recently, it also has been proven that customers’ desire to give information in virtual communities positively influences customers’ information sharing behaviour (Liou et al., 2015). Moreover, according to Evens and Van Damme (2016), people are willing to share personal information with news organisations in exchange for free news content, so when the benefits outweigh the disadvantages (Roeber et al., 2016). Furthermore, loyalty programs have a positive influence on consumers’ willingness to share personal information (Jai and King, 2016). Strikingly, one study showed that consumers would share their personal information with online search engines, e.g. Google, rather than with banks (Roeber et al., 2015).
According to Ajzen and Fishbein’s (1980) theory of reasoned action, an individual’s attitude influences his or her intention, and a positive attitude leads to a stronger intention. Moreover, several studies have demonstrated this: e.g. Curran, Meuter, and Surprenant (2003) have shown with their study that a customer’s intention to use online banking is influenced by a customer’s overall attitude toward self-service technologies influence. So, it is suspected that a customer’s intention to share data and use banking services of FinTech companies is also influenced by customer’s attitude towards new technologies.
2.2.2 Technology Readiness
Nowadays, new technologies and services are introduced and offered to consumers at a high pace, which seems to be a positive development. However, according to Mick and Fournier (1998), an increasing amount of new technologies does not only lead to customer satisfaction but can also lead to increasing customer frustration and disillusionment. Parasuraman (2000) described these negative concerns as psychological barriers that will have a negative influence on customer’s enthusiasm to embrace new technologies and affect people’s technology
Insecurity. According to Dabholkar (1996), customers could simultaneously have positive and negative feelings about technology-based services and these feelings can coexist. Therefore, instead of aggregating the four dimensions into one composite variable, this study examined the influence of each separate dimension on Customers’ intention to share data with FinTech companies. Technology Readiness is composed of the following four dimensions, as defined by Parasuraman (2000, p. 311).
(1) Optimism: a customer’s positive view that new technologies will offer “increased control, flexibility and efficiency” advantages to their daily life practices.
(2) Innovativeness: a customer’s “tendency to be a technology pioneer and thought leader”. It measures the extent to which customers think they are always the first trying out new technology-based services and if they believe others consider them as opinion leaders on technology-related issues.
(3) Discomfort: a customer’s feeling of “a lack of control over new technologies and a feeling of being overwhelmed by it.”
(4) Insecurity: a customer’s feeling of “distrust towards new technologies and a sceptical view about a new technology’s ability to work properly”. Although this dimension is somewhat related to discomfort, insecurity focuses on specific aspects of technology-based transactions instead of a lack of comfort with technology in general. This feeling of insecurity discourages customers to share personal information or do business with companies that could only be contacted via digital channels.
Both Optimism and Innovativeness are the real drivers of people’s propensity to embrace and use new technologies. According to Davis (1989), perceived usefulness, which is the
customer’s belief that technology will improve his or her performance, affects a customer’s intention to use new technologies. This idea that technology can improve his or her
performance can be related to the Optimism dimension of the Rechnology Readiness model, as these technology-optimistic people also view technology as something that could offer advantages to their daily life. Moreover, people’s level of innovativeness positively influences their adoption of new technologies, e.g. Wireless Internet, and negatively influences their perceived complexity about new technologies (Lu et al., 2005; Karahanna et al., 1999). Therefore, it is suspected that Optimism and Innovativeness will positively influence Customers’ intention to share data with FinTech companies.
This anxiety could be related to the Technology Readiness dimension of Discomfort and might be an inhibitor of Customers’ intention to share data with FinTech companies.
In 2000, Phelps, Nowak and Ferrel showed that the vast majority of their respondents would never share their annual household income or credit card data with marketers because of privacy concerns. However, this research has been done 16 years ago and focused on customer’s willingness to share data with marketers and not with nowadays emerging FinTech companies. So customers’ current opinion about sharing financial data could have changed. Privacy concerns could be related to the Technology Readiness construct of insecurity, as customers with a high score on insecurity are afraid their information will be seen by the “wrong” people. It is suspected that customers with a high score on insecurity only trust their bank with their personal data and do not want to share their bank account information with FinTech companies, as they think it is unsafe to do any kind of financial business online.
Parasuraman (2000) has shown that consumers’ Technology Readiness is related to making use of technology-based services and future desirability of these kinds of services. Furthermore, the Technology Readiness model has been researched as a determining factor in technology adoption and usage (Lam et al., 2008; Westjohn et al., 2009). The Technology Readiness model has also been applied in multiple studies with different contexts like the Construction industry (Jaafar et al., 2007), Online services (Taylor et al., 2002), Healthcare industry (Rosen et al., 2003) and Educational choices (Hendry, 2000). So, it has been shown that this model is highly applicable in different contexts. However, it has never been applied in the financial services sector, although this sector certainly has some significant differences. For example, according to Roboff and Charles (1998), consumers have a weak understanding of the security risks of online banking services, and they assume that their bank will protect them. Furthermore, they argue that consumers think their bank will be concerned about their privacy. Nowadays, more and more consumers are willing to interact with their bank online and get mobile access to financial services, but they also want to control what kind of data is collected and for what purpose their data will be used (Kobsa, 2002). However, sharing financial data with the other companies than a bank could lead to a greater chance of monetary loss due to transaction errors or bank account misuse (Lee, 2009).
H1a: Optimism will positively influence customers’ intention to share data with FinTech companies
H1b: Innovativeness will positively influence customers’ intention to share data with FinTech companies
H1c: Discomfort will negatively influence customers’ intention to share data with FinTech companies.
H1d: Insecurity will negatively influence customers’ intention to share data with FinTech companies.
2.2.3 Customers’ propensity to switch to FinTech companies
Banks are concerned with customers’ switching behaviour because normally this reduces their market share and profit (Ennew and Binks, 1996; Satish et al., 2011). Bansal and Taylor (1999) stated that customers’ switching behaviour consist of replacing the current service provider with another service provider. Switching behaviour reflects the customer’s decision to stop purchasing a particular service or patronizing the service firm completely (Boote, 1998). According to Edelman (2010), customers rely heavily on digital channel interactions nowadays, but their expectations are not entirely satisfied. Moreover, customers have become more open to competitive advancements and switch more quickly to another bank (Brun et al., 2014). Of course, there are many reasons why customers could switch. According to Stewart (1998), customers could switch because of charges and their implementation,
facilities and their availability, provision of information and confidentiality, and service issues relating to customers’ treatment. Gerrard and Cunningham (2000) investigated the switching behaviour of customers in the Asian banking market and found that switching behaviour in influenced by service failures, pricing, and inconvenience. A similar study was conducted in India, which showed that besides service failures and pricing, also a bank’s reputation and service quality, customer satisfaction, competition and customer commitment have a significant effect on customers’ switching behavior (Vyas and Raitani, 2014). Furthermore, other significantly tested determining factors of customer’s switching behaviour are
customers’ attitude towards switching (Bansal and Taylor, 1999) and the perceived costs of switching (Kim et al., 2013; Wang et al., 2011).
According to the expectancy theory, people will behave in ways that maximize their positive outcomes and minimize negative outcomes (Vroom, 1964; van Eerde and Thierry, 1996). Furthermore, based on Ajzen’s theory of planned behaviour (1991), an extension of Ajzen and Fishbein’s theory of reasoned action (1980), there is a positive relationship
and Zhang (2010), the behavioral intention has a significant influence on a person’s technology usage. Against this research background, it is assumed that customers will be more likely to switch to FinTech companies to use their innovative services if they have a stronger intention to share their data with these parties.
H2: Customers’ intention to share data with FinTech companies will positively influence Customers’ propensity to switch to Fintech companies.
2.2.4 The moderating role of Relationship quality
Previous research has shown that successful relationship marketing results in increased levels of customer retention and this strategy could also provide companies with a sustainable competitive advantage (Anderson and Narus, 1990; Roberts et al., 2003, Neto et al., 2011). However, understanding the relationship from the customer’s point of view has been
identified as an under-researched area of marketing, and given the importance of relationships in a financial service context, this study aims to research the quality of the relationship
between the bank and her customers from the customer’s perspective (De Wulf et al., 2001; Gwinner et al., 1998; Hennig-Thurau et al., 2002). The strength of a relationship between a customer and a company can be measured with a key indicator called relationship quality (Garbarino and Johnson, 1999). Roberts et al. (2003, p. 191) defined relationship quality as “a measure of the extent to which consumers want to maintain relationships with their service providers.” It measures the relationship quality between a firm and her customers, and views fulfilling customers’ needs as central to relationship success (Hennig-Thurau et al., 2002). The global construct of Relationship Quality consists of three intimately interconnected dimensions: Satisfaction, Trust and Commitment (Crosby et al., 1990; De Wulf et al., 2001; De Cannière et al., 2009; Athanasopoulou, 2009).
(1) Satisfaction: According to Parsons (2002), customers will be satisfied when the performances of a company meet their expectations. Lemon et al. (2002) have proven that satisfaction is a key determinant in a customer’s decision to continue or drop a relationship and be a loyal customer or not. So, relationship satisfaction can be defined as the customer’s satisfaction with the relationship, which is different than the customers’ satisfaction with the company in general (Palmatier et al., 2006).
(2) Trust: Trust can be defined as an individual’s confidence in the intentions,
in buyer-seller relationships between a company and a customer (Gundlach et al., 1995). So, trust can be seen as a key ingredient for building and maintaining successful long-term relationships (Berry, 1995; Singh and Sirdeshmukh, 2000).
(3) Commitment: The most common and used definition of commitment is the one of Moorman et al. (1992, p. 316) who defined commitment as “an enduring desire to maintain a valued relationship”, and measured it as a global construct. Dwyer et al. (1987) suggest that commitment will continue when the relationship has advantages for both sides. Furthermore, commitment is not only an essential ingredient for a long-term relationship but also a
customer’s expression of willingness to stay with a certain company (Moorman et al., 1993). So, a high-quality relationship consists of high levels of Satisfaction, Trust, and Commitment (De Wulf et al., 2001). Two often proven positive behavioural outcomes of a perceived high relationship quality are loyalty, e.g. expressed by customer’s repeat purchase behaviour, and positive word of mouth communication to other people within their social network (Hennig-Thurau et al., 2002).
However, it is important to determine what factors results in customers continued use of the services of their bank rather than the services of FinTech companies. According to Dwyer et al. (1987), a high level of satisfaction between exchange partners could result in the exclusion of other parties that provide the same benefits. Furthermore, previous research has shown that customers’ satisfaction with services increases their perceptions of the
disadvantages of switching suppliers (Gundlach et al., 1995; Patterson and Smith, 2003). Consequently, it was proposed that customers who are satisfied with their relationship with their bank will perceive higher switching costs and will use the services of their bank rather than switch to other unknown third parties’ services.
According to Zhou (2011), customers’ initial trust positively influences their intentions to use mobile banking. Similarly, Gu et al., (2009) indicate that when customers trust their bank, they can see the value of mobile banking and it encourages them to use it. Moreover, Gefen et al. (2003) associate high levels of trust with high levels of intention to use and Aydin and Ozer (2005) suggest that trust leads to a positive behavioural intention towards the other party. Also, Lau and Lee (1999) have proven that trust in a company favour a
customer’s positive buying intentions. According to Pavlou (2003), the trust will positively influence a customer’s attitude, which affects behavioural intention by minimizing fears about opportunistic behaviour. This opportunistic behaviour entails “unfair pricing, conveying inaccurate information, violations of privacy, unauthorized use of credit card information and unauthorized tracking of transaction” (Gefen et al., 2003, p. 55).
customers can perceive commitments made by individual advisors during a lifetime as indicative of a company’s practices and routines. So, whereas customers could have different individual financial advisors during their relationship with the bank, they still could have formed a stable and enduring relationship with the bank. In addition, customers who are committed to a relationship with a company are less likely to become an activist against the company (Hoyer and MacInnis, 1997). So, they are unlikely to complain about the company to other people or external agencies.
This study applies the Relationship Quality construct and investigates if Satisfaction, trust, and commitment negatively moderate the relationship between Customers’ intention to share data with FinTech companies and Customers’ propensity to switch to FinTech
companies.
H3a: Satisfaction will negatively moderate the relationship between Customers’ intention to share data with FinTech companies and Customers’ propensity to switch to FinTech
companies.
H3b: Trust will negatively moderate the relationship between Customers’ intention to share data with FinTech companies and Customers’ propensity to switch to FinTech companies. H3c: Commitment will negatively moderate the relationship between Customers’ intention to share data with FinTech companies and Customers’ propensity to switch to FinTech
companies.
2.2.5 Conceptual model
Figure 2. Conceptual model
2.3 Conclusion
3. Methodology
The previous chapter described and reviewed the financial services sector and relevant literature used to develop the hypotheses. In this chapter, the research design of this study is explained in order to answer and test these hypotheses. Firstly, more information is provided about how the quantitative data was collected and about the participants in this study.
Subsequently, the measurement instrument and control variables are outlined. Finally, more information is provided about the used software to analyse the data, and the ethical
considerations of this study are described.
3.1 Data collection
Quantitative data for this study has been collected using an online survey designed with the online survey tool called Qualtrics. Thereafter, this survey was disseminated via E-mail and Facebook. Advantages of an online survey are speed, low costs and efficiency (Cobanoglu & Cobanoglu, 2003). Furthermore, a survey method was applied because this enables a
researcher to use a small group of people to make inferences about the total population, which in this study consists of all Dutch people in the age range of 16 -55+ with a bank account (Holton and Burnett, 1997). A brief introduction was given about the goal of this research before the administration of the survey. Furthermore, the participants were allowed to answer the questionnaire by using a mobile device of their choice as long as it was able to connect with the Internet. A pilot test of the questionnaire has been done among five people in the age range of 16-34 and five individuals in the age range of 35-55+ to check whether the questions are clear and interpreted by everyone similarly, before the actual questionnaire was
distributed. These people were also asked to comment on the length and format of the questionnaire. Minor suggested wording changes were performed.
3.2 Sample
The participants in this study were Dutch customers of Dutch banks in the age range of 16 to 55+ years old. They were only included in the sample if they were a customer of a certain bank for at least 6 months to make sure they had developed a good opinion about the
to obtain a diverse participant group of sufficient size in a limited amount of time (Sadler et al., 2010). In total, over a two-week period, 318 customers of Dutch banks completed the online survey. Of these, 78 responses were excluded due to missing data; this resulted in a usable sample size of 240.
3.3 Measurement instrument
To test the conceptual model, a survey instrument was developed based on previously published empirical research. As the study was conducted in a Dutch-speaking environment, all measures previously developed in English were translated into Dutch using a standard back-translation procedure to make sure all questions were understood properly by Dutch respondents (Brislin, 1970). Since the survey is designed to determine customer’s attitudes towards new technology, a Likert-type scale was deemed the most useful approach (Kerlinger, 1986). Respondents were asked to respond on a 5-point Likert-type scale to the questionnaire items. To measure the construct of Technology readiness a Likert scale was used ranging from 1= Describes me not at all, to 5= Describes me extremely well. The construct of Customer’s intention to share data with FinTech companies was measured by using a Likert scale ranging from 1= Extremely unlikely, to 5= Extremely likely. The other constructs were measured on a Likert scale ranging from 1= Strongly disagree to 5= Strongly agree (Clark et al., 1998). A 5-point Likert scale was applied because coefficient alpha reliability with Likert scales with more than five points will decrease (Lissitz & Green, 1975). The survey started with a brief and clear explanation of the future legislative amendments, with regards to PSD2, to inform the respondents about this future scenario. Also, examples were given of the kind of innovative services FinTech companies could offer, so participants were able to make an informed choice about sharing their bank account data with these parties.
3.4 Control variables
In order to accurately test the hypotheses that relate the Technology Readiness dimensions to Customers’ intention to share data with FinTech companies, and the latter with Customers’ propensity to switch to FinTech companies, this study controlled for several other factors that could influence these relationships.
Firstly, it has been proven that people of generation X, born between 1965 and 1979, and generation Y, born between 1980 and 2000 have different preferences and values (Parment, 2013; Gibson et al., 2009). Generation Y, for example, grew up with technology, and the Internet and social media have a major impact on their lives (Nusair et al., 2013). Furthermore, their behavior could be a good first impression of how the customers of the future will behave (Bolton et al., 2013). According to Lee et al. (2010), younger people are also more innovative than older people. Additionally, previous research has shown that older people perceive online banking services as more difficult to use than younger people (Porter and Donthu, 2006; Cruz et al., 2010). Therefore, given the great likelihood that people from different generation have different levels of experience and difficulty with using new technologies and online services, which may indicate that young people are more eager to share their data with FinTech companies, the first control variable controlled for Age. Respondents’ age was measured on a 5-point scale ranging from 1= 16-24 to 5= 55+.
Secondly, previous studies have shown that Education level is negatively related with people’s anxiety feelings about technology (Igbaria and Parasuraman, 1989). Moreover, men have a stronger propensity to use novel and innovative technologies than women, and men are more likely to use mobile banking services than women (Chau and Hui, 1998; Nysveen et al., 2005; Koenig-Lewis et al., 2010). This could imply that higher educated people and men have stronger intentions to share their financial data in order to make use of the offered innovative services of FinTech companies than lower educated people and females. Therefore, this study controlled for Education level, measured on the following 4-point scale 1=High School, 2= Vocational Education, 3= University of Applied Science and 4=University, and for Gender: 1= Man, 2= Woman.
5= More than five years.
Finally, as proven by previous research, a great variety and number of products and services has a positive influence on customer satisfaction and loyalty (Srinivasan et al., 2002; Cho and Park, 2001; Pick and Eisend, 2014) and a customer’s propensity to switch (Neto et al., 2011). Furthermore, customers will be more loyal to a company when it is able to deliver a consistently superior service to them and as a result, customers will desire to use more services (Boshoff, 2005). Additionally, Mavri and Ioannou (2008) have researched customers’ switching behaviour in Greece and found that the quality of a bank’s service portfolio, both products and services, has a positive effect on decreasing switching behaviour. This couldimply that Customers’ propensity to switch to FinTech companies will be lower when these customers are already content with the current service portfolio of their bank. Therefore, this study applied the control variable Service portfolio, which was measured with items from Bahia and Nantel (2000) their Bank Service Quality instrument, which is specially designed for the banking sector.
So, the control variables Age, Gender, Education level, Duration of relationship and Service portfolio were applied to this study to purify the observed relationships among the variables of interest.
3.5 Data analysis
The collected data for this study was analysed by making use of IBM SPSS 23.0. In order to create a valid dataset, all responses with missing values were deleted. Then, by making use of descriptive analyses, the profile of respondents was outlined. Multiple regression analyses, including the control variables, were performed to test if the hypotheses were significant. These tests are appropriate statistical analyses because the variables of this research were measured at ordinal and interval level. Furthermore, structural equation modeling was applied to describe the goodness of fit of the conceptual model. However, these analyses and the corresponding results are explained in more detail in the findings section.
3.6 Ethical considerations
was used for research purposes only and no personal information had to be given to ensure anonymity and confidentiality. During the online survey, all respondents were able to quit at any time and had the opportunity to send an e-mail to the researcher in case anything was unclear. Moreover, the Newcastle University Business School has granted ethical approval for this study.
3.7 Conclusion
This chapter described and justified in detail the used sample and the employed methods and tools of both data collection and data analysis with reference to the research methods
4. Findings
The previous chapter described the used methodology to collect and analyse the collected data. In this chapter, the hypotheses are tested, and the findings are presented. First, demographic information about the respondents is analysed with descriptive statistical analyses to highlight the profile of the respondents. Then, descriptive statistical analyses are used to reduce the large amount of data in a sensible way and provide information about each variable. Furthermore, preliminary analyses were conducted to check the normal distribution, reliability, and validity of the dataset. Finally, each hypothesis is tested with the appropriate statistical analysis.
4.1 Profile of respondents
After two weeks, 318 respondents had participated in the online questionnaire. As mentioned before, 78 responses were deleted as a result of missing data. So, the remaining and usable dataset consisted of 240 valid responses. This dataset consists of 91 (37,9%) men and 149 (62,1%) women. So, men are slightly underrepresented in the sample. With respect to age, 116 (48,3%) were aged between 16-24 years; 24 (10%) were aged between 25-34 years; 18 (7,5%) were aged between 35-44 years; 50 (20,8%) were aged between 45-54 years and 32 (13,3%) were aged above 55 years. In terms of the highest academic qualifications attained by the respondents, 61 (25,4%) were holders of Master degrees, 93 (38,8%) had a degree from the University of Applied Sciences, 61 (25,4%) did Vocational Education and only 25 (10,4%) did High School at the moment of filling in the survey. The median of Education level was 3,00 (University of Applied Science). Strikingly, nearly all respondents (91,3%) had a relationship of more than five years with their bank. Furthermore, Rabobank (45%), ABN Amro (25,4%) and ING (22,1%) were the three most represented Dutch banks. Overall, the respondents were also quite satisfied with the current service portfolio offer by their bank, as the means for this control variable ranged from 3.85 to 4.27.
4.2 Descriptive results
The mean, standard deviation, Skewness, and Kurtosis of each specific item are presented in Table 1 of Appendix B. Remarkably, the means of all dimensions of the Relationship Quality construct are quite high ranging from 3.21 to 4.10 and the mean results of Customers’
intention to share data with FinTech companies are quite low ranging from 2.20 to 2.43. The standard deviations for all items are between 0.84 and 1.36.
The Skewness value and Kurtosis value were analysed to test if the data was
is an indicator for the asymmetry of the distribution of the data and is zero for a normal distribution, whereas Kurtosis is defined as “a measure of peakedness or heavy tails, or some kind of combination of the two” (Critchley and Jones, 2008, p. 416). Data can be interpreted as normal when both the Skewness and Kurtosis values of the measured items are between -1.96 and -1.96 (Blumberg, Cooper and Schindler, 2011). As shown in table 1 of Appendix B, all Skewness and Kurtosis values are exactly between this indicated range. Therefore, the data can be interpreted as normally distributed, which is also an assumption that should be covered for some analyses, e.g. a regression analysis, to be able to test the hypotheses in the right way.
4.3 Reliability analyses
All variables of the conceptual model were measured with multiple questions. Before a reliability analysis was done, the questions measuring each variable were checked to see if they correlate with each other. A new average variable was determined, that depicted all items measuring one variable up in one score when the items indeed correlate with each other and together had a high enough Cronbach’s alpha value. This Cronbach’s alpha value should be as high as possible (ranging from 0 to 1), but a value less than 0.5 is unacceptable (George and Mallery, 2003). Furthermore, a factor analysis was conducted to test how much every item of the questionnaire did explain the associated construct. The factor loadings are presented in Tables 1, 2 and 3 of appendix C. These factor loadings should be 0.70 of higher, according to the rule-of-thumb for reliability (Kolenikov, 2009). Fortunately, almost all factor loadings were around 0.70 or higher, which is a sign of good reliability of the questionnaire. Finally, the correlation results of all variables are incorporated in a correlation matrix, as shown in Table 1.
4.3.1 Customer’s intention to share data with FinTech companies
4.3.2 Technology Readiness
As previously mentioned, in order to reduce a significant number of survey questions, only the three items with the highest factor loadings on each of the Technology Readiness dimensions were applied instead of the whole 36-item scale developed by Parasuraman (2000). The correlation between these three items was tested to check if the model remained reliable.
Firstly, a correlation analysis was done to check whether the three items measuring Optimism of the Technology Readiness construct correlate with each other, as shown in Table 1 of Appendix C. The correlation analysis showed that the first two items did correlate
significantly (r= 0,697, p < .01). Furthermore, the first item did also correlate significantly with the third item (r=0,596, p < .001) and the second item did also correlate significantly with the third item (r= 0,646, p < .001). The reliability analysis on the three questions measuring Optimism revealed that the three items together had a α = 0,844. Therefore, an average variable was computed of these three items.
Secondly, a correlation analysis was done to check whether the questions measuring Innovativeness correlate with each other. The analysis showed that that the first item did correlate significantly with the second and third item (r= 0,706, p < .01 and r= 0,541, p < .001) and the second item did also correlate significantly with the third item (r= 0,567, p < .01). The reliability analysis on the three questions measuring Innovativeness showed that the three items together had a α = 0,821. This value is close to 1. Hence, an average variable was computed of these three items.
Thirdly, a correlation analysis was done to check whether the questions measuring Discomfort correlated with each other. This analysis showed that item 1 and item 2 did correlate significantly (r= 0,312, p < .001), item 1 and item 3 correlated significantly (r= 0,361, p < .001) and item 2 and item 3 correlated significantly (r= 0,268, p < .01). The reliability analysis on the three questions measuring Discomfort showed that the three items together had a α = 0,576. This coefficient is not as high as the previous coefficients, but high enough to compute these three items into one average variable.
Fourthly, a correlation analysis was done to check whether the questions measuring Insecurity correlated with each other. The correlation analysis indicated that item 1 and item 2 did correlate significantly (r= 0,507, p < .001), item 1 and item 3 correlated significantly (r= 0,365, p < .001) and item 2 and item 3 correlated significantly (r= 0,450, p < .01). A
reliability analysis showed that the three items together had a α = 0,702. Therefore, an average variable was computed of these three items.
customers’ Technology Readiness correlate with each other. This analysis, as illustrated in Table 1 of section 4.5, showed that item Optimism and Innovativeness did correlate
significantly (r= 0,671, p < .001), Optimism and Insecurity correlated significantly (r= -0,233, p < .001) and Insecurity and Discomfort correlated significantly (r= 0,474, p < .001). So, people who are optimistic towards new technologies are more likely to be also more innovative and have fewer feelings of insecurity, and vice versa. Furthermore, people who have more feelings of insecurity are also more likely to feel more uncomfortable with new technologies, and vice versa. The reliability analysis showed that the four variables together had a α = 0,379. This coefficient is below 0.5, and therefore, these variables were not computed into one variable.
4.3.3 Customer’s propensity to switch to FinTech companies
As shown in Table 3 of Appendix C, the correlation analysis showed that both items measuring Customers’ propensity to switch to FinTech companies correlated significantly with each other (r= 0,624, p < .001). So, when customers would switch to a FinTech company when this company offers them a better financial service than their bank, they are also more likely to switch to a FinTech company when this company offers them better economic conditions and vice versa. The reliability analysis showed that the two items together had a α = 0,768. So, an average variable was computed of these two elements.
4.3.4 Relationship quality
First, a correlation analysis was done to check whether the two items measuring Satisfaction of the Relationship quality construct correlated with each other, as shown in Table 2 of Appendix C. The correlation analysis showed that both items did correlate significantly (r= 0,640, p < .001). Moreover, the reliability analysis resulted in a α = 0,773. Therefore, a new average variable was computed of these two items.
Secondly, a correlation analysis was done to check whether the four items measuring Trust correlate with each other. This analysis showed that item 1 and 2 did correlate
reduced to 0,894. As this difference is so small, an average variable was computed of these four elements.
Thirdly, a correlation analysis was executed on the four items measuring
Commitment. This analysis showed that item 1 and 2 did correlate significantly (r= 0,509, p < .001), item 1 and 3 did correlate significantly (r= 0,560, p < .001), and item 1 and 4 did correlate significantly (r= 0,648, p < .01). Also item 2 correlate significantly with item 3 (r= 0,454, p < .001) and item 4 (r= 0,524, p < .001). Furthermore, item 3 and 4 correlate
significantly (r= 0,627, p < .001). A reliability analysis resulted in a α = 0,828. Therefore, a new average variable was computed of these four items.
In addition to these reliability and correlation analyses, a confirmatory factor analysis was conducted to test the proposed model’s goodness of fit when relationship quality is measured by Satisfaction, Trust, and Commitment. Therefore, the fit statistics of this proposed model were compared to other models in order to identify the model with the best fit. The structural model’s goodness of fit was evaluated using the Normed Fit Index (NFI), the Goodness of Fit Index (GFI), the Comparative Fit Index (CFI), The Root mean square residual (RMR) and the Relative chi-square. A model is regarded as acceptable if NFI is higher than .90, GFI is higher than .90, CFI is exceeding .93, RMR is less than .05 and the Relative chi-square is less than 5 (Byrne, 1994; Steiger, 1990; Schumacker and Lomax, 2004). The statistics for the proposed model showed strong goodness of fit, with NFI = .929 > .90, GFI = .923 > .90, CFI = .949 > .93, RMR = .047 < .05, and the Relative chi-square = 3.20 < 5. This model, as shown in figure 3, compares favourably to other tested models in which all underlying items were loaded on one single factor or covariance between two of the three factors was constrained to 1.
In the first tested model, all underlying items were loaded on one single factor. The evaluation of the model’s goodness of fit showed that NFI = .791 < .90, GFI = .769 < .90, CFI = .809 < .93, RMR = .091 > .05, and the Relative chi-square = 8.57 > 5. In the second model, covariance between Satisfaction and Commitment was constrained to 1, with the following fit statistics: NFI = .907 > .90, GFI = .900 > .90, CFI = .928 < .93, RMR = .314 > .05, and the Relative chi-square = 4.04 < 5. In the third model, covariance between Satisfaction and Trust was constrained to 1, with the following fit statistics: NFI = .896 < .90, GFI = .895 < .90, CFI = .917 < .93, RMR = .435 > .05, and the Relative chi-square = 4.51 < 5. In the last model, covariance between Commitment en Trust was constrained to 1, with the following fit
statistics: NFI = .907 > .90, GFI = .901 > .90, CFI = .927 < .93, RMR = .351 > .05, and the Relative chi-square = 4.05 < 5. As this shows, the proposed three-factor model is the only one with significant results on all required goodness-of-fit statistics.
4.4 Validity analyses
Validity refers to the degree to which one or more variables accurately represent what it intends to measure (Hair et al., 2006). In order to establish construct validity, both the convergent and discriminant validity were tested (Campbell and Fiske, 1959). The former shows to what extent the different items are measuring one variable correlate with each other, while the latter shows the extent to which a variable is truly distinct from another variable (Malhotra and Dash, 2010). The convergent validity was measured with the Average Variance Extracted (AVE), which should be higher than 0.5 to establish convergent validity (Fornell and Larcker, 1981). The AVE values of all variables were based on the measured factor loadings, as presented in Tables 1,2 and 3 of Appendix C, and were approximately 0.5 or higher. So, convergent validity was established, as shown in Table 1 on the next page.
Table 1. Correlation Matrix, AVE, √AVE (Diagonal bold entries)
**. Correlation is significant at the 0.01 level (2-tailed) *. Correlation is significant at the 0.05 level (2-tailed)
4.5 Hypotheses testing
In order to test the influence of the single independent variables Optimism, Innovativeness, Discomfort, and Insecurity on Customer’s intention to share data with a FinTech company, this study performed multiple regression analyses. Then, a linear regression analysis was conducted to test the influence of the independent variable Customers’ intention to share data with a FinTech company on the dependent variable Customers’ propensity to switch to FinTech companies. Thereafter, Structural Equation Modeling (SEM) was applied in order to indicate how well Optimism, Innovativeness, Discomfort and Insecurity together predict a Customer’s intention to share data with FinTech companies, to visualize the interrelationships among these variables and the variables Customers’ intention to share data and Customers’ propensity to switch, and to analyse the goodness of fit of the model. Finally, three multiple regression analyses were conducted to test the moderating effects of Satisfaction, Trust, and Commitment. All these variables were measured on interval level, so (multiple) regression analyses were suitable. Furthermore, all relevant control variables were applied to every analysis.
4.5.1 Customers’ intention to share data with FinTech companies
The first model, including Customers’ intention to share data with FinTech companies as the dependent variable, was initially tested with the control variables only, as shown in Table 2. However, the influence of age was not significant, B = -0.063, p = .207, the influence of Gender was not significant, B = -0.189, p = .174 and the influence of Education level was not significant, B = 0,053, p = .116. All control variables together explained 4,8% of the variance in Customers’ intention to share data with FinTech companies, as indicated by the R-squared value of .048.
Firstly, in order to examine whether or not customers’ level of optimism towards technology positively influences their intention to share data with FinTech companies, a multiple regression analysis was performed with Optimism regressed on Customer’s intention to share data with a FinTech company, as shown in model 2 of Table 2. A Durbin-Watson test showed a value of 1.945, which is between the critical values of 1.5 and 2.5, so there is an independence of observations (Durbin and Watson, 1971). The regression analysis was significant, R²=.133, F(4,236) = 23.033, p < .01. So, Customers’ Optimism does influence their intention to share data with FinTech companies, B = .376, p < .01. The R-squared value indicated that 8,5% of the total variation in the dependent variable was explained by
Secondly, to test whether or not customers’ level of innovativeness positively influences their intention to share data with FinTech companies, a multiple regression analysis was performed with Innovativeness regressed on Customers’ intention to share data with a FinTech company. The Durbin-Watson test showed a value of 2.044, which is between the critical values of 1.5 and 2.5, so the observations were independent. As shown in model 3 of Table 2, the regression analysis was significant, R² =.080, F(4,236) = 8.184, p < .01. So, Customers’ Innovativeness does positively influence their intention to share data with FinTech companies, B = .194, p < .01. Furthermore, only 3,2% of the total variation in Customers’ intention to share data was explained by Innovativeness. Innovativeness had a positive and significant effect on Customers’ intention to share data with FinTech companies, so H1b is confirmed.
Thirdly, a multiple regression analysis was also conducted to examine whether or not customers’ perceptions of Discomfort negatively influences their intention to share data with FinTech companies. A linear regression analysis was done with Discomfort regressed on Customers’ intention to share data with a FinTech company. The Durbin-Watson value was 1.871, which is also between the critical values of 1.5 and 2.5, so the observations were independent. As shown model 4 of Table 2, the regression analysis was significant, R² =.105, F(4,236) = 14.829, p < .01. So, Customers’ perceptions of Discomfort did negatively
influence their intention to share data with FinTech companies, B = -.286, p < .01. Moreover, 5,7% of the total variation in the dependent variable was explained by Discomfort. Hence, H1c is accepted as Discomfort had a negative and significant effect on Customer’s intention to share data with FinTech companies.
Table 2. Customers’ intention to share data with FinTech companies
**. Significant at the 0.01 level (2-tailed) *. Significant at the 0.05 level (2-tailed)
Model 1 Model 2 Model 3 Model 4 Model 5
B Std. Error B Std. Error B Std. Error B Std. Error B Std. Error
4.5.2 Customers’ propensity to switch to FinTech companies
The first model, including Customers’ propensity to switch to FinTech companies as the dependent variable, was initially tested with the control variables only, as shown in Table 3. These results indicated a significant influence of Gender on Customers’ propensity to switch, B = -.434, p = .002. Furthermore, the results showed a negative and significant effect of Service portfolio on Customers’ propensity to switch, B = -.231, p = .022, which implies that customers who are more content with the currently offered services by their bank are less likely to switch to the services offered by FinTech companies. Nevertheless, the influence of Age was not significant, B = -.017, p = .731, the influence of Education level was not
significant, B = .018, p = .830 and the influence of Duration of relationship was not significant, B = -.051, p = .645. All control variables together explained only 6,5% of the variance in Customers’ propensity to switch FinTech companies, as indicated by the R-squared value of .065
Thereafter, a multiple regression analysis was performed to examine the influence of the independent variable Customer’s intention to share data with a FinTech company on the dependent variable Customer’s propensity to switch to FinTech companies, as shown in model 2 of Table 4. The Durbin-Watson value was 1.934, which is between the critical values of 1.5 and 2.5, so the observations were independent. The regression analysis was significant, R² = .254, F(6,234) = 59.102, p < .01. So, a Customer’s intention to share data with FinTech companies will positively influence a Customer’s propensity to switch to FinTech companies, B = 0,451, p < .01. The R-square value indicated that 18,9% of the total variation in the dependent variable was explained by a Customer’s intention to share data. Moreover, the negative influences of customers’ Gender, B = -,345 p = .006, and customers’ perceived quality of their bank’s Service portfolio remained significant, B = -.230, p = .011. So, H2 is confirmed as Customers’ intention to share data with FinTech companies has a positive and significant effect on Customers’ propensity to switch to these companies.
Finally, a post hoc analysis was conducted to gain more insights about the effect of Gender on Customers’ propensity to switch. The independent variable Gender was measured at nominal level, and the dependent variable Customers’ propensity to switch was measured at interval level. Moreover, the variable Gender consisted of two categories: male and female (k=2). Therefore, an independent samples t-test was conducted in order to analyze whether men or women significantly differ in their propensity to switch to FinTech companies. However, first a Leven’s Test was conducted to test the equality of variances. This test
t(238) = 3.218, p = .001. So, the average propensity to switch of men (M= 3.21, SD=1.03) does significantly differ from the average propensity to switch of women (M = 2.77, SD = 1.04) with a mean difference value of .44, as shown in Appendix D.
4.5.3 Testing the model
Finally, a powerful and visual multivariate statistical approach called Structural Equation Modeling (SEM) was applied to analyse the structural interrelationships among the independent and dependent variables and specify the measurement error, as shown in Appendix E (Hair et al., 2006). SEM is a unique combination of both dependence and
interdependence techniques and is particularly useful if a dependent variable also becomes an independent variable in a subsequent relationship, which is the case with Customers’ intention to share data (Hair et al., 2006). Path analysis was used to obtain both information about the path values between the variables and information about the overall fit of the model. As shown in figure 4, only all previously tested variables were incorporated into the model to ensure the right level of overall fit of the model. The regression analysis on the influence of Innovativeness on Customers’ intention to share data conducted for H1b was significant. Strikingly, in this post hoc analysis this influence was not significant anymore, B = 0,079, p = .315. As previously mentioned in section 4.3.2 and shown in Table 1, there was a high multicollinearity between the variables Optimism and Innovativeness. A high
multicollinearity could increase the standard errors of the independent variables’ coefficients, which in turn could make some independent variables statistically insignificant whereas they are, in fact, significant. Therefore, based on this information, it is assumed that the influence of Innovativeness on Customers’ intention to share data is positive and significant, as
previously proved by H1b. The influence of Optimism (B = 0,271, p = .004), Discomfort (B = -0,171, p = .026) and Insecurity (B = -0,223, p < .001) on Customers’ intention to share data was still significant. Furthermore, only the covariances, as shown in the left side of figure 4, between Discomfort and Insecurity, Optimism and Insecurity, and Optimism and
Innovativeness were tested significantly. The R-squared value indicated that Optimism, Innovativeness, Discomfort and Insecurity together explained 21,8% of the variance in Customers’ intention to share data. Also, the influence of Customers’ intention to share data on Customers’ propensity to switch was tested significantly, B = 0,46, p < .001.
exceeding .93, RMR is less than .05 and the Relative chi-square is less than 5 (Byrne, 1994; Steiger, 1990; Schumacker and Lomax, 2004). The evaluation of the model’s goodness of fit showed that NFI = .979 > .90, GFI = .990 > .90, CFI = .990 > .93, RMR = .036 < .05, and the Relative chi-square = 1.80 < 5, that is non-significant at the p < .05 level. So, these results indicated a very good overall fit of the model, as all required criteria were met.
Figure 4. SEM
4.5.4 The moderating role of Satisfaction, Trust, and Commitment
Three moderator analyses using multiple regression were conducted to determine whether the relationship between a Customers’ intention to share data with FinTech companies and a Customers’ propensity to switch to FinTech companies was negatively moderated by
Satisfaction, Trust, and Commitment. These analyses were suitable because all relevant three were measured at interval level. In addition, an interaction term was created between
Customers’ intention to share data and each of the moderator variables to be able to perform a proper analysis. Furthermore, no strong correlation was found between Customers’ intention to share data and each of moderator variables. Moreover, the moderator variables were significantly related to Customers’ propensity to switch, as demonstrated in Table 1 of the validity analyses section.
all used independent variables were not strongly correlated with each other, as the VIF value < 10, so the data showed no multicollinearity. As shown in model 3, the effect of the
moderator was significant, R-squared change = .016, F(7,233) = 5.170, p < .05. So,
Satisfaction does negatively moderate the relationship between Customers intention to share data with FinTech companies and Customers propensity to switch to these companies, B = -.068, p = .024. Thus, H3a is accepted.
Second, the moderating influence of Trust on the relationship between Customers’ intention to share data and Customers’ propensity to switch was tested, as shown in model 4 of Table 3. The Durbin-Watson value was 1.965, which is between the critical values of 1.5 and 2.5, so again independence of observations was confirmed. Moreover, the VIF value of all variables was below 10, so all used independent variables were not strongly correlated with each other. As shown in model 4, the moderating effect was significant, R-squared change = .031, F(7,233) = 9.950, p < .01. So, Trust does negatively moderate the relationship between Customers intention to share data with FinTech companies and Customers
propensity to switch to these companies, B = -.101, p = .002. Therefore, H3b is confirmed. Finally, the moderator Commitment was applied to test the model, as illustrated in model 5 of Table 3. The Durbin-Watson value of 1.980 indicated independence of
observations. Furthermore, the VIF value of all variables was lower than 10. Therefore, no high multicollinearity between the independent variables was detected. As shown in model 5 the moderating effect of Commitment was significant, R-squared change = .065, F(7,233) = 22.203, p < .01. So, Commitment does negatively moderate the relationship between
Table 3. Customers’ propensity to switch to FinTech companies
**. Significant at the 0.01 level (2-tailed) *. Significant at the 0.05 level (2-tailed)
Model 1 Model 2 Model 3 Model 4 Model 5
B Std.
