Banks are becoming increasingly “mobile”, with access available at anywhere and at any time, as a result of the advancement of new information and communications technologies (ICTs) (SHAIKH; KARJALUOTO, 2015). One of the services born from these technologies is mobile banking, or m-banking, which is considered the third technological revolution in customer service (GOUVEIA, 2007) and has represented an important channel of relationship with customers (FEBRABAM, 2013).

Conceptually, m-banking can be described as a set of banking services offered through the use of a mobile device, connected to a mobile internet network or data plan, which enables users to make payments and perform virtually any banking transactions linked to the current account without the participation of bank employees (GARTNER, 2009; HANAFIZADEH; KHEDMATGOZAR, 2012).

The m-banking service includes full access to bank transactions and details of the current account, credit cards, payments, transfers of securities, applications in an investment account etc. The use of the mobile app does not generate expenses for the consumer, and the only cost to use m-banking is the transmission of data to the telephone operator if it is not connected to a Wi-Fi network (AL-JABRI; SOHAIL, 2012).

Several models have been applied to evaluate the acceptance of technology, with the ones mostly used being the Theory of Reasoned Action (TRA) by Fishbein and Ajzen (1975), which served as a basis for other models such as The Theory of Planned Behavior (TPB) (AJZEN, 1991), the Technology Acceptance Model (TAM) by Davis, Bagozzi and Warshaw (1989), the Model of the Information Technology Diffusion Process (STRAUB, 1994), The Theory of Decomposed Planned Behavior (Taylor and Todd, 1995) the Unified Theory of Acceptance and Use of Technology (UTAUT) (VENKATESH et al., 2003) and the Unified Theory of Acceptance and Use of Technology 2 (UTAUT 2) (VENKATESH; THONG; XU, 2012). In addition, the Diffusion of Innovations Theory (ROGERS, 2003), although it is not derived from TRA, has also been widely used to assess the acceptance of technologies in light of the attributes it utilizes to characterize innovations (relative advantage, compatibility, observability, complexity and testability).

Of the models above, one of the most used has been the TAM (DAVIS; BAGOZZI; WARSHAW, 1989), due to its parsimony and explanatory power (KULVIWAT et al., 2007). According to Davis et al. (1989), the acceptance of the technology, or the intention to use it, is influenced by the behavioral intention, which in turn, is formed by two cognitive evaluations: (1) the perceived usefulness; and (2) the ease of use. The first argues that the use of a certain computational technology (the use of some kind of software or information system) will improve the person’s performance in their job. The second proposes that the use of the evaluated technology will be free of effort.

The model infers that the perceived ease of use directly affects perceived usefulness, and both influence the attitude that an individual will have on the intention to use a particular system. The perceived usefulness and attitude regarding the use affect the behavioral intention of use, that is, the intention to use the system in the future, which will determine the actual use of the system. To increase the explanatory power of the TAM, Davis, Bagozzi and Warshaw (1989) proposed the possibility of including external factors prior to people’s cognitive evaluations, attitudes and intentions.

With regard to the application of these models for m-banking research, the literature shows a growing interest in investigating the theme (LUARN; LIN, 2005; RIIVARI, 2005; SUORANTA; MATTILA, 2004, LEE; KANG; MCKNIGHT, 2007; SULAIMAN; JAAFAR; MOHEZAR, 2007; GU; LEE; SUH, 2009; LIN, 2011; RAMOS; PIMENTA; RODRIGUES, 2010; SANTOS; VEIGA; MOURA, 2010; HANAFIZADEH; KHEDMATGOZAR, 2012; HANAFIZADETH, et al., 2014). Sulaiman, Jaafar and Mohezar (2007) suggest that Rogers’s (2003) Diffusion of Innovations Theory is one of the most suitable in the study of the adoption of m-banking. In contrast, Luarn and Lin (2005) and Pedersen (2005) believe that the TAM is the model that best applies to the context of m-banking. In general, all of these studies seek to identify which factors have the greatest effect on the intention to use m-banking. Thus, listed below are the most influential factors found in the literature that affect the consumer’s decision to accept new technologies together with the hypotheses of the proposed model to evaluate the intention to use m-banking in the Brazilian context.

In the virtual world, anonymity and temporal and spatial separation in online transactions involve great uncertainties and risks. Therefore, users need to build trust to mitigate perceived risk and thus facilitate their transactions (ZHOU, 2011a). According to Mayer, Davis and Shoorman (1995), trust can be understood as the desire of one of the parties of being vulnerable to the actions of another party based on certain expectations of how this second party, to which the trust is directed, will act in a given situation. According to the authors, the previously mentioned vulnerability is related to a pre-disposition to take risks, considering that something important may be lost. Rotter (1967) defines trust as an individual expectation that another individual’s word is trustworthy. Thus, trust would be the willingness of one party to accept vulnerability, but with the expectation that one can rely on the other party (MORGAN; HUNT, 1994).

In the context of e-commerce, trust has been widely investigated and received considerable attention from the literature (BELDAD; DE JONG; STEEHOUSER, 2010). Similar to online transactions, mobile transactions also involve large risks. Thus, the construction of mobile user trust is critical (VARNALI; TOKER, 2010) so that the individual is prone to use the service. However, compared to the abundant research on online trust, mobile trust is only beginning to receive attention (ZHOU, 2011a).

Siau and Shen (2003) observed that mobile trust is affected by factors associated with two aspects: supplier and mobile phone technologies. According to Li and Yeh (2010), application design affects trust in mobile technology through ease of use, perceived usefulness and customization. Vance, Christophe and Straub (2008) examined the effect of system quality, including visual appeal and navigation structure, as factors that affect user trust in the mobile environment.

Considering the impersonality of the online environment and the large number of threats present in this scenario (errors in transactions, viruses, phishing, hackers, etc.), we have a greater perception of risk, which diminishes the intention to use, making trust play a fundamental role to mitigate fears of this nature (WANG et al 2015; MALAQUIAS; HWANG, 2016). Therefore, trust is believed to be an important antecedent of use intention of digital financial systems, such as m-banking (SHAIKH; KARJALUOTO, 2015).

Articulating the trust construct further, the study performed by Dimitriadis and Kyrezis (2010) addresses the concept of trusting beliefs, which are described by means of four distinctive elements: (1) competence - the belief that the other party has the ability or power to do what needs to be done; (2) benevolence - the belief that the other party cares and is motivated with the other; (3) integrity - the belief that the other party makes good faith agreements, tells the truth, acts ethically, and keeps promises; and (4) predictability - the belief that the actions of the other party are consistent over time and can be predicted in certain situations. For authors, the “trusting beliefs” would be the cognitive and affective aspects of the trust construct and “trusting intention” would be the behavioral component of trust. Mayer, Davis and Shoorman (1995) map out in their study three of the factors cited above as those that explain most of the trust or, rather, of the trusted party, and they are: ability (which can be understood as competence/skill), benevolence and integrity.

Transposing these concepts into the context of m-banking, competence refers to the individual’s belief and the perception of the individual that managers of m-banking companies have the competence, ability and knowledge to understand their needs while managing their personal finances. Benevolence is the belief that one has the perception that bank executives care about them and act on their behalf. Integrity is the belief of the subject’s perception that the company follows a set of principles usually cultivated by responsible people (MCKNIGHT; CHOUDHURY; KACMAR, 2002). The perception of competence, benevolence, and integrity in m-banking can play an important role in determining individual attitudes to use the application. Furthermore, several other factors also influence the construction of trust. Familiarity, perceived usefulness, perceived ease of use, privacy, security and innovation are factors that have been pointed out as important antecedents to trust.

Familiarity refers to a person’s knowledge of a given technology and the experience they have with it - in the case of the present study, m-banking. According to Gefen (2000), familiarity is one of the factors that directly or indirectly influences the use intention of the Internet for both information and transaction purposes. It has been studied as a control variable in the relationship between trust and use intention (GEFEN; STRAUB, 2004). The impact of familiarity on the willingness to perform transactions through mobile channels has also been supported in a context of retail Internet banking (BHATTACHERJEE, 2002).

In addition, according to Luhmann (1979), familiarity is a requirement for trust as it assists the individual in understanding the environment within which expectations of trust can be explained. Gulati (1995) argues that familiarity helps build trust since it not only provides a framework for future expectations, but also enables people to come up with concrete ideas about what to expect based on previous interactions. The reason for this is that familiarity measures the degree to which previous experience was understood.

Afshan and Sharif (2016) also analyzed the direct and positive relationship that familiarity has on trust in their study, showing the direct link between these constructs.

The study by Gefen (2000), reports that in many cases, prior experience is the basis of trust, with familiarity possibly creating trust when the experience was favorable or ruining it if the experience were unfavorable. Thus:

H1: Familiarity has a direct and positive effect on trust in m-banking.

The construct perceived usefulness expresses how much an individual considers that a technology can improve their productivity or performance in a given task (DAVIS; 1989). Such a definition refers to the consumer’s perception of the use of banking applications via mobile devices in relation to the improvement of the result, making the experience of performing a financial transaction more efficient. This perception positively influences the intention to trust the use of m-banking, with the trusting intention being the behavioral aspect of the “Trust” construct (DIMITRIADIS; KYREZIS, 2010).

According to Afshan and Sharif (2016), individuals will adopt a particular technology if they perceive it as useful in performing a daily task, and will not adopt it in case it does not result any improvement in these tasks. Therefore, we constitute the following hypothesis:

H2: Perceived usefulness has a direct and positive effect on trust in m-banking.

The perception of ease of use is one of the main components of the TAM. It is defined as the belief that the individual forms regarding the absence of effort in learning to use a new technology. The greater the perception of m-banking’s ease of use, the more positive the consumer’s attitude will be towards the technology.

Because of restrictions on mobile devices, such as small screens, for example, it may be difficult for the user to operate m-banking services if service providers do not offer them a good interface. An easy-to-use m-banking app with a good interface and intuitive navigation would reflect the competence and benevolence of service providers, which would affect consumer trust. In addition, a simple mobile banking system would also reduce the time spent by the user to learn how to use the m-banking service and increase their sense of control over the application. This would help the consumer to focus on the main activity, which is to make financial transactions, improving their experience. The effect of perceived ease of use on trust was validated in previous studies (BENAMATI et al., 2010; VANCE; CHRISTOPHE; STRAUB, 2008). Thus:

H3: Perceived ease of use has a direct and positive effect on trust in m-banking.

Perceived security is the user’s perception that the service provider will meet safety requirements such as authentication, integrity, and encryption (KIM; FERRIM; RAO, 2008).

The customer’s perceived security when carrying out online transactions depends on how one understands the level of security measures taken by the seller (FRIEDMAN; KHAN JR, 2000). When a user finds security features and protection mechanisms (encryption, protection, SSL authentication, etc.) on the bank’s website and specifically in the m-banking application, one can recognize the seller’s intent to ensure the security requirements needed during online transactions (CHELLAPPA; PAVLOU, 2002). This would help the customer to make a purchase decision, since all of the above mentioned artifacts emphasize that the seller is making efforts to gain the consumer’s trust, reducing their risk perception. Therefore, the user’s security perception would increase trust towards the supplier, in this case, the banks.

On the other hand, many customers believe that online payment channels are not secure and can be intercepted (CRUZ et al., 2010; LA POLLA; MARTINELLI; SGANDURRA, 2013). This would reduce their trust level by discouraging them from carrying out online banking transactions. Therefore:

H4: Perceived security has a direct and positive effect on trust in m-banking.

Perceived privacy is the perception that the supplier will try to protect the consumer’s confidential information, collected during electronic transactions, and not to make unauthorized use or disclosure of such (KIM; FERRIM; RAO, 2008).

In online trading, during a transaction, the seller collects varied personal information such as name, e-mail, phone number and home address of the buyer (SWAMINATHAN; LEPKOWSKA-WHITE; RAO, 1999). In a bank transaction, in addition to the personal data of any transaction, information of bank statements, credit cards and investments is available, which, in the wrong hands, can jeopardize the user (MUKHERJEE, NATH, 2003). Thus:

H5: Perceived privacy has a direct and positive effect on trust in m-banking.

Innovativeness is the tendency for an individual to be a pioneer in adopting new technologies or being an opinion leader in relation to technology (AGARWAL; KARAHANNA, 2000; ROGERS, 2003).

According to Rogers (2003), individuals with greater innovativeness are able to cope with higher levels of uncertainty and develop more positive intentions regarding the acceptance of an innovation, such as a m-banking application. Therefore, users with high innovativeness are more likely to experience m-banking (MONTEZEMI; SAREMI, 2015). For being open-minded about new technologies and by having a more positive attitude to take risks, such individuals tend to quickly build trust in m-banking (ZHOU, 2011a). Therefore:

H6: Innovativeness has a direct and positive effect on trust in m-banking.

The use intention expresses the discreet probability of the consumer to use something specific in a period of time (DIMITRIADIS; KYREZIS, 2010). Based on models such as TAM (DAVIS; BAGOZZI; WARSHAW, 1989), TRA (FISHBEIN; AJZEN, 1975) and TPB (AJZEN, 1991), many e-commerce surveys have demonstrated that use intentions involving online transactions are a significant predictor of effective consumer participation in the operations (PAVLOU; FYGENSON, 2006). The relationship between intention and behavior is based on the assumption that human beings make rational decisions based on the information they have. Thus, the behavioral intention of a person to perform a behavior is the immediate determinant of the actual behavior of the person (FISHBEIN; AJZEN, 1980).

Pennington, Wilcox and Grover (2003) have suggested that consumers perceive the technical competence of a website in accordance with the understanding of the processes governing online transactions. If clients believe that banks possess the competence, skills and experience to provide adequate transactional services, such consumers are more likely to evaluate mobile banking services favorably, increasing trust in online transactions, which positively affects the intention to use m-banking. On the other hand, according to Afshan and Sharif (2016), trust plays a preponderant role in an e-commerce and m-commerce environment, due to the high degree of uncertainty and associated risks between the parties involved. Thus, trust becomes a factor related to the intention and use of m-banking. Therefore:

H7: Trust has a direct and positive effect on the intention to use m-banking.

Figure 1 illustrates the discussed hypotheses and the proposed theoretical model.

Figure 1 Source: Adapted to the Brazilian context based on DIMITRIADIS; KYREZIS (2010).

We performed a confirmatory factor analysis (CFA) to test the validity, unidimensionality and reliability of the scales used in the measurement model. After refinement of the model, in which 8 items were removed (1 of usefulness, 1 of privacy, 1 of security, 1 of familiarity, 2 of innovativeness and 2 of trust), the final measurement model was composed of 33 items and presented fit indices considered as good, with the RMSEA 0.051, CFI 0.92, IFI 0.92, and TLI 0.91. Assessed jointly, the final indices suggest a satisfactory fit of data for the proposed model (HU; BENTLER, 1999).

The face validity for all scales was guaranteed during the development of the research instrument (choice of scales already used in the literature, careful translation and pre-tests). To verify the nomological validity, we analyzed the correlation matrix between the constructs, with all correlations being significant and in the expected direction. In relation to the convergent validity, we calculated the average variance extracted for each construct (AVE). All AVE values were between 0.59 and 0.73, evidencing the convergent validity of the scales used. Regarding the internal consistency and reliability of the scales, all met the minimum levels of reliability considered adequate by the literature (FORNELL; LARCKER, 1981), with all values ranging from 0.72 to 0.90 for the alpha coefficient and from 0.71 to 0.90 for composite reliability. Finally, all the shared variances were lower than the variance extracted by the items that measure the constructs, indicating adequate discriminant validity.

Regarding the structural model, all indices indicated a good fit of the model to the data. The ratio χ2/d.f. was 2.63. Incremental goodness-of-fit indexes were greater than 0.90, with CFI 0.91, TLI 0.90 and IFI 0.91. The absolute fit indices presented values below the limit of 0.08 established by the literature (HU; BENTLER, 1999), also indicating a good fit of the model. The RMSEA was 0.064 (C. I. from 0.057 to 0.068) and the SRMR was 0.069. Given the presented indices, we concluded that the fit of the model was satisfactory.

After checking the fit of the measurement and structural models, we evaluated the coefficients estimated for the causal relationships between the constructs (Table 2). The verification of each of the hypothesis of the research was carried out with the analysis of the magnitude, direction and significance of the standardized coefficients estimated by means of the structural model. Except for hypothesis 7, all others were verified.

Table 2

Proposed relationship	Standardized Coefficient	p-value	Hypothesis Verified
H1: Familiarity → Trust	0.10	0.048	Yes
H2: Perceived usefulness → Trust	0.13	0.011	Yes
H3: Perceived  ease of use → Trust	0.15	0.005	Yes
H4: Perceived Security → Trust	0.60	< 0.001	Yes
H5: Perceived Privacy → Trust	0.43	< 0.001	Yes
H6: Innovativeness → Trust	-0.04	0.938	No
H7: Trust → Use Intention	0.73	< 0.001	Yes

Source: research data.

In the proposed model, thirteen items measured the trust construct, with the results indicating a good fit of the model and providing evidence of the relevance of the proposed relations. Regarding the seven hypotheses tested in this study, six were tested with a significance level of 0.05 or lower. The proposed model was able to explain 60.5% of the variability of the trust construct and 53.4% of the variability of the use intention, indicating that the antecedents evaluated are able to explain a good part of consumers’ trust and use intention with respect to m-banking services.

The first hypothesis, H1, pointed out that familiarity had positive and direct impact on trust. We can infer that consumers more familiar with banking applications tend to trust the channel. By analyzing the data, we observed that 31% of the respondents reported that they knew their bank’s m-banking service through the bank’s institutional campaigns; 15%, through the media; and 12% were presented to it by the manager of their account. We can assume that more than 50% of the respondents had access to information to get to know their bank’s m-banking solution and feel familiar with it, and this would have contributed to a positive impact on trust in m-banking.

In the case of H2, the study also indicates that there is a positive and direct relationship between usefulness and trust. We can infer that the consumer who perceives the usefulness of m-banking tends to have their trust positively affected. Considering the current lifestyle of consumers, characterized by a constant concern with time, which is increasingly scarce, it is understood that there is a greater perception about the usefulness of applications that optimize time, i.e., of technologies that can improve individual productivity and performance. This agrees with the study by Davis, Bagozzi and Warshaw (1989), which indicates that the perception of usefulness is a factor that can lead the individual to use a new technology.

As for H3, it is also confirmed, indicating that the ease of use positively affected trust and, indirectly, the use intention, corroborating the results of studies such as Zhou (2013) and Hanafizadeh et al. (2014). That is, the easier the handling of m-banking services, the greater the user’s propensity to rely on the app. Ease is a perception that human beings have that they will expend little effort to perform a particular task. By observing the data, we verified that respondents aged 31 to 40 correspond to 47% of the sample and participants younger than 30 years of age, 18%. We can infer that more than half of the participants grew in the age of computers, suggesting that these users find it easy to deal with new technologies such as m-banking. Another fact that may have influenced the results is the fact that 55% of respondents stated that they had a family income above ten thousand BRL. Usually, the greater the purchasing power, the greater the chances that these people will have a higher degree of schooling, which may also reduce the perception of some difficulty in using m-banking. In addition, the smaller the effort to learn how to handle the application, the greater will be the acceptance of it, thus justifying the confirmation of H3.

In the case of security, consumers who have the perception that the application gives them security have shown greater trust in m-banking, confirming H4, where there was a positive effect of security on trust. We can assume that in the user’s perception, if the application is secure and is on a trustworthy platform, the sense of security in using the application leads one to trust in m-banking. These results resemble those by Dimitriadis and Kyrezis (2010) and by Hanafizadeh et al. (2014), in which security is one of the antecedents that positively impact trust. The fact that m-banking applications are available on official platforms with which the user has already established a trust relationship, as in the case of Apple’s App Store customers, can assist in forming the perception of security regarding this channel of relationship with the bank. Another assumption that can be made is that the perception that the device is always with the person gives them a sense of security, since, in most cases, he/she is the only user.

According to Wu, Huang and Hsu (2013) and McKnight and Chervany (2002), privacy is one of the barriers to consumer acceptance of online transactions. However, in contrast to the research by Wu, Huang and Hsu (2013), in the present study, it had a positive effect on trust (H5). One possible explanation suggests that participants’ perception of privacy actually has a direct and significant effect on trust, illustrating that, users who feel secure in protecting their data and privacy tend to trust more in online applications and transactions.

In relation to H6, innovativeness did not present direct and significant effects on trust, indicating that trust in m-banking would not be related to the degree of consumer innovativeness. The fact that this hypothesis was rejected corroborates the results obtained in the study by Dimitriadis and Kyrezis (2010), in which the classification of innovativeness of respondents had no impact on trusting the new technology. A possible explanation for the results found here is that, for the most part, the participants in this study are young people, younger than 40 years of age, with high schooling levels and high purchasing power, i.e., with greater propensity for innovation (innovativeness of the research participants, in an aggregated way, presented a high mean of 4.00 on a scale of 1.00 to 5.00, with a very low standard deviation of 0.23). This may have generated a perception that m-banking would not be such a new technology and, therefore, the degree of innovativeness of respondents did not have a significant effect on trust.

Another possible analysis for this result is that the tendency to innovativeness that an individual presents, in the context of the adoption of mobile banking and phone banking transactions, is not related to trust. In this case, other factors such as security and privacy weigh more as antecedents, since the experience in the use of mobile banking, if negative, can have unpleasant consequences, reducing the intention to experiment with a certain technology only by the propensity to innovativeness, corroborating with the finding also by Zhou (2011b).

Finally, H7 pointed out that trust positively and directly impacts the use intention of m-banking. This corroborates studies that speak of the direct and positive relationship between trust and use intention by consumers (MCKNIGHT; CHERVANY, 2002; DIMITRIADIS; KYREZIS, 2010; KIM; SHIM; LEE, 2009). The confirmation of H7 aligns with the arguments by Marett et al. (2015), which postulates that the consumer builds their trust beliefs based on the institutions that they relate with. When such beliefs are positively reinforced, they directly influence their trusting intention, thus reinforcing trust and subsequently their intention and/or attitude of using the service. Based on the study by Marett et al. (2015), it can be said that, if the respondent has a positive belief in relation to their bank, they present an intention to trust in their m-banking and, therefore, this will positively affect their intention to use the application.