Role of Celebrity Endorser on Brand Credibility, Loyalty, and Commitment on Mobile Banking Users

ABSTRACT


Introduction
The invasion of e-wallets and digital banking has posed a significant threat to the banking industry.Banks use various methods and efforts to introduce their products to the general public, including advertising.Over the last 40 years, global advertising and promotion have evolved (Belch & Belch, 2021).Outside of the United States, advertising spending increased from $55 billion in 1980 to nearly $214 billion in 2002 and $394 billion in 2019.Spending on advertising and other forms of promotion will continue to rise as companies worldwide seek to market their products and services to customers, including consumers and businesses.According to Belch and Belch (2021), there has been a shift in advertising spending from traditional media such as television, radio, and print to digital formats such as online search, image and video advertisements, and social media advertisements.The shift to mobile marketing occurs as consumers spend more time on their mobile devices than on traditional media such as newspapers, magazines, and television (Dash, Kiefer, & Paul, 2021).
One strategy that companies can use to promote their products is by using celebrity endorsers.Much previous marketing research has discussed celebrity endorsers' effectiveness in promoting products.Research also shows that advertisements using celebrity endorsers are more effective than other forms of advertising (Malik & Qureshi, 2016), and the research results show that celebrity endorsements can positively influence consumer perceptions and purchase intentions (Gupta, Kishore, & Verma, 2015).The research conducted by (Ha & Lam, 2017) also explained that using celebrity endorsements can increase the positive effect on trustworthiness, persuasiveness, believability, and likeability.
Companies use celebrity traits and charisma to create an analogy with product specialization to position a brand in the minds of target consumers.As a result, celebrity endorsements have emerged as one of the most popular advertising strategies for promoting various products, services, or brands (Schimmelpfennig & Hunt, 2020).In Indonesia, several large banks, from state-owned to private, have tried celebrity endorser marketing for their mobile banking advertisements.Several celebrities, including Agnez Monica, Isyana Saraswati, Noah Band, Luna Maya, and senior actress Kristin Hakim, have participated in mobile banking advertisements.(Hussain, Melewar, Priporas, Foroudi, & Dennis, 2020) examined how celebrity endorser credibility affects trust, loyalty, and commitment to Indian banking services.General banking services were studied, with celebrity credibility moderated by gender and profession.The study divided celebrities into film stars and athletes, reflecting India's endorsement landscape.The researcher modified variables from "Celebrity Endorsement, Brand Credibility, and Brand Equity", which examined how endorser credibility boosts brand credibility.According to that study, endorsement is a communication tool that signals to the endorsed brand, giving the brand credibility.Thus, credible endorsers boost brand credibility.
Celebrity credibility and trust are linked.The combined effects could result in high brand loyalty and commitment among mobile banking users.The celebrity profession is an essential moderator in this study.Traditionally, the gender of the endorser has been considered necessary, but few empirical studies have addressed this variable."The paucity of research on endorser gender effects is somewhat surprising, given that persuasion research shows that men and women respond differently to male and female communicators".As a result, the gender of celebrity endorsers is the current study's second moderator.
Based on previous research, a limited number of studies have examined the effects of celebrity endorsers on financial services.Furthermore, no specific research focuses on celebrity endorsers' influence on mobile banking.Therefore, this study examines the role of celebrity endorsers in shaping brand credibility, commitment, and loyalty among mobile banking users in Indonesia.The study will also explore the moderating effects of gender and profession of celebrities on the relationship between celebrity credibility and brand credibility, as well as test whether brand credibility mediates the influence of celebrity credibility on loyalty and commitment.The subsequent sections of this article provide a detailed explanation of the development and conceptual framework of the hypothesis, followed by a description of the testing methodology.The article's conclusion discusses conclusions, managerial implications, limitations, and future research directions.

Subjects
The method used in this study is quantitative because data on the dependent variable (DV) and independent variable (IV) were obtained using measuring instruments, the results of which were numerical scores.Statistical analysis was used to draw conclusions and interpret the relationship between the dependent and independent variables.Quantitative research is conducted when the researcher wishes to measure variation in an event, situation, problem, or idea, when the information collected is primarily for quantitative variables, and when the analysis aims to determine the level of the variable.

Measurement
Previous research on trust, loyalty, and commitment on banking service providers in India used an online survey as the study's assessment method.Table 1 shows the measuring device used for each variable in this study.Each participant's responses on a seven-point Likert scale, where seven equals "strongly agree" and 1 equals "strongly disagree," are used to determine the content validity of each assessment question.

Structural Equation Model
In this research, the researcher will use the Partial Least Squares (PLS) method with formative constructs as one of the component-based Structural Equation Modeling (SEM) methods.This is done to predict and develop theories or concepts that have been previously researched, unlike covariance-based SEM, which aims to test and confirm existing theories.PLS-SEM is used to predict and identify the primary variable, Celebrity Credibility, using directly observable indicators.This method is chosen due to the relatively large sample size.However, the relationship between celebrity credibility variables has a weak theoretical framework, which will be used for the second-order confirmatory factor analysis.
This research utilizes Second Order Confirmatory Factor Analysis, which examines the relationships between the indicators and latent variables.The research model is based on hypotheses developed from the theoretical framework of previous studies.The Second Order analysis involves two levels of factor analysis that explain the correlations among factors at the first level (Ganaie & Bhat, 2020).This analysis confirms the initial hypotheses, specifically exploring how the dimensions contributing to celebrity credibility influence customer loyalty, commitment, and trust in mobile banking applications.
The measurement model is used to assess the model's validity and reliability.The Kaiser-Meyer Olkin Measure of Sampling Adequacy (KMO) will be used as a measurement indicator in validity testing to determine the applicability of components.It should be greater than 0.5.In addition, the Component Matrix, which comprises factors from all variables, will be used as an index.It should be greater than 0.5.Reliability testing will employ Cronbach's alpha value parameter, with each variable meeting the 0.6 criteria.A Cronbach's alpha value of less than 0.6 indicates poor reliability, 0.7 is acceptable, and greater than 0.8 indicates good and acceptable dependability.
A structural model is used to project the causal relationship between latent variables.Various measures, such as the coefficient of determination (R2) and predictive relevance (Q2), can be used to evaluate the structural model.

The Hypothesis Test
Because the one-tailed test employed in this study's hypothesis testing has a significance level of 5%, a T-statistic value greater than 0.65 is recommended (Hair et al., 2017).The P value represents the probability that the null hypothesis is true.The pvalue threshold for the relationship between constructs in this study is less than 0.05.

Respondent Characteristics
Individuals must be 18 to open a bank account and utilize mobile banking services.To guarantee that respondents meet the criteria, they are requested to fill out screening questions about their age, the mobile banking service they use, and whether they know the celebrity sponsor of the mobile banking service.A total of 240 responses were collected, but 39 did not match the criteria, resulting in only 201 being measured further.The following is the distribution of respondents: Table 2 shows 111 men (55%) and 90 women (45%) of the 201 respondents answered.Therefore, it can be concluded that the majority of respondents in this study were female.Most respondents are aged 31 to 40 and have a bachelor's degree.
Respondents utilized the BRI mobile banking service with the name BRImo the most, and Noah Band was the most prominent celebrity endorser.

Measurement Model
The measurement model is an element of the path model that includes indicators and their relationship to the construct.This model measures the factor loading of the items used to measure the discriminant validity, convergent validity, and composite reliability.

Discriminant Validity
With the Fornell-Lacker test, discriminant validity can be examined.Table 3 displays the results of the Fornell-Lacker test, which indicates that the model is errorfree.This is because the correlation value between a variable and the variable itself is more significant than between a variable and other variables.

Convergent Validity Test
The study's model was evaluated using partial least squares (PLS).Hair et al. ( 2017) assessed the significance of the pathways using bootstrap resampling with SmartPLS version 3 and 5000 resamples.Due to the non-normal distribution of our data and the unique requirements of our respondents, we have chosen to employ the PLS method.This procedure suits non-normally distributed data and tiny populations.
We began by evaluating the measurement model.The measurement burden on their respective items exceeds the 0.50 threshold Holland (1999) suggested.This result indicates that each item is significantly laden on its respective construct.The composite reliability (CRs) exceeded the recommended 0.70 threshold.Nevertheless, the average variance extracted (AVE) results disclose that several variables' values fall below the suggested cut-off value of 0.05.According to Fornell and Larcker (1981), the AVE is a conservative and stringent measurement.Therefore, the convergent validity of a concept is acceptable even if the AVE value is below the suggested value if it has a high CR score.In addition, the alpha scores indicate that all variables used in this study are reliable, as each variable's value exceeds the recommended score of 0.6 (Malhotra, 2020).The factor loading value provides information about the convergent validity test.Table 3 shows indicators with factor loading values 0.7, such as L3 (0.516) and C1 (0.531).Both indicators can still be used because they fall within the acceptable range of 0.4 to 0.7.

Composite Reliability, Cronbach's Alpha, and AVE
Cronbach's Alpha (CA) and Composite Reliability (CR) values for the seven variables are shown in Table 4.All variables have CA and CR values greater than 0.7, indicating that the seven variables are reliable.The variable is valid because the AVE value on seven variables is more significant than 0.7.

Structural Model
In order to test hypotheses, a structural model examines the connections between the various research variables.This measurement indicates a theoretical component or concept associated with a route model.This research uses structural models' measurement, which includes not only the measurement of latent variables but also the assessment of the path between variables.

Coefficient of Determination R-square (R2)
According to Shmueli and Koppius (2011), the coefficient of determination, or R2, measures the model's explanatory power since it computes the variance explained by each endogenous component.According to Hair et al.'s research from 2019, higher values of the R2 suggest a greater explanatory power, which ranges from 0 to 1. Q2 value is a statistic used to determine a model's predictive power.Blindfolding is a technique that removes individual data points from a data matrix, replaces the eliminated points with the mean, and produces model parameters.For all variables, the blindfolding strategy estimates the missing data points.Higher Q2 values indicate greater prediction accuracy due to tiny differences between the original and anticipated values.Q2 values about 0.02 -0.15, 0.15 -0.35, and above 0.35 are often considered path models into minor, medium, and large predictive relevance.The coefficient of cross-determination (Q2) measures the model's prediction quality.Table 7 shows the best Q2 relevance value of 43.9% in predicting Brand Credibility.

Results of Hypothesis Testing
The path coefficient value of the construct of celebrity credibility to brand credibility is 0.638, indicating a positive relationship, indicating that as celebrity credibility increases, so does brand credibility.The t-value of 13.125 is greater than the t-table of 1.96, and the p-value of 0.000 is less than 0.05, indicating that the hypothesis of celebrity credibility having a significant effect on brand credibility is accepted.
The path coefficient construct of brand credibility on loyalty has a value of 0.845, indicating a positive relationship; that is, as brand credibility increases, so does loyalty.
The t-value of 24.548 is greater than the t-table of 1.96, and the p-value of 0.000 is less than 0.05, indicating that the hypothesis that brand credibility has a significant effect on loyalty is accepted.
The value of the path coefficient construct of brand credibility towards commitment is 0.848, indicating a positive relationship, indicating that as brand credibility increases, so does commitment.The t-value of 24.441 is greater than the ttable of 1.96, and the p-value of 0.000 is less than 0.05, indicating that the hypothesis that brand credibility has a significant effect on commitment is accepted.
The brand credibility construct's path coefficient value, which mediates celebrity credibility on loyalty, is 0.539, indicating a positive relationship.The t-value of 10.746 is greater than the t-table of 1.96, and the P-value of 0.000 is less than 0.05, indicating that the hypothesis, in which brand credibility mediates the relationship between celebrity credibility and loyalty, is accepted.
The brand credibility construct's path coefficient value, which mediates celebrity credibility on commitment, is 0.541, indicating a positive relationship.The t-value of 10.771 is greater than the t-table of 1.96, and the p-value of 0.000 is less than 0.05, indicating that the hypothesis, where celebrity credibility mediates the relationship between celebrity credibility and commitment, is accepted.
The celebrity profession path coefficient value that moderates celebrity credibility on brand credibility is 0.125, indicating a positive relationship.The t-value of 1.269 is less than the t-table of 1.96, and the p-value of 0.205 is more significant than 0.05, indicating that the hypothesis is rejected.In general, celebrity status as an actor/actress or singer does not significantly strengthen the relationship between celebrity and brand credibility (Hani, Marwan, & Andre, 2018).
The celebrity gender path coefficient value that moderates celebrity credibility on brand credibility is -0.163, indicating a negative relationship.The t-value of 1.756 is less than the t-table of 1.96, and the p-value of 0.08 is more significant than 0.05, indicating that the hypothesis is rejected.In general, celebrity gender, whether female or male, does not affect the relationship between celebrity credibility and brand credibility.

Effect of Celebrity Credibility on Brand Credibility
According to the findings of this study, celebrity credibility has a significant impact on brand credibility.This is consistent with previous research by (Hussain et al., 2020), which demonstrated that having a positive association with one stimulus (such as a celebrity) can help increase other stimuli, such as a brand.These factors may be necessary in forming a more long-lasting brand perception.Consumers who lack expertise in a specific product/service may rely on celebrity endorsements if the brand is deemed credible.Celebrity credibility has successfully influenced customer attitudes toward service providers, including influencing trust, increasing brand reviews, and increasing brand recognition (Purohit & Arora, 2022).Endorsing a credible celebrity improves customer perception of quality and uniqueness, resulting in customer loyalty to the service provider.

The Effect of Brand Credibility on Loyalty
According to the findings of this study, brand credibility has a significant impact on loyalty.This is consistent with previous research, which found that credibility significantly impacts customer loyalty.Trust significantly influences loyalty, which contributes to consistency and competence in future partnership outcomes, ensuring the benefits generated by both parties in the relationship (Litiver & Wardaya, 2022).

The Effect of Brand Credibility on Commitment
According to the findings of this study, brand credibility has a significant impact on commitment.Customer commitment entails considering future costs and benefits.Customers who feel compelled to continue using the service are less likely to switch to competing services if the brand believes the value it receives is greater than that of other services.As a result, advertisements featuring credible celebrities can increase positive perceptions of service providers and help customers increase subjective utility.In this case, customers are more likely to maintain long-term relationships with products or services endorsed by celebrities with high credibility.

Mediation effect of Brand Credibility on the effect of Celebrity Credibility on Loyalty
According to the findings of this study, brand credibility moderates the effect of celebrity credibility on loyalty.This is consistent with (Hussain et al., 2020), which found that endorsements from credible celebrities add value to a brand.According to (Hussain et al., 2020), celebrities humanly give the brand identity.This process supports the meaning transfer model, which aids in developing consumer-brand relationships.As a result, positive relationship outcomes such as trust in the endorsed brand emerge, driving commitment to the service provider.

Mediation effect of Brand Credibility on the effect of Celebrity Credibility on Commitment
The results of this study indicate that brand credibility mediates the effect of celebrity credibility on commitment.The findings of this study are consistent with previous research in which celebrity endorsements that promote trust ensure that service providers are trustworthy, which is the first step toward developing long-term commitment between consumers and brands.

The Moderating Effect of Celebrity Profession on the Effect of Celebrity Credibility on Brand Credibility
The findings of this study indicate that the celebrity profession does not affect the influence of celebrity credibility on brand credibility.This is consistent with research conducted by (Knoll & Matthes, 2017), who investigated how education changes a person's perspective on gender, finding that the higher a person's education, the more likely that person is to see gender as more equal.Customers with higher education in Indonesia's big cities dominated the sampling sample for respondents in this study, so they do not have a gender perception.

The Moderating Effect of Celebrity Gender on the Effect of Celebrity Credibility on Brand Credibility
The findings of this study indicate that celebrity gender does not affect the influence of celebrity credibility on brand credibility.This is in line with research conducted by (de Oliveira Santini et al., 2020) on the influence of a person's profession on public views regarding the information provided, which found that people do not consider a person's profession to be important in determining the credibility of information, but rather how a person conveys information.

Conclusion
According to this study, celebrity credibility affects brand credibility.Credible celebrities boost quality and originality, increasing customer loyalty.Brand Credibility also boosts loyalty.Brand trust ensures stability and competency in future partnerships, benefiting both parties.Brand Credibility affects Commitment and Loyalty.Celebrity endorsements increase customer loyalty.Brand credibility mediates celebrity credibility and customer loyalty for mobile banking app users.Credible celebrities boost customer loyalty.Brand credibility buffers celebrity credibility's effect on mobile banking app users' customer loyalty.Through brand-related marketing methods, quality information can boost brand credibility.This study demonstrated no substantial moderating effect of the celebrity profession on brand credibility.Singers have a more significant impact on brand credibility than actors.Celebrity credibility and brand credibility are not moderated by celebrity gender.Male celebrities affect brand credibility more than female celebrities.This research shows that celebrity and brand credibility impact each other through brand credibility.Celebrity occupation and gender also affect Brand Credibility, but not much.
This research is essential since celebrity endorsers in financial services, especially banking, are rarely studied.This study illuminates celebrity endorsements in Indonesian mobile banking app ads.This research aims to help banking decision-makers choose a credible celebrity endorser for mobile banking ads, affecting brand credibility, loyalty, and customer commitment to the app.
This study has several limitations.Firstly, there is a bias due to the generalization of the study, as the respondents mainly consisted of individuals from Generation X and Y, who may have different perspectives on celebrities endorsing banking products.Secondly, the study did not explore celebrities' specific media channels to promote the endorsed products, even though different media channels have varying impacts and effectiveness.Thirdly, in Indonesia, celebrity endorsers are predominantly actors/actresses and singers, whereas, in reference journals from India, athletes are also recognized as a separate category with a significant impact on credibility.The following research is expected to reach a broader range of respondents from various backgrounds and generations to capture different perspectives on celebrity endorsers in mobile banking advertisements.Furthermore, future research can diversify data collection methods.In addition to questionnaires, research can be conducted using different data collection methods such as in-depth interviews, direct observations, or document analysis.These various approaches can provide richer and deeper insights.

Bibliography
Dash, Ganesh, Kiefer, Kip, & Paul, Justin. (2021).Marketing-to-Millennials: Marketing 4.0, customer satisfaction and purchase intention.Journal of Business Research, Figure 1 Various Celebrity Endorsers of Mobile Banking in Indonesia The claims of this bank can be trusted.•Over time, my experience with this mobile banking service provider makes me hope that the bank will continue to uphold its commitment, neither more nor less.• This mobile banking service provider has a trusted reputation.• This mobile banking service provider does not pretend to be something it is not.

Table 6 Coefficient of Determination R-square (R2)
According to the R2 value inTable 6, Celebrity Credibility has a 48.7% influence on Brand Credibility.Meanwhile, the Brand Credibility variable has a 71.9% influence on Commitment to the Service Provider.Brand Credibility has a 71.5% effect on service provider loyalty.Coefficient