1. Introduction
Achieving sustainable development has become paramount given the current sustainability crises globally. Among the sustainability crises is the lack of financial access, which exacerbates poverty and inequality and slows the pace of economic activities in both developed and developing countries. The innovation in digital finance has emerged as a critical driver of Sustainable Development Goals (SDGs) through universal financial inclusion, which holds significant promise for both developed and developing economies [
1], making digital finance especially the mobile money industry of utmost importance to development practitioners and policymakers. The emergence, adoption growth, and continuance usage of MPT in emerging economies, particularly Ghana has positioned the industry as a viable alternative to generate revenue as the mobile transaction value has increased by about
$81 billion [
2]. Rapid mobile device penetration has fueled mobile payment growth, positioning Ghana as a leading mobile money market. As of 2022, Ghana recorded 55.3 million mobile money accounts, with active transactions totaling GHS 122 billion [
3].
While there has been growing interest in mobile payment technology and several studies have been conducted, there remain significant gaps in the extant literature. Firstly, the literature connecting TBL dimensions (economic and social environmental impact) on perceived behavioral control, subjective norms, and attitudes for behavioral intention and the adoption of sustainable behaviors through the usage of mobile payment technology (MPT) is limited. Past studies concentrate on only TPB and other dominant adoption theories [
4,
5,
6,
7]. Also, there is fragmented evidence on the intention and adoption of a sustainable behavior gap [
8,
9]. Some past studies have demonstrated that not all intentions lead to actual behaviors [
10], a gap this study seeks to fill. Therefore, moving beyond psychological behavioral factors affecting behavioral intentions and adoption, this study makes a significant contribution to extending TBL dimensions for sustainable behaviors.
Secondly, some scholars have argued that the adoption of MPT does not always translate into financial inclusion and, hence, presents contradictory findings [
11,
12], and its contribution to sustainable development through financial inclusion may be unclear. Thus, the argument is centered on the high cost of transaction charges, and soft loan accessibility traps users into perpetual indebtedness which may potentially have a negative impact on financial inclusion when users make trade-off decisions between continuance usage and discontinuance usage. This uncertainty calls for further research, although there has been a significant positive relationship obtained by past studies on this subject matter [
13,
14].
Presently, the mediation role of perceived behavioral control, subjective norms, and attitude are scarce in past studies [
15], making this study novel in this direction. Thus, individuals’ behavioral intentions can be greatly impacted by the economic, social, and environmental impact of MPT and not merely perceived behavioral, subjective norms, and attitudes. Past studies have demonstrated the possible mediation effects of perceived behavioral, subjective norms, and attitudes on the beliefs and behavioral intention for mobile payment through beliefs [
16]. Supporting this is the claim by Ajzen that other factors might influence behavioral intention through TBP constructs [
17]. Again, Altawallbeh et al. [
18] argued in their study that there exists a mediation effect of TPB constructs, salient beliefs, and e-learning intention. The growing consumer awareness of sustainability issues triggers how consumers make adoption decisions.
Furthermore, it is unclear whether the taxation policies by African governments in the mobile payment industry moderate the relationship between the adoption of sustainable usage behaviors and financial inclusion. Research has investigated the consumers’ perception of MTTP for mobile money uptake [
19], the adoption factors of MTTP in advancing continuance usage [
20,
21], MTTP on price value and customer satisfaction [
22], the impact of MTTP on informal workers [
23], and MTTP as a moderator between financial well-being and usage [
19]. While some of these studies relied on only qualitative approaches [
19], little is known about its moderation effects on the relationship between adoption and financial inclusion. Several calls to remove MTTP have been made because its negative externalities on usage and financial inclusion reflect its potential to drive adoption decisions and financial inclusion [
24,
25].
Additionally, there is a lack of theoretical grounding on the nexus of MTTP, the adoption of sustainable usage behaviors, and financial inclusion due to the predominance of previous study reports and literature reviews [
19,
26,
27,
28]. Again, there are no well-established theories that have built a conceptual framework in previous studies regarding this relationship. Additionally, the current literature emphasizes the direct effect of MTTP on the continuance usage of MPT [
29,
30]. Hence, by filling this gap, this study applies a PLS-SEM approach to provide empirical evidence on the subject matter. Also, the sustainability TPB was used to develop a conceptual framework to investigate the nexus of MTTP, the adoption of sustainable usage behaviors, and financial inclusion.
Notwithstanding the gains to financial inclusion that are attributed to the mobile payment technology in Ghana, many fear MTTP, the issue of fraud, and the other challenges that have marred the industry and made the usage of the MPT unpopular; hence, appropriate strategies must be adopted to sustain the industry; among these strategies is the role of consumer behavior. As consumers are increasingly aware of prevailing sustainability issues and align their adoption decision in that regard, this study is significant in uncovering the antecedents of behavioral intention, the adoption of sustainable usage behaviors of the MPT, its implication on financial inclusion, and how MTTP moderates the nexus of the adoption of sustainable usage behaviors of the MPT and financial inclusion. Given the unique socioeconomic and cultural landscape of Ghana’s mobile payment industry, this would provide valuable insights to stakeholders and service providers with the impetus of translating behavioral intention into financial inclusion. From the viewpoint of development partners and policymakers’ efforts to achieve the UN Sustainable Development Goals by 2030, the mobile payment industry’s contributions could be promoted and sustained by utilizing the findings of this study.
The rest of the paper is organized as follows:
Section 2 presents the literature and hypothesis development;
Section 3 details the research methodology, and
Section 4 and
Section 5 present the results and discussion, respectively. The last section,
Section 6, contains the conclusion, theoretical contribution, practical implication and limitation, and suggestions for further research.
4. Results
This part of the data analysis presents the PLS-SEM result; it evaluates the research hypotheses concerning the antecedents of sustainable behaviors toward financial inclusions and the moderation role of the mobile transaction tax policy. It also details the findings of the measurement and the structural model.
4.1. Common Method Bias
Common method bias (CMB) is often a challenge in survey research because both independent and dependent variables are measured by the same survey using the same response technique [
152], making the reliability and validity of the empirical findings questionable, so methods to identify it, avoid it, and control it are essential [
153,
154]. CMB control is grouped into two ways, that is, procedural and statistical control [
152].
For procedural controls, the study designed a clear questionnaire by conducting a pilot study to test the questionnaire. Ambiguous and unclear wording was revised before the main study. Also, the constructs in our conceptual model were categorized in distinct sections using varied scales to maintain a proper sequence of the indicators. Additionally, all the questions designed as measures of the constructs maintained their unique terminologies identified in the literature. Thus, there was no compatibility in the questions asked on TBL dimensions, TPB constructs, and financial inclusion.
For the procedural control of CMB, Harman’s single-factor test was employed to determine the degree of the measurement error in the constructs under study that is caused by the measurement method. It determines whether a single factor accounts for the majority of the variance. The result obtained from the first unrooted factor revealed multiple factors. A single factor explained up to 27.9 percent of the variance which is below the threshold of 50% [
155]. This signifies the absence of CMB issues in this study.
4.2. Respondents’ Demographic Profile
This section of the study reports the demographic characteristics of the respondents. The data obtained are used for the PLS-SEM analysis presented in
Table 2. A total of 320 responses were received from consumers of MPT from Accra and Tamale. Of this sample, the majority of the respondents are males and well educated, with nearly half having a bachelor’s degree. Most earn between GHS 3001 and 4000 monthly and use mobile payment services frequently, with 78.8% engaging 10–14 times per week. MTN Mobile Money is the preferred platform (78.8%), aligning with broader trends in mobile payment adoption among middle-income, educated demographics in Ghana [
156,
157].
4.3. The Measurement Model
The research explored several steps to ensure the accuracy of the study findings. First, the pilot study helped refine the measurement scale. Additionally, experts were recruited to verify the research instruments alongside carefully selecting the instruments from previous studies published in high-quality journals. Finally, we performed confirmatory factor analysis (CFA) and exploratory factor analysis (EFA) to establish the reliability and validity of the measurement model. Thus, the EFA measures consist of the values of the various indicators’ loadings and Cronbach’s alpha, and composite reliability was used to determine convergent reliability, internal consistency, and reliability.
Table 3 below shows the various indicators’ outer loadings, varying from 0.644 to 0.934. Also, for internal consistency, the composite reliability values fall within the acceptable range of 0.831 to 0.978. The latent variable average variance explained (AVE) has values ranging from 0.636 to 0.938, which is above the recommended level [
158,
159]. It also confirms that the variables observed present good convergent validity. The Cronbach’s alpha coefficients which were also used to assess the internal consistency fall within the threshold of 0.735 and 0.967 [
160].
Table 4 also consists of the cross loadings.
Additionally, confirmatory factor analysis was used to determine discriminant and convergent validity. As such, the average variance extracted (AVE) and cross-loading evaluated convergent validity, while the Fornell and Larcker criterion and the Hetrotrait–Monotrait Ratio (HTMT) were employed to assess discriminant reliability, which is the robustness of the empirical distinctiveness.
Table 5 presents the Fornell and Larker criterion which shows that all the indicators’ square root of their AVE was higher than their correlation with other variables and demonstrates the required discriminant validity. On the other hand, HTMT compares average correlations between different constructs (heterotrait) to those within the same construct (monotrait); that is, it uses the geometric correlations between the items used to assess a given construct, and the results are shown in
Table 6. All the values were below the threshold of <0.9 [
161]. Consequently, the analysis of the measurement model showed the quality of the latent constructs and measurement items.
4.4. The Structural Model
After analyzing the measurement model and reliability and validity have been established, the structural model analysis was conducted to evaluate the relationships of the formulated hypotheses using the coefficient of determination (R
2), predictive relevance (Q
2), and the statistical significance of the path coefficients. However, we first determine the variance inflation factor (VIF). A variance inflation factor (VIF) value exceeding 3.3 may indicate potential multicollinearity issues and could be a sign of common method bias (CMB). Thus, if all VIF values in the inner model are 3.3 or lower, the model is generally considered free from CMB concerns.
Table 7 shows that the variance inflation factor (VIF) met the recommended guidelines by Hair et al. [
161]. Constructs with VIF values lower than 3.3 were considered free from multicollinearity issues. All constructs in the study exhibited VIF values between 1.0 and 1.609, meeting the recommended threshold [
162].
For the predictive relevance of the structural model, the values of the Q
2 were used. According to Hair et al. [
162], Q
2 values should significantly exceed zero to indicate that the exogenous structure predicts the endogenous structure. Our results to achieve the dependent variable and the financial inclusion of mobile payment technology showed cross-validation values exceeding zero, as presented in
Table 8. These values indicate that the test results were good, signifying the predictive relevance of the framework. Again, the study assessed the predictive power of the model, where out-of-sample prediction is resolved using PLSpredict on the SmartPLS software version 4. [
163]. Accordingly, the mean absolute error (MAE) and root mean squared error (RMSE) for each indicator were evaluated against a baseline linear regression (LM) model. As presented in
Table 8, the PLS-SEM model demonstrates strong predictive capability, with all indicators showing lower RMSE values compared to the naïve LM benchmark [
163].
Finally, the strength and significance of the correlation between variables were assessed using the R
2 value, indicating a substantial, moderate, or weak relationship if the R
2 value surpassed 0.67, 0.33, or 0.19, respectively [
164]. In our analysis, the R
2 value demonstrated a high and moderate relationship. For example, perceived behavioral control explains up to 45.5%, subjective norm up to 56%, attitude up to 33%, intention to adopt behavioral behavior up to 66%, the adoption of sustainable usage behaviors up to 33.1%, and financial inclusion up to 59.2%. Therefore, all criteria were considered met for the structural model. A bootstrapping method was further conducted to test the formulated research hypotheses through the significance of the path coefficients. We conducted bootstrapping using 5000 subsamples randomly extracted from the data set at a 95% confidence interval. This was used to assess the statistical significance and magnitude of the interrelationships between the constructs within the model using SmartPLS 4 [
165]. The results of the bootstrap subsamples are presented in
Table 9 in the form of significance levels, t-values, and the path coefficients for each model parameter.
4.5. Hypotheses Testing and Path Coefficients
4.5.1. Direct Effects
All the hypotheses formulated from the TPB and its extended constructs effectively explained the intention, the adoption of sustainable behaviors, and financial inclusion. Firstly, concerning the influence of TPB constructs on the intention to adopt sustainable behaviors through the usage of MPT, our findings showed that perceived behavioral control (β = 0.210, t = 4.517, and p = 0.000) and attitude (β = 0.231, t = 3.30, and p = 0.001) were statistically significant with positive coefficients, except for subjective norms (β = 0.088, t = 1.229, p = 0.219). Thus, hypotheses H1 and H3 were supported, while H2 was not supported.
Secondly, the relationship between behavioral intention to adopt and the actual adoption of sustainable usage behaviors through MPT revealed a positive and statistically significant relationship (β = 0.223, t = 4.728, and p = 0.000). Hence, hypothesis H4 was supported. To test the relationship between sustainable usage behavior and financial inclusion, the empirical result was positive and statistically significant (β = 0.567, t = 8.956, and p = 0.000). Thus, hypothesis H5 was supported.
Thirdly, for the influence of the TBL dimensions on TPB constructs, the results of the perceived economic impact effect on perceived behavioral control, subjective norms, and attitude consisting of hypotheses H6a, H6b, and H6c were all positive and statistically significant at (β = 0.366,
t = 4.421, and
p = 0.000), (β = 0.556,
t = 8.038, and
p = 0.000), and (β = 0.224,
t = 2.795, and
p = 0.003), respectively. The findings on the perceived social impact relationship on perceived behavioral control, subjective norms, and attitude consisting of hypotheses H7a, H7b, and H7c are (β = 0.132,
t = 1.680, and
p = 0.093), (β = 0.187,
t = 2.357, and
p = 0.018), and (β = 0.249,
t = 2.823, and
p = 0.005). Additionally, the results of the perceived environmental impact influence on perceived behavioral control, subjective norms, and attitude consisting of hypotheses H8a, H8b, and H8c are (β = 0.132,
t = 2.360, and
p = 0.019), (β = 0.133,
t = 3.530, and
p = 0.000), and (β = 0.498,
t = 7.601, and
p = 0.000). All triple bottom line constructs on TPB constructs are positive and showed statistical significance. Therefore, all hypotheses are supported. Again, the path diagram of the study model, consisting of the path coefficients and the statistical significance, is presented in
Table 9 and
Figure 2.
4.5.2. Mediation Analysis (Indirect Effects)
The mediation analysis was conducted to establish whether the triple bottom line dimension affects the TPB constructs and translates into the behavioral intention to adopt the sustainable usage of MPT. With this, the influence of a mediating variable can determine how behavioral intention behaves and relates to independent variables, that is, the triple bottom line dimensions. The results obtained are presented in
Table 10. It shows that the TPB constructs mediate the relationship between the triple bottom line dimensions and the intention to adopt sustainable usage behaviors.
The findings of the perceived economic impact effect on TPB constructs, which in turn influence behavioral intention to adopt sustainable behaviors through MPT, consisting of hypotheses H10a, H10b, and H10c were all positive and statistically significant at (β = 0.132, t = 1.680, and p = 0.093, (β = 0.366, t = 4.421, and p = 0.000), and (β = 0.223, t = 4.727, and p = 0.000), respectively. The results of the impact of perceived social impact on TPB constructs, which in turn derive behavioral intention to adopt sustainable behaviors, with hypotheses H11a, H11b, and H11c showed positive and statistically significant values of (β = 0.113, t = 2.159, and p = 0.031), (β = 0.368, t = 4.422, and p = 0.000), and (β = 0.234, t = 2.975, and p = 0.003), respectively. Additionally, regarding the perceived environmental impact influence on TPB constructs, the hypotheses H12a, H12b, and H12c with (β = 0.132, t = 0.352, and p = 0.0019), (β = 0.210, t = 4.517, and p = 0.000), and (β = 0.234, t = 2.975, and p = 0.003) were all positive and statistically significant. All TPB constructs effectively mediate the relationship between TBL dimensions and the behavioral intention to adopt the sustainable behaviors of MPT.
4.5.3. Moderation Analysis
Finally, we performed a moderation analysis to determine how the mobile transaction tax policy affects the relationship between adopting sustainable usage behaviors and financial inclusion through MPT using a 2-stage approach since our moderator is a formative construct [
166]. The findings of the interaction between the mobile transaction tax policy and sustainable usage behaviors had a significant negative effect on financial inclusion, with values of (β = −0.234,
t = 2.975, and
p = 0.003), as shown in
Table 11.
6. Conclusions
This study contributes to knowledge in financial inclusion, consumer behavior, and sustainable development, particularly within the context of mobile payment technology. There is consensus in the literature that the emergence, adoption, growth, and continuance usage of mobile payment technology have led to the achievement of some of the SDGs [
25,
169,
170]. However, the sustainability of the mobile payment industry is jeopardized by multiple barriers coupled with the implementation of mobile payment taxation in the industry. Accordingly, sustainable policies and behavioral strategies are needed to sustain the mobile payment industry toward financial inclusion and sustainable development.
The extant literature holds inconsistent views on the intention behavior gap and the impact of mobile payment usage on financial inclusion, especially from the lance of TBL dimensions’ influence on TPB constructs. To respond to these gaps, we sought to (1) explore the antecedents of consumers’ intention and the sustainable usage behaviors of mobile payment technology and its implication for digital financial inclusion and (2) examine the moderation role of the mobile payment transaction tax policy on the relationship between the sustainable usage of mobile payment technology and digital financial inclusion.
By using PLS-SEM to test the twelfth hypothesis, the following results were obtained. We found that all the constructs of the sustainability TPB including the perceived economic impact, the perceived social impact, the perceived environmental impact, perceived behavioral control, and attitude, intention, adoption, and financial inclusion were all supported except subjective norms. We also confirmed that MTTP negatively affects the nexus between the sustainable usage of mobile payment technology and digital financial inclusion. This study demonstrates that sustainability TPB can be employed to explain financial inclusion, especially with the use of MPT, which accounts for the large population who benefits from its usage.
6.1. Theoretical Contribution
This study makes a significant theoretical contribution to consumer behavior, taxation, financial inclusion, and the sustainability literature by addressing unexplored sustainable behaviors to ensure the sustainability of the MPT.
Accordingly, this thesis extends the adoption of the technology literature, specifically by studying the antecedents of adopting sustainable behaviors. The innovation of this part of the study lies in the integration of TPB with TBL dimensions to study the sustainable behaviors of the MPT. While TPB has been widely explored with other matured theories in extant studies, sustainability TPB serves as a foundational framework for application in the consumer behavior literature toward financial inclusion. In this subject matter, this study provides valuable insights to policymakers and guides industry practitioners and relevant stakeholders.
The lack of consensus regarding intention-adoption of sustainable behavior gab has also been examined in this study.
In addition, this study addresses the call for MTTP removal. It confirms the public outcry that MTTP in the mobile payment industry inhibits the smooth transition of sustainable behaviors to financial inclusion in developing countries where mobile payment technology is a vital part of filling in the financial access gap left by traditional banking institutions. Particularly, this study enriches the current literature by providing a comprehensive understanding of how MTTP affects sustainable usage and financial inclusion and how it can be reviewed to provide avenues for revenue generation while encouraging usage simultaneously.
This study explores the broader impacts of taxation strategies on the economic impact, environmental outcomes, political support, and social inclusion in Africa’s MTTP setting. Consequently, this study affirms the TBL dimensions in the financial literature, offering a replicable model for developing economies to design sustainable mobile taxation strategies. This research advances theoretical discourse in the MTTP frame while providing actionable insights to promote financial inclusion and equity in taxation policy development. Other studies can use this framework to study a related phenomenon in a new context.
6.2. Managerial and Practical Implication
The MPT’s sustainable performance contributes to the broader universal financial inclusion and consequently sustainable development; hence, the findings of this study are critical for service providers, as businesses are developing sustainable business models that guarantee long-term viability in terms of economic, environmental, and social inclusion that meets the needs of the growing demands of the environmentally conscious consumer. This study offers considerable managerial and practical contributions and implications of the PLS-SEM approach.
Firstly, this study provides insight into the most pertinent factors affecting the sustainable behaviors of MPT and its implication on financial inclusion. The findings confirm that sustainable behaviors significantly influence financial inclusion. It highlights the positive influence of the TBL dimensions on attitude, perceived behavioral control, and subjective norms on consumers’ decisions to continue the usage of MPT instead of conventional payment methods, aligning with sustainable principles. The importance of sustainable behaviors in the mobile payment industry is paramount for maintaining the sustainable development of the mobile payment industry, hence providing support in achieving sustainable development. Therefore, MPT service providers must be proactive in designing friendly interfaces with sustainable features that will positively influence consumers’ attitudes. They should also engage in public awareness by creating and educating consumers on the benefits and usage of MPT.
Secondly, industry players in the mobile payment industry must have a clear vision to contribute to financial inclusion, actively promote sustainable initiatives, and invest in resources necessary for attracting usage. The results show that policies implemented in the mobile payment industry such as the MTTP possess a negative externality and are disincentive to sustainable usage and financial inclusion. Hence, it is imperative for service providers to vehemently reject this tax levy by advancing their disapproval and therefore negotiate with policymakers for the best alternative means for revenue generation in the industry.
Because of these considerable implications, the results present policy implications for sustaining mobile payment technology. First, the multiple characteristics of factors of sustainable behaviors of the MPT that influence financial inclusion suggest that finding the antecedents from a sustainability perspective is essential. Specifically, the MTTP implemented in the mobile payment industry increases transaction costs, leaving consumers to opt for alternative payment methods that negatively affect financial inclusion. By eliminating mobile taxation in the mobile payment industry, governments can support at least eight UN SDGs, demonstrating the transformative potential of well-structured policies in fostering innovation, inclusivity, and sustainability in the mobile payment sector, or policymakers should consider withdrawing this policy or review it to meet the broader stakeholder needs.
6.3. Limitations and Recommendations for Further Studies
While the thesis has utilized a robust theoretical foundation and has presented valuable theoretical contributions, as well as practical and policy implications on the dynamics of sustaining the mobile payment technology in developing countries, especially in Ghana, it is not without limitations, and these limitations present opportunities for further research.
First, this study relied on data from Ghana to draw generalizations which can potentially limit the findings. Therefore, future studies can use the frameworks applied in this study to conduct comparative studies using different contexts where mobile payment technology is widely adopted to validate the findings and offer additional contextual insights on the subject matter. This will ensure a better understanding of how different geographical variations in terms of culture, national policies, or regional context might affect the sustainability of mobile payment technology and its implication on financial inclusion to broaden the scope of this concept’s applicability.
Also, our study captured only behavioral factors and TBL dimensions influencing behavioral intention for sustainable mobile payment technology and consequently financial inclusion. However, other factors could also play a role like cost, technical issues, etc. Future studies can incorporate different models to provide a comprehensive understanding of the intricate relationship between behavioral intention, the adoption of sustainable mobile payment technology, and financial inclusion
The last limitation is that the PLS-SEM applied in this study cannot account for non-linear relationships and complex dynamics. We recommend that future studies integrate MCDM methods and statistical methods using the framework examined in this study.
Lastly, the study made use of only adopters of MPT; we acknowledge this as a limitation and recommend that future studies incorporate both adopters and non-adopters to provide a robust relationship between attitudes, intentions, and behaviors.