FinTech Adoption, Bank Profitability, and Stability: The Mediating Role of Operational Cost Efficiency in Bangladesh's Commercial Banking Sector

arXiv:2611.04298 · q-fin.GN · Submitted 2026-11-03 · Read on arXiv

Listen

Radio episode about this paper

Transcript

Introduction to the show: ident: Quantitative Finance. The day in quantitative finance: models of markets, risk and prices.

Ruth: We're looking at the paper: "FinTech Adoption, Bank Profitability, and Stability".

Warren: Here's the gist of the paper. FinTech adoption presents no significant relationship with profitability, stability or cost efficiency, while cost efficiency does not significantly transmit the effect of FinTech to either outcome Research Objectives The study explores the direct impact of a composite FinTech Adoption Index on bank profitability (ROE) and bank stability (ln(Z-score)) along with whether operational cost efficiency (CIR) acts as a mediator.

Ruth: The research questions investigate the direct effects of FinTech adoption on ROE and ln(Z-score), its impact on CIR, and whether CIR mediates the relationship between FinTech adoption and both outcomes.

Warren: Hypotheses Testing The study tests several hypotheses regarding the relationships between the variables.

Ruth: The primary hypotheses include H1, which posits that FinTech adoption (FTI) is positively associated with bank profitability (ROE), and H2, which suggests that FTI is positively associated with bank stability ln(Z-score).

Title and authors: Ruth: They’re not just stopping at saying the direct effects are insignificant; they’re pointing out that the current measurement of FinTech adoption might be too broad for what actually matters in this market.

Warren: They want to move beyond just using that one composite index and start comparing it against this entropy-weighted index more rigorously to see if performance findings are driven by specific digital components.

Iris: That makes sense because as we saw, the entropy measure showed a significant negative link to return on equity, so they think that’s where the real signal might be hiding in the data.

Ruth: They're pushing for models that account for how bank performance evolves over time, not just taking a snapshot with their ten-year panel design.

Warren: And they’re looking at using two-step System GMM estimation to handle the persistence of those variables better, because a static estimator would cause some bias when you only have a short panel like that.

Iris: That's good; it checks for identification issues by seeing what variation actually does the work and what assumptions they are making about the data structure.

Ruth: They’re also interested in building an operational cost efficiency prediction module that tests the mediation hypothesis more directly, looking at whether it truly mediates between FinTech adoption and risk-taking.

Warren: And they want to test those mediation effects in other contexts, maybe comparing how much operational efficiency really mediates the relationship in different banks or different countries.

Iris: They also suggest incorporating macroeconomic variables like GDP growth rate into the models because that variable consistently shows significance for both raising return on equity and lowering operational cost efficiency.

Ruth: It seems they’re calling for a more granular approach to testing, moving from one big composite measure to comparing different ways of measuring adoption and looking at dynamic modeling techniques.

Warren: Basically, they want the next generation of research to look deeper into the specific components of digital infrastructure and how bank performance actually changes over time.

Iris: It points toward a need for more nuanced econometric testing, focusing on identifying which specific drivers—like those macroeconomic factors or different aspects of technology—are truly responsible for the observed effects.

Ruth: So, they’re advocating for a shift in focus from just measuring adoption volume to understanding the actual quality of that adoption and its dynamic impact on the bank's core performance metrics.

The paper's summary: Warren: To wrap up this look at FinTech Adoption, Bank Profitability, and Stability: The Mediating Role of Operational Cost Efficiency in Bangladesh's Commercial Banking Sector, the main thing is that they didn't find a strong direct link between FinTech adoption and those bank results based on their primary measure.

Ruth: Yeah; the paper suggests that just having more technology isn't automatically translating into better profits or stability for these banks in Bangladesh right now, based on their main analysis of the composite index.

Iris: And they also couldn't find that operational cost efficiency variable acting as a middleman to explain why FinTech adoption affects those outcomes in this specific model structure.

Warren: That means the direct connection between FinTech adoption and return on equity, or that log of the Z-score for stability, isn't strong based on their main regression results.

Ruth: So it’s a call for banks to focus on organizational change and employee capabilities because the technology itself isn't delivering the immediate performance boost they hoped for.

Iris: They’re also pointing toward a need for more nuanced econometric testing, focusing on identifying which specific drivers—like those macroeconomic factors or different aspects of technology are truly responsible for the observed effects.

Warren: And they want future research to look deeper into the specific components of digital infrastructure and how bank performance actually changes over time, not just taking one big composite score as the main driver.

Ruth: It’s a call for deeper analysis on how performance changes dynamically and across different types of banks to get a fuller picture of how technology actually impacts them.

Iris: And they also suggest incorporating macroeconomic variables like GDP growth rate into the models because that variable consistently shows significance for both raising return on equity and lowering operational cost efficiency.

Warren: That leaves us thinking about how these findings compare to other markets, maybe testing if this relationship holds true in countries with different levels of digital maturity.

The paper's improvements: Ruth: So, to wrap up this discussion on FinTech Adoption, Bank Profitability, and Stability: The Mediating Role of Operational Cost Efficiency in Bangladesh's Commercial Banking Sector, we’ve seen that the direct links between FinTech adoption and bank performance haven't been strongly confirmed by their main measure.

Warren: And the important caveat is that operational cost efficiency didn't play a role as a mediator in either profitability or stability outcomes, according to their analysis.

Iris: It suggests we might need to look deeper into the specific components of digital transformation rather than just taking one overall FinTech score as the main driver of performance.

Ruth: And for what this means practically, banks need to focus on internal changes and skills development because the technology itself isn't delivering the promised performance boost yet.

Warren: And regulators need to look beyond just adoption numbers and focus more on the quality of that adoption, checking if it's leading to actual efficiency improvements in their systems.

Iris: I just think the focus on identifying those specific drivers is crucial because without that, you’re just looking at a bunch of noise across all the different digital indicators.

Ruth: Yeah; that's right; the authors are basically saying the current tools aren't enough to tell us what works in this adoption phase.

Warren: It’s a good reminder that we need to be careful about how we measure things when you try to link them directly to bank results.

Iris: So the lesson is that operational efficiency isn't just some passive background variable; it needs to be actively managed alongside the tech rollout for any real payoff.

Ruth: Exactly; we’ll keep an eye on this paper as we look at how these specific digital components actually interact with bank risk in our next deep dive.

Conclusion: Ruth: So, to wrap up this discussion on "FinTech Adoption, Bank Profitability, and Stability: The Mediating Role of Operational Cost Efficiency in Bangladesh's Commercial Banking Sector," we’ve seen that the direct links between FinTech adoption and bank performance haven't been strongly confirmed by their main measure.

Warren: And the important caveat is that operational cost efficiency didn't play a role as a mediator in either profitability or stability outcomes, according to their analysis.

Iris: It suggests we might need to look deeper into the specific components of digital transformation rather than just taking one overall FinTech score as the main driver of performance.

Ruth: And for what this means practically, banks need to focus on internal changes and skills development because the technology itself isn't delivering the promised performance boost yet.

Warren: And regulators need to look beyond just adoption numbers and focus more on the quality of that adoption, checking if it's leading to actual efficiency improvements in their systems.

Iris: They’re also pointing toward a need for more nuanced econometric testing, focusing on identifying which specific drivers—like those macroeconomic factors or different aspects of technology—are truly responsible for the observed effects.

Ruth: It’s a call for deeper analysis on how performance changes dynamically and across different types of banks to get a fuller picture of how technology actually impacts them.

Warren: That leaves us thinking about how these findings compare to other markets, maybe testing if this relationship holds true in countries with different levels of digital maturity.

Iris: I just think the focus on identifying those specific drivers is crucial because without that, you’re just looking at a bunch of noise across all the different digital indicators.

Ruth: Yeah, that's right. The authors are basically saying the current tools aren't enough to tell us what works in this adoption phase.

Warren: It’s a good reminder that we need to be careful about how we measure things when you try to link them directly to bank results.

Iris: So, the lesson is that operational efficiency isn't just some passive background variable; it needs to be actively managed alongside the tech rollout for any real payoff.

Ruth: Exactly. We’ll keep an eye on this paper as we look at how these specific digital components actually interact with bank risk in our next deep dive.

Warren: Next week, we’re shifting gears and looking at how complex networks, like those seen in financial systems, actually affect the valuation of things like corporate bonds.

Shabab Sabiqa, Progga Biswas Paramab

University of Barishal · Eden Mohila College

q-fin.GN

Submitted: 2026-11-03

Updated: 2026-11-03

Comments: 23 pages, 8 tables, 1 figure. JEL: G21, G28, G32, O33

License: http://arxiv.org/licenses/nonexclusive-distrib/1.0/

Importance score: 72/100

The gist: FinTech adoption presents no significant relationship with profitability, stability or cost efficiency…

Key concepts

FinTech Adoption Index (FTI)
This is a composite measure used to gauge how much banks in Bangladesh have adopted financial technology solutions. It combines various digital indicators into one score to represent the overall level of tech integration within commercial banking operations.
Bank Profitability (ROE)
Return on Equity (ROE) measures how effectively a bank uses its shareholder money to generate profit. The study investigated whether increased FinTech adoption directly influences this measure, finding no statistically significant link.
Operational Cost Efficiency (CIR)
This concept refers to how well a bank manages its day-to-day expenses relative to its revenue. It measures the efficiency of internal processes, and the research tested if FinTech adoption improved or changed this efficiency.
Mediation Analysis
This statistical technique checks if one variable (FinTech Adoption) influences another outcome (Profitability/Stability) only through a third variable (Cost Efficiency). The study found that operational cost efficiency did not significantly mediate the relationship between FinTech and performance.

Summary

FinTech adoption presents no significant relationship with profitability, stability or cost efficiency, while cost efficiency does not significantly transmit the effect of FinTech to either outcome

Research Objectives

The study explores the direct impact of a composite FinTech Adoption Index on bank profitability (ROE) and bank stability (ln(Z-score)) along with whether operational cost efficiency (CIR) acts as a mediator. The research questions investigate the direct effects of FinTech adoption on ROE and ln(Z-score), its impact on CIR, and whether CIR mediates the relationship between FinTech adoption and both outcomes. p. 4

Hypotheses Testing

The study tests several hypotheses regarding the relationships between the variables. The primary hypotheses include H1, which posits that FinTech adoption (FTI) is positively associated with bank profitability (ROE), and H2, which suggests that FTI is positively associated with bank stability ln(Z-score). p. 2 Furthermore, the study tests the hypothesis that operational cost efficiency (CIR) mediates the relationship between FinTech adoption and bank profitability (H4) and between FinTech adoption and bank stability (H5). p. 4

Empirical Findings on Direct Effects

The main dynamic System GMM regression results show that FTI is statistically insignificant for ROE (β = 0.0023, t = 0.07) and ln(Z-score) (β = 0. p. 00328, t = 0.65), thus not supporting H1 and H2. p. 6 The robustness check using the entropy-weighted index (FTIEWM) shows that FTIEWM is significantly negatively associated with ROE (β = −0. p. 40570, t = −2.08, p < 0.05), suggesting an absence of a direct profitability effect under this measure. p. 2

Mediation Analysis Results

The analysis for the mediating role of operational cost efficiency (CIR) indicates that FTI is insignificant in the Model 3 first stage and CIR in Models 4 and 5, meaning neither condition for mediation is met. This null result is replicated under FTIEWM in both mediated models, lending confidence to the conclusion regardless of FinTech measurement. p. 3

Theoretical Implications

The findings qualify rather than overturn Transaction Cost Theory and Diffusion of Innovation Theory. The direct impacts were mostly null, consistent with the early stages of the adoption process where fixed costs remain high and are likely to be the reason for cost savings expected by these theories not yet being realised. The lack of mediation suggests that the role of operational efficiency has yet to be activated in this context. p. 3

Managerial and Policy Implications

The findings indicate that while the introduction of FinTech solutions is not enough in isolation to drive measurable returns for banks, the process probably requires some support from organisational change, employees' capabilities and processes. To Bangladesh Bank and the regulators, the message is not to slow down FinTech growth but to pair it with closer supervisory attention to adoption quality, as the intensity of adoption is not an accurate measure of better bank performance. p. 1

Limitations and Future Research

This study is limited in scope by its single-country design, a ten-year panel which may understate effects unfolding over a longer adoption horizon, and by the composite FinTech measure which is crossvalidated with an alternative based on the entropy measure of digital indicators but still captures only a small number of digital indicators. Future studies may investigate whether these are found in other countries, if the relationship between FinTech and performance is nonlinear and not linear, and if the impact of FinTech differs between state-owned and private banks, or across banks of different sizes. p. 5

Green — the paper says it, and the page is printed. Red — no passage carries it, so treat that sentence as unverified. The same check runs on your own text: check a summary against its source, $5.

Improvements for AI systems

  1. Incorporate a measurement-specific sensitivity assessment into FinTech adoption modeling by explicitly comparing results from the composite FinTech Adoption Index (FTI) against an entropy-weighted index (FTIEWM). This allows AI systems to determine if performance findings are driven by the specific components of digital infrastructure rather than just a general measure, as noted in The robustness check qualifies this picture: FTIEWM is significantly related to lower ROE and improved cost efficiency.

  2. Develop predictive models that account for the dynamic persistence of bank performance by utilizing Two-step System GMM estimation. This addresses the issue that the lagged dependent variable would induce Nickell (1981) bias under a static estimator given the short panel (T = 10) and ensures more accurate estimates of how FinTech adoption impacts ROE and ln(Z-score).

  3. Create an operational cost efficiency prediction module that tests the mediation hypothesis by comparing direct effects with mediated models. This allows the AI to evaluate whether operational cost efficiency fully mediates between FinTech and risk-taking in specific contexts, as suggested by in Pakistani and Bangladeshi banks, operating efficiency fully mediates between FinTech and risk-taking.

  4. Implement a dynamic control mechanism for macroeconomic variables by prioritizing GDP Growth Rate as the most consistently significant predictor across models. This ensures that the AI systems correctly weight the impact of economic expansion on both profitability and cost-to-income ratios, noting that "GDPGR is the most consistently significant predictor, raising ROE (β = 0.4002, p < 0.01) and lowering CIR (β = −1.2326, p < 0.01)."

Related papers