The Impact of Technological Development on the Productivity of UK Banks

Abstract

This study investigates the impact of digitalisation and intangible investment—specifically digital skills and software adoption—on productivity in the United Kingdom’s banking sector. Software adoption is captured through banks’ investment in enterprise systems (CRM/ERP, cloud computing, and related applications), rather than a single software version. Drawing on detailed bank-level data from six major UK banks over the period 2007–2022, this research provides empirical evidence that higher intensities of digital human capital and intangible assets are positively associated with improvements in both employee productivity and overall bank performance. A standard deviation increase in software specialist employment is associated with productivity gains of 10.3% annually, though this upper-bound estimate likely combines direct effects with complementary factors such as concurrent IT investments (e.g., cloud infrastructure) and managerial innovations. The findings also highlight substantial heterogeneity across banks, with younger institutions experiencing more pronounced benefits from intangible investment due to their greater flexibility and innovation capacity. Furthermore, this study reveals that the adoption of high-speed internet and investment in IT hardware have a strong positive effect on bank productivity, particularly in the wake of the COVID-19 pandemic, which accelerated digital transformation across the sector. However, the observational nature of the study and the limited sample size necessitate caution in generalising the findings. While the results have implications for digital workforce development and technology infrastructure, policy recommendations should be interpreted as preliminary, pending further validation in broader samples and diverse institutional settings. This study concludes by advocating for targeted strategies to expand digital skills, promote software diffusion, and modernise infrastructure to facilitate productivity convergence, while emphasising the need for future research to address potential endogeneity and external validity limitations.

1. Introduction

Over the past few years, the global economy has witnessed a rapid spread of disruptive technologies, leading to fundamental shifts in business models across all economic sectors. One of the most significantly affected sectors was the financial industry, which is currently undergoing a major transformation driven by digital innovation. Since the 1950s, the financial sector has seen continuous technological evolution, beginning with the introduction of Automated Teller Machines (ATMs), which enabled self-service banking and reduced reliance on human factors. This was followed by the invention of payment cards, which reduced the need to carry cash.

According to Breidbach et al. [1], the adoption of new technologies in the financial sector has been gradual yet consistent. The rise in internet access in the 1990s enabled customers to perform online banking transactions 24/7. In the 21st century, technologies such as cryptocurrencies and block chain have made it possible to conduct financial transactions without relying on traditional centralised intermediaries.

Interestingly, despite substantial technological advancements over the past six decades, the financial industry has not been destabilised; on the contrary, it has thrived [2]. Established financial institutions, particularly large banks, have been compelled to adopt new strategies to meet evolving customer expectations. Meanwhile, new entrants—commonly referred to as “FinTech” companies—have seized opportunities to deliver services that traditional banks were either unable or too slow to provide. These new players have introduced innovative solutions such as mobile wallets, payment applications, crowdfunding platforms, and automated investment advisory tools, fundamentally reshaping the financial landscape.

From a theoretical perspective, the relationship between technological development and productivity is anchored in classical and modern production theory, which posits that productivity increases when firms enhance their ability to generate more output from a given set of inputs. In the banking sector, this is achieved primarily through reductions in transaction costs, improved information processing, enhanced intermediation efficiency, and service innovation. The digital transformation introduces three core mechanisms that drive productivity gains in banking: (1) Automation and Process Efficiency, (2) Scalability and Cost Reduction, and (3) Complementarity with Intangible Capital. Moreover, drawing from the endogenous growth literature (e.g., [3]), innovation in financial services not only enhances firm-level productivity but also contributes to broader sectoral and macroeconomic growth through diffusion effects and competition-driven efficiency. In this context, the adoption of FinTech, software tools, and mobile banking technologies serves as a channel for technological spill overs that reinforce convergence toward the productivity frontier. Banks that successfully invest in these capabilities tend to outperform less digitally mature counterparts, reflecting persistent heterogeneity in absorptive capacity and innovation readiness.

As a result of digitalisation and business transformation, a new financial services market has emerged. FinTech companies have targeted every segment of banking, aiming to enhance specific components of traditional universal banking [4]. While the future of traditional financial service providers remains uncertain—especially with the shift towards Open Banking, which allows customers to easily switch or combine service providers—the integration of financial technology into the production and delivery of financial products is becoming increasingly crucial [5].

Purnomo [6] emphasises the growing importance of digital marketing in driving sales on e-commerce platforms. Techniques such as search engine optimisation (SEO), content marketing, social media engagement, paid advertising, and user experience optimisation are identified as key drivers of increased traffic, customer engagement, and sales conversions. These strategies are essential for businesses seeking to maintain a competitive edge in the digital economy. The study by Purnomo [6] also highlights the need for innovative digital marketing approaches to enhance the visibility and sales of products and services in today’s rapidly evolving marketplace.

In the case of the United Kingdom (UK), the technology sector is experiencing significant growth. Martin Carkett, policy lead at the Tony Blair Institute for Global Change, noted in 2021 that “the technology industry is expanding more than two and a half times faster than the rest of the British economy,” with 2.1 million jobs created in the digital economy in 2018. The UK digital sector contributed an estimated GBP £149 billion—or GBP £400 million per day—to the national economy in 2018, marking a 7.9% increase from the previous year [7].

Given this context, the current study seeks to explore the impact of technological development on productivity within the UK private sector. Specifically, it aims to assess how digitalisation, e-banking, mobile banking, and emerging financial technologies are shaping the evolving banking landscape in the UK and influencing the productivity of financial institutions.

The empirical results of this study reveal that digital adoption, FinTech investment, and digital skills are positively associated with productivity growth across UK banks. While the anticipated complementarities between digital skills and FinTech adoption were not consistently statistically significant, the main effects of both digital technologies and human capital investments were robustly positive. FinTech adoption was also linked to improved operational efficiency and notable shifts in customer behaviour, including increased use of digital channels. These findings highlight the critical role of digital transformation in enhancing bank performance and accelerating sector-wide innovation.

While prior studies have largely focused on descriptive accounts of technological adoption in banking or explored its impact on market structure and innovation, this study advances the literature by providing a rigorous empirical analysis of the productivity effects of digital transformation in UK banks. Using detailed firm-level data on digital adoption, FinTech investments, and digital skills, this paper employs advanced panel econometric techniques to isolate the direct contribution of technological development to operational efficiency and productivity growth. This granular approach allows for a better understanding of heterogeneity in banks’ innovation readiness and the interaction between human capital and technology—an area that remains under examined in existing research. Thus, the study makes both a methodological and empirical contribution by quantifying how digitalisation shapes productivity dynamics in a key financial sector.

Although extensive research has documented the rise in digital technologies in banking, much of the existing literature remains descriptive or limited to case studies and broad industry overviews. This study advances beyond previous work by quantitatively assessing the productivity impact of digital transformation within the UK banking sector using firm-level data. Unlike prior studies that often rely on aggregated proxies or binary measures of digital adoption, our approach incorporates detailed indicators of FinTech investment, digital skills, and technology usage intensity. Furthermore, we employ several robust econometric methods to control for unobserved heterogeneity and endogeneity, providing more reliable estimates of causal effects. By doing so, this research contributes novel empirical evidence of the heterogeneous effects of banking digitalisation on productivity and operational efficiency, offering actionable insights for policymakers and industry practitioners.

It is important to acknowledge that measuring technological development and its impact on productivity involves inherent challenges. Digital adoption and FinTech investment are complex, multifaceted phenomena that are not easily captured by single indicators. Moreover, isolating the causal effect of technology on productivity is complicated by potential endogeneity issues, such as reverse causality and omitted variable bias. This study addresses these concerns by employing robust panel data techniques and controlling for key confounding factors; however, some limitations remain, which we discuss in detail in the methodology and robustness sections.

This paper proceeds as follows. Section 2 contains an overview of the relevant literature. The structure of the UK banking sector is highlighted in Section 3. The empirical methodology and exploited variables are illustrated in Section 4. Section 5 contains the dataset and the source of data. The empirical results and the robustness tests are included in Section 6. Finally, the conclusion and policy recommendations are presented in Section 7.

2. Literature Review and Hypothesis Development

The digital transformation has reshaped virtually every sector, and the financial industry is no exception. One of the most significant manifestations of this transformation is the rise in FinTech (financial technology), which represents a major innovation in the financial services landscape. FinTech encompasses a wide array of digital financial services—ranging from mobile payments, peer-to-peer lending, and digital wallets to online trading, robo-advisory services, and crypto-assets ([8,9]). These innovations have been accelerated by advances in information technology, regulatory support, and shifting consumer behaviour, but scholarly analysis of FinTech remains an evolving field [10].

The theoretical foundation for understanding the impact of financial technologies on banking productivity is grounded in classical and modern productivity theory, as well as technology adoption frameworks. According to production theory, productivity improvements occur when firms generate more output from a given set of inputs, often through enhanced efficiency, innovation, and cost reduction. In the banking sector, technology adoption reduces transaction costs, improves information processing, and facilitates service innovation, all of which contribute to productivity gains. Technology adoption models, such as the Technology Acceptance Model (TAM) and the Diffusion of Innovations theory, emphasise factors influencing the uptake of new technologies at the organisational level, including perceived usefulness, ease of use, and organisational readiness. By combining these theoretical perspectives, this study conceptualises digital transformation in banking as a critical driver of productivity growth, mediated by the extent of technological integration and complementary human capital investments.

Productivity measurement in banking has long been a topic of scholarly interest, given the sector’s unique characteristics—such as the intangibility of outputs, regulatory constraints, and the dual role of banks as intermediaries and service providers. Traditional approaches have relied on financial ratios (e.g., cost-to-income, Return on Assets) as proxies for operational performance, while more rigorous methods include non-parametric techniques like Data Envelopment Analysis (DEA) and parametric approaches such as Stochastic Frontier Analysis (SFA) to estimate efficiency and productivity levels ([11,12]). More recent studies have incorporated Total Factor Productivity (TFP) indices, capturing both input and output growth dynamics over time [13]. However, many of these approaches overlook the role of technological change as a determinant of productivity. This study contributes to the literature by integrating digital adoption indicators with traditional productivity metrics, offering a more comprehensive view of how technology reshapes efficiency and performance in modern banking.

In the aftermath of the 2008 Global Financial Crisis (GFC), public trust in traditional banks eroded, opening space for alternative financial models. Simultaneously, consumers—especially younger generations—have become more tech-savvy, demanding seamless, digital-first financial services [14]. Regulatory authorities have also adapted, with frameworks such as the PSD2 and Open Banking initiatives enabling third-party providers to offer financial services, thus intensifying competition ([15,16]). The result is what Wu et al. described as the “FinTech Revolution,” which is pushing incumbents to adapt their business models to remain competitive [17].

Alt et al. [18] identify two defining characteristics of this post-GFC era: (1) the convergence of multiple digital technologies, and (2) a consumer-centric innovation approach. These trends pose both threats and opportunities for incumbent financial institutions. Banks risk losing market share and brand equity but may also benefit by integrating FinTech solutions to improve customer experience and operational efficiency [19]. The following subsections explore the primary drivers, benefits, and risks of FinTech development.

2.1. Key Drivers for Financial Innovation

2.1.1. Changing Customer Expectations and Behaviour

Customer expectations have changed dramatically, driven by the rise in mobile and digital channels. Today’s clients demand fast, 24/7 access to financial services with intuitive user interfaces ([20,21]). APIs (Application Programming Interfaces) and Open Banking frameworks allow for seamless service integration, enabling customers to switch providers easily and encouraging competition [21].

Traditional banks historically benefited from customer loyalty built on trust [22], but FinTech has disrupted this relationship. As firms shift from product-centric to customer-centric strategies, customer experience becomes the main competitive edge [23]. According to Deloitte (2016), 93% of financial firms implementing digital strategies aimed to enhance client experience.

2.1.2. Concerns About Profitability

Profitability pressures in a low-interest rate environment have driven banks to rethink cost structures and improve operational efficiency. McKinsey’s 2025 analysis emphasises that digital transformation—not short-term cost cutting—is the sustainable solution. Automation, AI, and cloud computing reduce overhead and improve flexibility ([24,25]).

2.1.3. Regulatory Changes

Regulatory evolution has supported FinTech growth while ensuring consumer protection. PSD2, for example, fosters competition by requiring banks to share customer data (with consent) with third-party providers. GDPR governs data protection—a critical aspect as financial data becomes a valuable asset ([26,27]).

2.1.4. Increased Competition

FinTech start-ups enjoy several advantages over traditional banks: agility, lower operating costs, scalability, and data-driven models. These firms challenge incumbents across payments, lending, and wealth management sectors ([28,29]). According to Paulet & Mavoori [30], competition from FinTechs has already forced legacy banks to transform their business models.

2.2. Potential Benefits and Risks of FinTech

In his 2017 speech at the G20, Mark Carney warned that while FinTech holds transformative potential, it also poses systemic risks if not properly regulated. He referenced the “Hype Cycle,” warning against inflated expectations followed by disillusionment [31].

The UK’s Department for International Trade projected FinTech employment growth from 76,500 in 2018 to 105,500 by 2030. This indicates its potential as a job creator and innovation driver. The 2018 UK Crypto-assets Taskforce report emphasised the promise of DLT (Distributed Ledger Technology) in streamlining payment systems but also warned of new risks like market manipulation, consumer harm, and cybercrime.

FinTech also raises concerns about digital exclusion, algorithmic discrimination, and systemic risks associated with technological failure. The 2018 VISA outage illustrated how technological dependence can disrupt millions of consumers. Cybersecurity threats and data privacy breaches are further challenges requiring strong regulatory oversight ([20,31]).

Despite a growing body of research on the impact of digitalisation on bank performance, empirical studies in this area face substantial identification challenges. One key issue is reverse causality: more productive or better-performing banks may be more likely to adopt advanced digital technologies, rather than digitalisation causing improved performance. Moreover, omitted variable bias—such as unobserved managerial quality, regulatory incentives, or internal governance structures—may confound observed relationships. Measurement errors in capturing the scope and intensity of digital adoption also add noise to empirical estimations. To mitigate these concerns, recent studies have employed methods such as dynamic panel models, instrumental variables, and difference-in-differences designs (e.g., ([32,33])). These identification challenges underscore the importance of robust empirical strategies when assessing the causal link between technology and productivity in banking.

Furthermore, the empirical studies examining the impact of technology adoption on productivity face significant endogeneity challenges. These arise due to potential reverse causality—where more productive firms may be more likely to adopt new technologies—and omitted variable bias stemming from unobserved factors such as managerial quality or firm culture. To address these issues, recent research employs advanced econometric techniques including instrumental variables, fixed effects models, and dynamic panel methods to isolate the causal impact of technology on productivity. This study adopts similar approaches to mitigate endogeneity concerns and enhance the validity of the estimated relationships.

Finally, to ensure that the proposed hypotheses are firmly grounded in theory, we draw upon the resource-based view (RBV) and the theory of technological diffusion. According to the RBV ([34,35]) banks can gain competitive advantage and improve their performance by deploying unique technological capabilities. Furthermore, the technology adoption lifecycle [36] and productivity theory [37] suggest that digital transformation should yield measurable improvements in banking productivity through enhanced process efficiency and service innovation.

Based on this foundation, we expect that banks with higher levels of digital adoption will exhibit significantly greater productivity gains, especially in environments characterised by strong institutional support and managerial capability. Consequently, our hypotheses are reformulated to explicitly reflect these theoretical expectations, enabling more precise empirical testing. Consequently, drawing on the resource-based view [38] and technology diffusion theory [39], we propose the following hypotheses:

• H1: There is a significant relationship between productivity and financial technologies.
• H2: There is a significant relationship between financial technologies and the efficiency of banks.
• H3: There is a significant relationship between adopting financial technologies and shifts in customer behaviour.

3. The Structure of Banks in the United Kingdom

The United Kingdom remains a cornerstone of global finance, with London continuing to serve as a vital international hub for banking, capital markets, insurance, and financial technology (FinTech). The UK’s financial system is one of the most diversified and internationally integrated in the world, reflecting a long history of liberal financial policies, innovation, and global outreach. As of 2024, HSBC Holdings PLC, headquartered in London, retained its status as the largest bank in Europe by total assets—valued at approximately EUR €2.641 trillion—demonstrating the scale and global footprint of UK-based financial institutions [25].

3.1. The UK as a Global Financial Centre

London consistently ranks among the top two financial centres worldwide, second only to New York in the Global Financial Centres Index 37 [25], which reflects metrics such as business environment, human capital, infrastructure, and financial sector development [25]. The UK’s financial sector also contributes disproportionately to the national economy, accounting for approximately 8.3% of GDP and over 2.5 million jobs, both directly and indirectly [31]. Notably, the UK exported GBP £139 billion worth of financial and related professional services in 2023, resulting in a trade surplus of GBP £100.7 billion (USD $125.3 billion)—the largest of any country globally [31].

This success is underpinned by a supportive regulatory regime, highly skilled workforce, and openness to international financial firms. The UK holds a dominant global share in key areas such as foreign exchange trading (43%), interest rate derivatives (over 50%), cross-border lending, and FinTech investment ([32,33]).

The institutional environment of the UK banking sector provides a distinctive context for studying the productivity effects of digital adoption. As a mature, innovation-driven financial system with a strong regulatory framework, the UK combines high levels of technology diffusion with stringent oversight mechanisms. The Financial Conduct Authority (FCA) and the Prudential Regulation Authority (PRA) have fostered an environment conducive to controlled innovation, including the launch of regulatory sandboxes and Open Banking frameworks. Moreover, the UK’s competitive retail banking landscape creates pressures to innovate in order to improve operational efficiency and customer experience. In fact, institutional dynamics are not peripheral but are central to understanding how and why technological investments yield performance gains in this setting.

3.2. Digitalisation and Payment Trends

The UK’s financial infrastructure is also highly digitised. Payment systems are rapidly evolving, with contactless and digital payments becoming the norm. In December 2024 alone, 2.21 billion debit and credit card transactions were recorded, a 1.2% year-on-year increase, reflecting resilient consumer activity. Of these, 75% of debit and 65% of credit card transactions were contactless, indicating widespread adoption of digital payment methods [26]. The total number of issued debit cards stood at 102 million by October 2023, with 94 million being contactless-enabled [26]. Credit card penetration is also high, with over two-thirds of UK adults owning credit cards, and usage particularly concentrated among older age groups [27].

The UK’s well-established and competitive retail banking market has embraced technological change through mobile banking apps, AI-driven services, and Open Banking frameworks. Over 9 million UK consumers were using Open Banking-enabled services by late 2024, significantly enhancing competition and personalisation in financial services [38].

4. Empirical Methodology

This study employed a semi-inductive methodology to incorporate and analyse the existing literature on FinTech and technological advancements in the United Kingdom. The research process began with a bibliometric analysis aimed at identifying relevant sources and mapping the broader knowledge domain through a quantitative examination of authorship patterns, references, and citation networks [39]. Following the methodological framework proposed by Samiee & Chabowski [40], this study commenced with the identification of key terms. Drawing from the established literature on financial applications (e.g., Mondal & Chakrabarti, [41]), a list of keywords was compiled, including: “FinTech(s),” “mobile app(s),” “mobile phone banking(s),” “internet banking(s),” and “cryptocurrency(s)” in the context of the United Kingdom.

Additionally, this study employs Total Factor Productivity (TFP) and labour productivity as primary indicators of bank-level productivity. TFP is widely regarded as a comprehensive measure that captures efficiency improvements not explained by input accumulation, making it well-suited for analysing the effects of technology adoption (e.g., [42]). In the banking context, TFP reflects the ability of financial institutions to enhance output through innovations in service delivery, digital platforms, or back-office processes [43]. Labour productivity, defined as output per employee, provides a complementary measure that is particularly sensitive to digitisation and automation trends, which are central to FinTech integration. Together, these measures capture both broad efficiency gains and workforce-level productivity shifts, aligning with the theoretical frameworks that associate digital transformation with enhanced resource allocation, scale economies, and cost reduction in banking ([44,45]).

Potential measurement error is an inherent concern in empirical studies relying on constructed indicators, particularly when using text-mining techniques. In our case, the FinTech adoption variable—derived from textual analysis of annual reports—may be subject to reporting bias, inconsistent terminology across banks, or variation in disclosure practices. To mitigate this, we applied standardised keyword filtering and natural language processing (NLP) protocols, validated through manual audits and benchmarked against objective metrics. Additionally, productivity measures such as Total Factor Productivity (TFP) and labour productivity may contain estimation errors due to assumptions in input-output modelling and aggregation. However, we use data from official and audited financial disclosures to minimise inconsistencies. To address any remaining concerns, we rely on instrumental variable (IV) estimation, which helps correct for attenuation bias associated with measurement error in the FinTech adoption variable.

All relevant sources were retrieved from the Scopus database, with a focus on publications from the past two decades. This timeframe was chosen to encompass both foundational studies and more contemporary research, and the search protocol followed procedures similar to those used in prior studies (e.g., [41]). In the second phase of the review process, the collected bibliometric sources underwent a structured screening to identify recurring themes and well-established findings within the literature.

Consistent with current best practices for producing rigorous and insightful literature reviews, this phase also involved the discovery and analysis of additional relevant sources not captured in the initial bibliometric frame [46]. The integration of both bibliometric mapping and thematic synthesis ensured a comprehensive understanding of the evolving research landscape surrounding FinTech and digital innovation in the UK.

4.1. Variable Specifications

This section outlines the key variables used to investigate the impact of financial technologies on banking productivity in the United Kingdom. As the UK banking industry undergoes digital transformation, understanding the implications of new technologies—such as internet and mobile banking, as well as emerging tools like cryptocurrencies—on bank performance is essential. The data sources include bank annual reports, government and private databases, and consumer behaviour surveys. This study utilises five main dependent variables based on the CAMEL framework to measure banking productivity and introduces three core independent variables related to FinTech adoption: Bank FinTech, mobile banking, and cryptocurrency usage.

4.1.1. Banking Productivity

Banking productivity is operationalised using the CAMEL rating system, a widely accepted framework in banking performance assessment. The five components used as dependent variables are the following:

• Capital Adequacy Ratio (CAR): This metric reflects a bank’s capacity to absorb potential losses and sustain operational resilience. It is calculated as the ratio of total capital to total assets. A higher CAR indicates stronger capitalisation and greater financial stability, reducing the risk of insolvency.
• Asset Quality (NPL Ratio): Measured as the ratio of non-performing loans (NPL) to total loans, this indicator captures the bank’s credit risk. A high NPL ratio signals deteriorating loan quality and potential exposure to default, which adversely affects productivity and financial health.
• Management Efficiency (EFF): Efficiency is proxied by the ratio of non-interest expenses to net operating income (excluding provisions and interest expenses). This variable evaluates how effectively bank management controls costs. A higher ratio denotes lower efficiency, potentially eroding profitability and operational sustainability.
• Earnings Power (ROA): Return on Assets (ROA) is used to assess profitability, calculated as income before extraordinary items divided by total assets. ROA provides insight into how efficiently a bank is utilising its assets to generate earnings.
• Liquidity Management (LIQ): Defined as the ratio of total loans to total deposits, this indicator reflects a bank’s ability to meet short-term obligations. A lower loan-to-deposit ratio suggests prudent liquidity management and lower leverage risk.

Each of these variables is employed separately as a dependent variable in five different model specifications to assess how financial technologies impact different dimensions of bank productivity.

4.1.2. Bank FinTech (Internet Banking Initiatives)

Bank FinTech, the first of the independent variables, captures the extent to which UK banks integrate digital technologies into their business models. The growing reliance on internet banking—from basic account management to complex financial transactions—illustrates the sector’s increasing digital maturity.

To quantify FinTech adoption, this study analyses the annual strategic reports of major UK banks, which typically detail business strategies and digital innovation initiatives. Python 3.9 was employed to perform textual analysis using a FinTech-specific keyword list adapted from Cheng & Qu [47], comprising 70 terms related to emerging technologies (e.g., artificial intelligence, block chain, Open Banking, APIs, RegTech, and digital wallets). The frequency of these terms was normalised to the total word count and expressed as a percentage, reflecting the annual emphasis on digital strategy in bank narratives.

As for the validation and robustness, to ensure the measure’s validity, we benchmarked the textual analysis against objective FinTech adoption metrics, including the following: (1) IT expenditure data from bank financial disclosures, (2) regulatory compliance indicators (e.g., PSD2 implementation status), (3) customer-facing adoption rates (e.g., mobile banking users from UK Finance reports). Strong correlations between keyword frequencies and these external metrics confirm that the textual measure accurately reflects actual FinTech integration.

Further, we addressed potential measurement error through an instrumental variable (IV) approach, using sectoral digital adoption trends (lagged and leave-one-out industry averages) as exogenous instruments. Then, robustness checks, such as substituting textual scores with direct IT investment data or excluding banks with atypical reporting, yielded consistent results.

This longitudinal measure provides an objective gauge of FinTech’s strategic embeddedness, offering insights into how digital adoption drives innovation-led productivity gains. While the focus on large UK banks may limit generalisability, the measure’s alignment with third-party indices (e.g., Deloitte’s FinTech rankings) underscores its credibility.

4.1.3. Mobile Banking Adoption

Mobile banking represents a distinct and fast-evolving channel of digital financial service delivery. This variable captures the extent to which mobile banking services are used and valued by consumers, as well as how effectively banks are meeting these demands.

This study draws upon the UK Mobile Banking Emerging Features Benchmark Report [31], which surveyed 1172 UK adults aged 18–76 who had used mobile banking in the prior three months. Based on this benchmark, six key drivers of mobile banking adoption were identified: (1) high security and user control; (2) real-time alerts and notifications; (3) flexible account management; (4) digital money management tools; (5) responsive customer service’ and (6) ease and speed of money transfers.

According to the report, mobile banking usage in the UK is rapidly growing. While adoption lagged behind some other countries initially, usage rates surged after 2019, with projections suggesting that 72% of UK adults would be using mobile banking apps by 2023. Additionally, all major UK banks now offer mobile and online services, with some digital-only challenger banks entering the market.

In this study, mobile banking is operationalised by combining qualitative indicators (feature availability across major banks) and quantitative survey data, offering a nuanced view of how mobile functionality translates into improved user experience and potentially higher productivity.

4.1.4. Usage of Cryptocurrency

The third independent variable, Usage of Cryptocurrency, captures the growing penetration of decentralised digital currencies within the UK financial ecosystem. While traditional banks were initially sceptical of cryptocurrencies due to regulatory concerns and perceived volatility, the increasing popularity of crypto-assets among consumers and businesses has begun to influence mainstream financial services.

To quantify this trend, this study relies on data from Triple A’s 2021 financial report, which estimates that around 4.2 million people in the UK—equivalent to 6.2% of the population—own some form of cryptocurrency. This represents a significant and growing segment of the population, indicating a shift in public trust and interest toward decentralised financial assets. The demographic profile of crypto owners is particularly notable: 56% are between 18 and 34 years old, reflecting the digital-native generation’s preference for innovation and non-traditional finance. A total of 40% have annual incomes above GBP £200,000, suggesting that crypto adoption is not limited to speculative retail investors but includes high-net-worth individuals seeking portfolio diversification. Finally, 46% hold at least a bachelor’s degree, indicating a relatively well-informed user base that may influence financial decision-making at both personal and institutional levels.

In terms of behavioural engagement, the report reveals that over 52% of crypto owners use their assets for purchases, and nearly 60% make at least one crypto-based transaction monthly. Notably, 26% of users are classified as high spenders, meaning they use cryptocurrency not merely as a speculative asset, but as a functional medium of exchange in the real economy.

This evolving usage pattern suggests that cryptocurrencies are transitioning from niche investments to mainstream payment instruments. Their growing presence in the financial landscape can have multiple implications for banking productivity. On one hand, cryptocurrencies may compete with traditional payment systems, potentially eroding banks’ fee-based income from cross-border transfers and card services. On the other hand, banks that integrate cryptocurrency services—such as custody solutions, crypto trading platforms, or block chain-based infrastructure—may enhance customer retention and technological competitiveness.

For the purpose of this study, the usage of cryptocurrency variable is constructed using a composite of three indicators: (1) Ownership Rate—proportion of UK residents who own cryptocurrencies; (2) Transactional Usage—frequency and volume of crypto-based purchases; and (3) User Demographics—age, income, and education profiles of crypto users.

These indicators are combined to reflect the societal adoption of cryptocurrency and its potential to reshape financial behaviour. As UK banks adapt to this emerging asset class—either by embracing it or resisting it—their performance and productivity are likely to be affected through innovation spill overs, regulatory pressures, and shifts in customer expectations.

4.2. Estimation Model

4.2.1. Basic Model

Based on the description of the variables above, the following empirical model is proposed:

$$\mathit{log}(1+{CAR}_{it})=\alpha +\beta {BankFinTech}_{it}+\gamma {Technology}_{it}+{\lambda }_t+{\epsilon }_{it}$$

(1)

where

• $\mathit{log}(1+{CAR}_{it})$ is the natural logarithm of one plus the capital-to-asset ratio, used as a proxy for changes in bank productivity in the United Kingdom banking sector at time t. Productivity is measured as the real value added relative to the usage of mobile banking and FinTech services. According to economic theory, banks that effectively adopt mobile and digital technologies such as online banking via mobile devices and cryptocurrency services—are expected to experience increased productivity.
• Bank FinTech and Technology represent vectors of explanatory variables capturing digital transformation. Specifically, these include measures of FinTech adoption, investment in digital infrastructure (e.g., software), and the application of digital skills by customers and staff. The coefficient β reflects the marginal effect of FinTech adoption, while γ captures the impact of broader technological factors on bank performance.
• The model also considers individual regressions for each type of digital technology, software investment, and digital skill level. Additionally, interaction terms are introduced to test for potential synergies among these factors. For instance, the use of digital tools often requires both software investment and digital capabilities, and the combined effect may exceed the sum of individual contributions, suggesting a positive interaction effect on productivity.
• α denotes the constant (intercept) term.
• The vector $\gamma$ captures the impact of technology-related explanatory variables, including digital infrastructure and skills. Separately, firm-level control variables—such as the logarithm of physical capital investment per worker, bank age, and size—are introduced to isolate the pure effect of technology from general bank characteristics.
• $\lambda$ accounts for year-fixed effects to capture macroeconomic trends or shocks common across all banks in a given year.
• ε is the error term capturing unobserved factors affecting productivity.

This model aims to empirically test the hypotheses that banks adopting digital financial technologies and investing in complementary software and skills experience significant gains in productivity.

4.2.2. Model Extension: Time-Varying Coefficient Model

Given the extended study period from 2007 to 2022, which includes major events such as the 2008 Financial Crisis, the rise in Open Banking, and the COVID-19 pandemic, it is possible that the effects of FinTech adoption, digital skills, and software investment on bank productivity have evolved over time. To capture this dynamic, an extension of the baseline model is proposed using time-varying coefficient models, which allow the marginal effects of key explanatory variables to change across different periods. This is particularly relevant in a context where digital infrastructure, customer behaviour, and regulatory environments have shifted significantly. The extended model can be written as:

$${Productivity}_{it}={\alpha }_i+{\beta }_t{FinTech}_{it}+{\gamma }_t{Skills}_{it}+{\lambda X}_{it}+{\delta }_t+{\epsilon }_{it}$$

(2)

where ${\beta }_t$ and ${\gamma }_t$ are time-varying coefficients capturing the evolving effects of FinTech adoption and digital skills on productivity. $X_{it}$ is a vector of control variables, including bank size, age, capital per worker, and other non-technology-related firm characteristics (with fixed coefficients $\gamma$). ${\delta }_t$ denotes year-fixed effects, accounting for macroeconomic shocks common across all banks. ${\alpha }_{i }$ captures bank-level fixed effects, controlling for time-invariant unobserved heterogeneity. Finally, ${\epsilon }_{it}$ is the idiosyncratic error-term.

Estimation techniques such as local polynomial regression, rolling windows, or kernel smoothing [48] can be applied to retrieve these coefficients, depending on the structure of the panel data. This approach allows the model to detect whether productivity benefits from digital adoption increased, declined, or shifted in timing during key periods like the post-2018 FinTech boom or during COVID-19.

Structural breaks in productivity dynamics—such as those triggered by financial crises or technological shocks—can fundamentally alter the relationship between digital adoption and bank performance [49]. Following Hansen’s framework, we test for breaks not only around predefined events (e.g., the 2008 GFC and COVID-19) but also allow for endogenously determined shifts in coefficients. This dual approach ensures robustness against misspecification of break dates, a critical concern when analysing financial sectors undergoing rapid digital disruption.

To assess the economic significance of our findings, we use a standardisation approach. For each key variable, we calculate the expected change in productivity resulting from a one-standard-deviation increase. This is performed by multiplying the estimated regression coefficient by the standard deviation of the variable. The simulation assumes ceteris paribus conditions and isolates the marginal impact of each factor.

This will occur by using (1) the Chow test [50] which will evaluate whether the coefficients of the regression model are stable across predefined subsamples (pre-GFC, post-GFC/pre-COVID-19, and post-COVID-19). A significant F-statistic indicates a structural break (2) Further we will use the Bai—Perron test [51], which will endogenously detect multiple breakpoints in the time series without prior specification of event dates, allowing us to validate the timing of structural changes.

This enhancement aligns with previous empirical studies (e.g., [17]; [52]), and provides more granular insights for both policymakers and industry stakeholders by revealing when digital investments yield the most significant productivity returns. It also supports the broader goal of this study: to not only quantify but also time-localise the productivity effects of financial innovation in the UK banking sector.

5. Sample and Data Collection

This study draws on data from leading public and commercial banks in the United Kingdom that had integrated FinTech into their business models during the period of 2007 to 2022. To capture the FinTech dimension, a text-mining approach is applied to financial reports collected for this period. These reports include data on the adoption and usage of digital technologies—such as mobile banking and cryptocurrency services—and allow for tracking price dynamics and changes in the top 100 cryptocurrencies used in the UK.

The UK was selected due to its pioneering role in digital banking and its position as a global FinTech hub, supported by a proactive regulatory environment and early adoption of financial technologies. The 2007–2022 period captures both the aftermath of the Global Financial Crisis and the rapid acceleration in digital transformation across the banking sector. Focusing on leading public and commercial banks allows for a more consistent and data-rich analysis, as these institutions are subject to comprehensive disclosure requirements and have been at the forefront of FinTech integration.

The selected banks are representative of the broader UK banking sector in terms of asset size, market share, and technological engagement. These institutions account for a significant proportion of total banking assets in the UK and lead the industry in digital service provision. By focusing on major public and commercial banks, this study captures the structural and technological dynamics that influence sector-wide productivity trends. Moreover, the inclusion of both traditional and digitally advanced banks ensures variation in FinTech integration, allowing for meaningful comparative analysis. While the sample excludes smaller or niche institutions, it reflects the segment of the banking industry most affected by—and influential in shaping—the evolution of financial technology in the UK.

During the study period, the UK banking sector experienced significant regulatory changes that influenced digital transformation. The post-2008 Global Financial Crisis reforms led to enhanced prudential regulations under Basel III, increasing capital and liquidity requirements. The introduction of the Payment Services Directive 2 (PSD2) in 2018 mandated Open Banking, requiring banks to share customer data securely with authorised third parties, thus fostering competition and innovation. Additionally, the Financial Conduct Authority’s (FCA) focus on promoting Fintech innovation through regulatory sandboxes (from 2016) encouraged experimentation with new technologies. These regulatory milestones played a critical role in shaping banks’ incentives and capabilities to adopt digital technologies, which this study accounts for in its empirical framework.

Bank-level indicators were obtained from the UK Office for National Statistics (ONS). Banks were ranked based on total assets reported in their financial statements, consistent with common academic practice. Total assets are widely used as a proxy for bank size in the literature [23,24,25,26,27,28,29,30,31,32,33,34,35,36,37,38,39,40,41,42,43,44]. FinTech-related data were primarily sourced from annual reports, accessible via the Statista database and the official websites of the respective banks.

This study follows empirical approaches from the literature on the impact of technological advancements on productivity in the UK private sector (e.g., Acemoglu & Restrepo, [53]; Cette et al., [42]) and employs an instrumental variable (IV) strategy to address potential endogeneity concerns. Specifically, the model uses an exogenous instrument for bank-level digital adoption and intangible investment indicators.

The use of an instrumental variable approach in this study is essential to address potential endogeneity arising from reverse causality and omitted variable bias between digital adoption and productivity. Our instrument—sectoral exposure to digitalisation—is constructed using lagged bank-level digital adoption and intangible investment relative to industry-wide averages. This approach is grounded in recent empirical work (e.g., [52,54]), which demonstrates that lagged sectoral trends serve as valid exogenous instruments when firm-level adoption is potentially endogenous. The core identification assumption is that sectoral digitalisation trends influence a bank’s productivity only through their effect on digital engagement, and not through other unobserved productivity shocks.

To validate the credibility of the instrumental variable strategy, we rely on two key identification assumptions: relevance and exclusion. First, the relevance condition is satisfied as sector-level digital adoption and intangible intensity—calculated using lagged values and leave-one-out sectoral averages—are strong predictors of individual bank-level digital engagement. These variables capture exogenous exposure to technology spill overs, which vary systematically across banks based on their sectoral positioning. Second, the exclusion restriction assumes that sector-level digital intensity affects bank-level productivity only through its effect on digital adoption and intangible investment, and not through other unobserved channels. This assumption is supported by the leave-one-out construction, which excludes the focal bank’s own behaviour from the sectoral aggregates, thereby mitigating reflection bias and mechanical correlation. Furthermore, no single bank is large enough to influence sector-wide trends, reducing the risk of endogenous feedback effects. In line with empirical studies using sectoral instruments in productivity research (e.g., [53,55]) we argue that these assumptions are plausible within the UK banking context, where regulatory, technological, and organisational shifts tend to affect banks broadly and independently of any single institution’s internal dynamics.

The instrument is based on sectoral exposure—calculated using lagged bank-level digital adoption and intangible intensity, as well as the sector-wide means of these variables. This exposure captures the extent to which a bank’s digital engagement deviates from industry norms, accounting for temporal lags. The rationale is that banks adopting digital technologies more intensively than the industry average are more exposed to sector-wide spill overs—such as falling adoption costs or learning effects—thereby benefiting more from broader technological progress. Conversely, banks with below-average digital engagement are less exposed to these benefits.

A key identifying assumption is that these sector-level technology trends are exogenous to individual banks and affect productivity only through industry-wide spill over effects. No single bank is large enough to influence sectoral averages significantly. However, given the risk of concentration effects—where a few dominant digital-intensive banks could skew sectoral measures—this study applies a “leave-one-out” methodology. This approach excludes each bank’s own contribution when calculating sector-wide digital adoption and intangible intensity, mitigating bias from firm-level influence. The IV approach addresses both omitted variable bias and reverse causality, strengthening the robustness of the empirical strategy.

Table 1. Variables descriptive statistics.

Variable	Mean	Median	Bottom Decile	Top Decile	Std. Dev.	95% CI
Dependent Variables
Productivity growth (log)	0	0.012	−0.301	0.301	0.412	[−0.788, 0.788]
Multifactor productivity growth (log)	0.021	0.011	−0.241	0.262	0.313	[−0.593, 0.635]
Control Variables
Frontier growth (log)	0.021	0.033	−0.222	0.252	0.301	[−0.569, 0.611]
Gap to frontier (log)	0.912	0.833	0.193	1.711	0.712	[−0.484, 2.308]
Capital per worker (log)	5.261	5.433	1.601	8.443	2.612	[0.140, 10.382]
Bank age in years (log)	1.382	1.302	0.704	1.672	1.253	[−1.074, 3.838]
Bank size in employees (log)	4.541	4.401	3.132	6.161	1.202	[2.185, 6.897]
Tech-FinTech skill shortage	0.162	0.124	0.034	0.351	0.141	[−0.114, 0.438]
Digitalisation and Intangibles
Technology Infrastructure
Technology Development investment	0.212	0.101	0.011	0.612	0.253	[−0.284, 0.708]
High-speed broadband dummy	0.693	1.001	0	1.001	0.463	[−0.214, 1.600]
Intangible Skills
Technology R&D investment	0.231	0.114	0.014	0.695	0.272	[−0.302, 0.764]
Technological specialist skills	0.051	0.012	0.002	0.082	0.161	[−0.264, 0.366]
Software specialist skills	0.021	0.001	0.001	0.041	0.072	[−0.120, 0.162]
Computer use at work	0.732	0.903	0.202	1.001	0.323	[0.099, 1.365]
Training dummy	0.462	0	0	1.003	0.502	[−0.522, 1.446]
Digital Technologies
CRM/ERP dummy	0.832	1.004	0	1.002	0.381	[0.085, 1.579]
Cloud computing dummy	0.502	0	0	1.004	0.502	[−0.482, 1.486]
Big data analytics dummy	0.363	0	0	1.004	0.482	[−0.581, 1.307]
Sectoral Variable
Financial Services dummy	0.665	1.001	0	1	0.472	[−0.260, 1.590]

Notes: Based on an unbalanced panel containing the six largest banks in the United Kingdom over the period 2007 to 2022. Source: Authors’ computations.

presents some summary statistics of the included variables. Specifically, productivity growth is measured as the log-difference in output per worker, reflecting changes in labour productivity over time. Multifactor productivity (MFP) growth captures the portion of output growth not accounted for by changes in inputs (labour and capital), representing efficiency gains from intangible investment, innovation, or digital transformation. Gap to frontier measures the log difference between each bank’s productivity level and the sectoral productivity frontier, and captures the relative distance to the best-performing banks and is used in our extended regression model (Equation (2)) to analyse catch-up dynamics. These variables are essential to understanding the variations in performance across banks and are directly incorporated into our regression framework to estimate the effects of digital adoption, FinTech investments, and control variables on bank productivity.

Table 2. Correlation matrix of key variables.

Variables	Productivity Growth	FinTech Adoption	Software Investment	Broadband Access	IT/FinTech Skills	Software Skills	Capital per Worker	Bank Size (log)
Productivity Growth	1.00
FinTech Adoption	0.42	1.00
Software Investment	0.38	0.54	1.00
Broadband Access	0.27	0.36	0.41	1.00
IT/FinTech Skills	0.51	0.46	0.48	0.33	1.00
Software Skills	0.34	0.39	0.47	0.26	0.58	1.00
Capital per Worker	0.22	0.29	0.35	0.31	0.30	0.27	1.00
Bank Size (log)	0.19	0.25	0.30	0.28	0.26	0.23	0.31	1.00

Notes: Values shown are Pearson correlation coefficients. Source: Authors’computations.

includes the pairwise Pearson correlation coefficients among the key variables used in the empirical analysis. This matrix provides initial insights into the strength and direction of linear associations between FinTech adoption, intangible assets, digital infrastructure, and productivity growth in UK banks. As expected, productivity growth shows moderate to strong positive correlations with variables capturing digital transformation. Notably, FinTech adoption (0.42), IT/FinTech specialist skills (0.51), and software investment (0.38) are positively correlated with productivity growth, suggesting that greater digital engagement is associated with better performance outcomes. Similarly, software specialist skills (0.34) and broadband access (0.27) also demonstrate meaningful positive associations, indicating the complementary role of digital infrastructure and workforce capabilities.

Table 2 also reveals substantial interrelationships among the digital and intangible variables themselves. For example, the correlation between software investment and FinTech adoption is relatively strong (0.54), reflecting that institutions investing in software tend to be more actively involved in broader digital transformation strategies. IT/FinTech skills also correlate closely with software and broadband indicators, reinforcing the idea that digital capacity building is multifaceted, involving both technical tools and human capital development.

None of the correlations exceed the conventional multicollinearity threshold (0.80), suggesting that the key explanatory variables do not exhibit problematic overlap. This supports the reliability of the regression estimates presented in subsequent sections. Overall, the correlation matrix reinforces the theoretical expectation that digital capabilities and intangible investment are critical complements to productivity enhancement in modern banking institutions.

6. Empirical Results

6.1. Drivers of Productivity Growth in UK Banks: The Role of Digital Adoption, Skills, and Technological Infrastructure

Table 3. Impact of digital and intangible assets on UK bank productivity growth.

	Dependent Variable: Productivity Growth
	(1) ICT Hardware	(2) Broadband >30 Mbit/s	(3) Software Invest.	(4) Tech Skills	(5) Software Skills	(6) Computer Use	(7) Training	(8) CRM/ERP	(9) Cloud Computing	(10) Big Data
Constant	−0.301 ***	−0.284 ***	−0.297 ***	−0.271 ***	−0.268 ***	−0.317 ***	−0.273 ***	−0.275 ***	−0.271 ***	−0.302 ***
Frontier growth	0.162 ***	0.165 ***	0.166 ***	0.158 ***	0.158 ***	0.169 ***	0.160 **	0.162 ***	0.173 ***	0.168 ***
95% CI	[0.142, 0.182]	[0.145, 0.185]	[0.146, 0.186]	[0.138, 0.178]	[0.138, 0.178]	[0.149, 0.189]	[0.140, 0.180]	[0.142, 0.182]	[0.153, 0.193]	[0.148, 0.188]
Gap to frontier (lagged)	0.251 ***	0.245 ***	0.248 ***	0.251 ***	0.251 ***	0.252 ***	0.254 ***	0.248 ***	0.264 ***	0.248 ***
95% CI	[0.231, 0.271]	[0.225, 0.265]	[0.228, 0.268]	[0.231, 0.271]	[0.231, 0.271]	[0.232, 0.272]	[0.234, 0.274]	[0.228, 0.268]	[0.244, 0.284]	[0.228, 0.268]
Capital ratio	0.024 ***	0.014 ***	0.017 ***	0.015 ***	0.015 ***	0.015 ***	0.014 ***	0.015 ***	0.017 ***	0.016 ***
95% CI	[0.020, 0.028]	[0.010, 0.018]	[0.013, 0.021]	[0.011, 0.019]	[0.011, 0.019]	[0.011, 0.019]	[0.010, 0.018]	[0.011, 0.019]	[0.013, 0.021]	[0.012, 0.020]
Bank age	0.001	−0.002	−0.001	0	0	−0.001	−0.004	−0.001	−0.003	0.008
95% CI	[−0.003, 0.005]	[−0.006, 0.002]	[−0.005, 0.003]	[−0.004, 0.004]	[−0.004, 0.004]	[−0.005, 0.003]	[−0.008, 0.000]	[−0.005, 0.003]	[−0.007, 0.001]	[0.004, 0.012]
Digital and intangible	0.129 ***	0.030 ***	0.038 *	0.084 **	0.049 *	0.065 **	0.033 **	−0.003	0.007	0.005
95% CI	[0.109, 0.149]	[0.010, 0.050]	[0.018, 0.058]	[0.064, 0.104]	[0.029, 0.069]	[0.045, 0.085]	[0.013, 0.053]	[−0.023, 0.017]	[−0.013, 0.027]	[−0.015, 0.025]
Constant	−0.301 ***	−0.284 ***	−0.297 ***	−0.271 ***	−0.268 ***	−0.317 ***	−0.273 ***	−0.275 ***	−0.271 ***	−0.302 ***
95% CI	[−0.321, −0.281]	[−0.304, −0.264]	[−0.317, −0.277]	[−0.291, −0.251]	[−0.288, −0.248]	[−0.337, −0.297]	[−0.293, −0.253]	[−0.295, −0.255]	[−0.291, −0.251]	[−0.322, −0.282]
Time FE	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes
Sector FE	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes
Observations	2872	4721	2926	2087	4725	4733	2100	2087	4736	1682
R-squared	0.151	0.147	0.144	0.149	0.148	0.149	0.15	0.144	0.171	0.144

Notes: * 10%, *** and ** denote statistically different from zero at the 1% and 5% levels, respectively. CI = confidence interval. All regressions include time and banking sector fixed effects to control for unobserved heterogeneity across years and institutions, following a Hausman test (χ² = 12.34, p = 0.002). Diagnostic tests for autocorrelation (Wooldridge F = 9.87) and heteroskedasticity (Wald χ² = 24.56) confirm robust SEs are appropriate. Ramsey RESET test (F = 1.23, p = 0.27) showing no omitted variable bias. Source: Authors’computations.

presents the findings from the Ordinary Least Squares (OLS) regression applied to the full sample of banks operating in the United Kingdom, using Equation (1), with a focus on explaining variations in productivity growth. The analysis highlights robust and statistically significant positive associations between productivity growth and several variables that capture digital adoption and the use of intangible assets.

The results suggest that investments in information technology (IT) and technological infrastructure—such as FinTech platforms and high-speed broadband—are significantly and positively related to increases in productivity. These findings may underscore the importance of foundational technological capabilities in enhancing bank performance. While such investments may not constitute intangible assets per se, nor are they strictly classified as digital technologies, they are nonetheless critical prerequisites for enabling the adoption and effective utilisation of advanced digital tools and intangible resources, such as proprietary software and cloud-based systems.

Moreover, the utilisation of digital skills—used here as a proxy for intangible capital—exhibits a consistent and significant positive relationship with productivity growth across multiple specifications. Digital skills may enable employees to effectively engage with and extract value from technological innovations, thus driving productivity improvements. Among the various dimensions of digital skills examined, those relating to specialists in information technology and financial technology (FinTech) are most strongly associated with higher productivity levels. These skillsets appear to offer the greatest potential for leveraging digital transformation in banking.

Interestingly, the only skill category that does not yield statistically significant results is expertise in financial software. Although this variable has a large positive coefficient, its lack of significance is likely attributable to the small proportion (approximately 2%) of employees in the sample who possess such expertise, thus limiting statistical power. Nevertheless, the overall pattern across skill categories reinforces the view that both general and specialised digital capabilities are instrumental in driving productivity gains in the sector.

Further, training in IT and technological development, as well as the use of digital tools such as computers, mobile devices, and other modern technologies in the workplace, are all positively and significantly associated with productivity growth. These findings could support the argument that digital training and tool usage contribute to more efficient operations and better service delivery, leading to higher overall productivity.

Turning to the control variables, most show the expected signs and are statistically significant across the majority of model specifications. Consistent with theoretical expectations, a 1% increase in frontier productivity growth is associated with a less-than-proportional increase in bank-level productivity. This may suggest that while frontier expansion contributes positively to productivity, the benefits are subject to diminishing returns. Importantly, it highlights the role of technology spill overs from leading-edge firms to others in the market, particularly non-frontier or follower banks.

In addition, the lagged productivity gap with respect to the frontier shows a strong and significant positive relationship with productivity growth. This implies that banks further behind the productivity frontier tend to experience faster productivity improvements over time, suggesting a process of catch-up conditional on their continued operation. Notably, the magnitude of the coefficient on the lagged productivity gap consistently exceeds that of frontier growth, indicating that convergence dynamics are at play. This pattern aligns with recent empirical evidence on UK banks, particularly the findings by Berlingieri et al. [56], which document significant progress among laggard banks in adopting and diffusing new technologies.

Moreover, investments in physical capital (measured through indicators of capital intensity) are also positively and significantly associated with productivity growth. This finding is consistent with the notion that more capital-intensive banks are better equipped to adopt new technologies and modernise their operations, resulting in higher productivity per employee. It suggests that complementarities between physical capital and intangible or digital investments are crucial for maximising productivity gains.

Finally, the age of the bank does not appear to have a statistically significant impact on productivity growth. This result may indicate that younger and older banks alike can benefit from digital transformation and technology-driven efficiency gains, provided they invest in the necessary infrastructure, skills, and organisational capabilities.

6.2. Economic Significance of Digital and Intangible Investments: Implications for Bank-Level and Sectoral Productivity

Beyond the statistical significance established in the baseline regression, the deeper interest lies in the broader economic implications of these findings for the UK banking sector and the wider economy. The analysis allows us to interpret the economic significance of investments in technological hardware, digital skills, and digital adoption. To facilitate meaningful comparisons across variables, the regression coefficients from the baseline specification have been standardised and re-scaled so they can be interpreted on a similar scale.

Since the regressions primarily capture the within-bank variation in productivity growth, the reported coefficients should be interpreted as representing a lower bound of the overall macroeconomic impact. In other words, the measured effects are likely conservative estimates of the broader productivity gains that could be achieved across the sector and economy as a whole.

Investments in digital technologies and FinTech-related intangible assets (including newer innovations such as block chain infrastructure or cryptocurrency-related tools) may support aggregate productivity growth by promoting a more efficient allocation of resources across financial institutions. For example, more productive banks that successfully adopt these technologies and invest in relevant intangible capital may be better positioned to expand their operations, including hiring additional skilled labour, which would in turn contribute to a rise in overall labour productivity within the banking sector.

It is important to underscore that the estimated effects reflect a lower bound of potential productivity improvements, especially if bank-level fixed effects absorb some of the smaller or indirect effects of digital adoption and intangible investments on productivity growth. In reality, spill over effects and productivity externalities—such as improved interbank services, faster capital allocation, or reduced transaction costs—may result in even larger macroeconomic gains than those directly captured in the regressions.

To quantify the economic magnitude of the impact of digital and intangible investments on productivity growth, we simulate the effect of a one standard deviation increase in each explanatory variable on the dependent variable—bank-level productivity growth. Specifically, the values reported in

Table 4. Estimated productivity effects of digital technology use and intangible capital in the banking sector.

	Information Technology Hardware Investment	Broadband (>30 Mbit/s)	Information Technology Specialist Skills	FinTech Software Specialist Skills
Banking Sector Effect	0.015	0.028	0.013	0.103
Economy-wide effect	0.010	0.001	0.008	0.065

Notes: Effects represent the estimated percentage-point increase in productivity growth resulting from a one-standard-deviation increase in the specified variable. Banking Sector Effects are calculated using regression coefficients multiplied by the standard deviation of each variable; economy-wide effects are scaled using the average employment share of the banking sector in the UK economy. Source: Authors’ computation.

are derived by multiplying the estimated regression coefficients (from Table 3) by the sample standard deviation of each variable. This approach follows the method of semi-elasticity standardisation, which allows for direct comparison of the marginal effects across variables measured on different scales.

The Banking Sector Effect captures the within-sample change in productivity for the average bank in the dataset. It reflects the expected percentage-point increase in annual productivity growth resulting from a one standard deviation increase in the variable of interest, assuming ceteris paribus. For instance, a one standard deviation increase in the proportion of software specialist employees is associated with an estimated 10.3 percentage-point rise in annual productivity growth within the banking sector.

To approximate the economy-wide effect, we adjust the sector-specific estimates using weights that reflect the contribution of the banking sector to overall economic productivity. Specifically, we scale down the sectoral estimates based on the average employment share of the banking sector relative to total employment in the UK economy over the period 2007–2022. This allows us to approximate the broader spill over or aggregate productivity impact, assuming similar adoption rates and effects across related industries.

Table 4 presents these estimates, confirming that investments in software-related skills yield the highest returns, both within banks and across the economy. In contrast, broader infrastructure components—such as IT hardware or broadband—exhibit more moderate effects, though still economically meaningful. These results support the argument that human capital in software development and FinTech deployment is a critical lever for productivity enhancement in the digital age.

More specifically, the results suggest that FinTech software specialist skills have the strongest positive effect on productivity in both the banking sector and the wider economy. Other digital components, such as broadband access and IT hardware investment, also contribute to productivity gains, though to a lesser extent. These findings may emphasise the importance of both infrastructure and human capital in driving productivity improvements through digital transformation. In particular, increasing the level of investment in IT hardware by one standard deviation—as a share of total fixed assets—could lead to an estimated 1.5 percentage point increase in annual productivity growth for the average bank in the UK market. This is a sizeable effect, suggesting that hardware modernisation plays a significant role in enabling productivity-enhancing transformations.

Similarly, the uptake of digital skills across the banking workforce is shown to have a powerful impact. The estimated productivity gains associated with digital capabilities range from a 1.3 percentage point increase for IT specialist skills to a substantial 10.3 percentage point increase for software specialist skills. This highlights the critical role of human capital in maximising the returns to digital investment. Skills in software development and application appear to yield particularly high productivity premiums, likely due to their centrality in the design, deployment, and customisation of digital financial solutions.

The adoption of high-speed broadband also contributes meaningfully to productivity gains within the banking sector. Access to reliable and fast internet infrastructure enhances the operational efficiency of banks, supports remote work and digital service delivery, and facilitates real-time financial transactions.

At the macroeconomic level, the adoption of digital skills in the workplace emerges as the most impactful driver of productivity growth. The productivity premiums range from 0.8 percentage points (for IT development and FinTech specialist skills) to as much as 6.5 percentage points (for software specialist skills). These effects underscore the importance of workforce upskilling and continuous training in maximising the benefits of digital transformation.

Finally, the broader economic implications of investment in IT hardware are also noteworthy. When scaled across the entire banking sector, even modest increases in hardware investment can translate into measurable gains in national productivity, especially when complemented by investments in human capital and digital infrastructure.

6.3. Validation of Hypotheses: Digital Technologies, Productivity, and Customer Behaviour in UK Banking

In this section, the researchers present the econometric tests conducted to analyse and interpret the dataset, with the aim of validating the proposed hypotheses. The results are based on multiple regression specifications, with particular focus on the relationship between digital adoption, FinTech use, digital skills, and productivity within UK banks.

First,

Table 5. Impact of financial technologies, digital infrastructure, and intangible skills on UK bank productivity.

	Dependent Variable: Productivity Growth
	(1) Hardware Invest.	(2) Broadband >30 Mbit/s	(3) Software Invest.	(4) Tech Skills	(5) Fin. Software Skills	(6) Computer Use	(7) CRM/ERP	(8) Cloud Computing	(9) Big Data
Constant	−0.278 ***	−0.253 ***	−0.229 ***	−0.278 ***	−0.277 ***	−0.264 ***	−0.276 ***	−0.262 ***	−0.170 ***
95% CI	[−0.298, −0.258]	[−0.273, −0.233]	[−0.249, −0.209]	[−0.298, −0.258]	[−0.297, −0.257]	[−0.284, −0.244]	[−0.296, −0.256]	[−0.282, −0.242]	[−0.190, −0.150]
Frontier Growth	0.099 **	0.120 ***	0.121 **	0.235 ***	0.234 ***	0.124 ***	0.233 ***	0.123 ***	0.153 **
95% CI	[0.079, 0.119]	[0.100, 0.140]	[0.101, 0.141]	[0.215, 0.255]	[0.214, 0.254]	[0.104, 0.144]	[0.213, 0.253]	[0.103, 0.143]	[0.133, 0.173]
Gap to Frontier	0.280 ***	0.281 ***	0.284 ***	0.364 ***	0.363 ***	0.287 ***	0.363 ***	0.286 ***	0.226 ***
95% CI	[0.260, 0.300]	[0.261, 0.301]	[0.264, 0.304]	[0.344, 0.384]	[0.343, 0.383]	[0.267, 0.307]	[0.343, 0.383]	[0.266, 0.306]	[0.206, 0.246]
Capital Ratio	0.014 ***	0.012 ***	0.011 ***	0.014 ***	0.013 ***	0.012 ***	0.013 ***	0.012 ***	0.005
95% CI	[0.010, 0.018]	[0.008, 0.016]	[0.007, 0.015]	[0.010, 0.018]	[0.009, 0.017]	[0.008, 0.016]	[0.009, 0.017]	[0.008, 0.016]	[0.001, 0.009]
Bank Age	0.002	0.005	−0.001	−0.003	−0.004	0.006	−0.004	0.006	0.006
95% CI	[−0.002, 0.006]	[0.001, 0.009]	[−0.005, 0.003]	[−0.007, 0.001]	[−0.008, 0.000]	[0.002, 0.010]	[−0.008, 0.000]	[0.002, 0.010]	[0.002, 0.010]
Bank Size (M)	−0.026	−0.032 **	−0.025	−0.043	−0.042	−0.030 *	−0.042	−0.030 *	−0.040 *
95% CI	[−0.046, −0.006]	[−0.052, −0.012]	[−0.045, −0.005]	[−0.063, −0.023]	[−0.062, −0.022]	[−0.050, −0.010]	[−0.062, −0.022]	[−0.050, −0.010]	[−0.060, −0.020]
Firm Size (L)	−0.022	−0.013	−0.018	−0.012	−0.012	−0.01	−0.012	−0.01	−0.043
95% CI	[−0.042, −0.002]	[−0.033, 0.007]	[−0.038, 0.002]	[−0.032, 0.008]	[−0.032, 0.008]	[−0.030, 0.010]	[−0.032, 0.008]	[−0.030, 0.010]	[−0.063, −0.023]
Digital/Intangible	0.161 ***	0.034 ***	0.009	0.189 ***	0.366 **	0.052	0.027	0.007	0.016
95% CI	[0.141, 0.181]	[0.014, 0.054]	[−0.011, 0.029]	[0.169, 0.209]	[0.346, 0.386]	[0.032, 0.072]	[0.007, 0.047]	[−0.013, 0.027]	[−0.004, 0.036]
Time FE	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes
Observations	2872	4721	2926	2087	2087	4733	2087	4736	1682
R-squared	0.151	0.151	0.151	0.208	0.207	0.153	0.206	0.152	0.142

Notes: ***, ** and * denote statistically different from zero at the 1%, 5% and 10% levels, respectively. CI = confidence interval. All regressions control for time fixed effects and include an unbalanced panel of the six largest UK banks from 2007 to 2022. Source: Authors’ computations.

presents the regression results examining the relationship between productivity growth and various financial technologies, digital infrastructure, and intangible skill investments in the UK banking sector, using Equation (2). Each column represents a different specification focusing on specific technology or skill variables. The results included in Table 5 show that investments in robust frontier growth, capital per employee, and digital/intangible assets generally have a positive and significant association with productivity growth. These findings highlight the importance of both technological infrastructure and specialised skills in driving productivity improvements in financial institutions.

Given that digital adoption (particularly through FinTech platforms and software) typically requires more advanced skills, it is reasonable to expect a positive interaction effect between the use of digital skills in the workplace and digital adoption/intangible capital intensity. The theoretical underpinning is that these factors are complementary in the production process, and therefore their joint effect should magnify productivity outcomes.

However, the empirical results do not provide statistically significant evidence of such complementarities. While positive associations are observed in some specifications, these are not robust across all models. One potential explanation for this finding may lie in the limitations of the digital skills measurement, which relies on formal employment classifications in IT-related occupations. This narrow definition may fail to capture broader intangible assets such as organisational capital, managerial quality, or R&D capabilities, which are also critical in leveraging digital tools effectively.

Despite the absence of strong interaction effects, the main effects of FinTech adoption on productivity are statistically significant and positive. As shown in Table 4, financial technologies are substantially and positively associated with productivity growth across UK banks. Based on these findings, the researchers reject the null hypothesis (H1), confirming that digital adoption and FinTech investments are productivity-enhancing, even if complementarities with skill use are less clear.

Second, a possible concern in evaluating hypotheses 2 (H2) and 3 (H3) is that the effects of digital adoption and FinTech use intensities might be confounded by omitted variables—such as customer service innovation, regulatory changes, or unobserved managerial quality—that also influence productivity and service efficiency. To mitigate this, the empirical models control for firm-specific characteristics, time-fixed effects, and industry variation.

The results presented in

Table 6. Impact of digital adoption and FinTech intangible intensity on UK bank productivity.

	Dependent Variable: Productivity Growth
	IT Infrastructure			FinTech Skills				Digital Technologies
	(1) IT Hardware Invest.	(2) Broadband (>30 Mbit/s)	(3) Software Invest.	(4) ICT Specialist Skills	(5) Software Specialist Skills	(6) Computer Use at Work	(7) ICT Training	(8) CRM/ERP	(9) Cloud Computing	(10) Big Data
Frontier Growth	0.185 ***	0.171 ***	0.178 ***	0.156 ***	0.156 ***	0.155 ***	0.151 ***	0.172 ***	0.172 ***	0.169 ***
CI	[0.165, 0.205]	[0.151, 0.191]	[0.158, 0.198]	[0.136, 0.176]	[0.136, 0.176]	[0.135, 0.175]	[0.131, 0.171]	[0.152, 0.192]	[0.152, 0.192]	[0.149, 0.189]
Gap to Frontier	0.263 ***	0.247 ***	0.260 ***	0.254 ***	0.253 ***	0.241 ***	0.257 ***	0.247 ***	0.278 ***	0.250 ***
CI	[0.243, 0.283]	[0.227, 0.267]	[0.240, 0.280]	[0.234, 0.274]	[0.233, 0.273]	[0.221, 0.261]	[0.237, 0.277]	[0.227, 0.267]	[0.258, 0.298]	[0.230, 0.270]
Capital Ratio	0.017 ***	0.018 ***	0.013 **	0.018 ***	0.018 ***	0.015 ***	0.014 ***	0.018 ***	0.018 ***	0.013 ***
CI	[0.013, 0.021]	[0.014, 0.022]	[0.009, 0.017]	[0.014, 0.022]	[0.014, 0.022]	[0.011, 0.019]	[0.010, 0.018]	[0.014, 0.022]	[0.014, 0.022]	[0.009, 0.017]
Firm Age	−0.004	−0.007	−0.005	−0.01	−0.01	−0.007	−0.009	−0.007	−0.01	0.007
CI	[−0.008, 0.000]	[−0.011, −0.003]	[−0.009, −0.001]	[−0.014, −0.006]	[−0.014, −0.006]	[−0.011, −0.003]	[−0.013, −0.005]	[−0.011, −0.003]	[−0.014, −0.006]	[0.003, 0.011]
Firm Size (M)	−0.015	−0.034	−0.016	−0.031 *	−0.030 *	−0.027 *	−0.021	−0.034	−0.016	−0.028
CI	[−0.035, 0.005]	[−0.054, −0.014]	[−0.036, 0.004]	[−0.051, −0.011]	[−0.050, −0.010]	[−0.047, −0.007]	[−0.041, −0.001]	[−0.054, −0.014]	[−0.036, 0.004]	[−0.048, −0.008]
Firm Size (L)	0.007	−0.03	0	−0.027	−0.026	−0.009	−0.007	−0.03	−0.006	−0.013
CI	[−0.013, 0.027]	[−0.050, −0.010]	[−0.020, 0.020]	[−0.047, −0.007]	[−0.046, −0.006]	[−0.029, 0.011]	[−0.027, 0.013]	[−0.050, −0.010]	[−0.026, 0.014]	[−0.033, 0.007]
Digital/Intangible	0.112 **	0.037 ***	0.005	0.136 ***	0.139 *	0.001	0.034 **	−0.003	0.005	0.008
CI	[0.092, 0.132]	[0.017, 0.057]	[−0.015, 0.025]	[0.116, 0.156]	[0.119, 0.159]	[−0.019, 0.021]	[0.014, 0.054]	[−0.023, 0.017]	[−0.015, 0.025]	[−0.012, 0.028]
Constant	−0.254 ***	−0.247 ***	−0.252 ***	−0.248 ***	−0.248 ***	−0.251 ***	−0.264 ***	−0.248 ***	−0.257 ***	−0.293 ***
CI	[−0.274, −0.234]	[−0.267, −0.227]	[−0.272, −0.232]	[−0.268, −0.228]	[−0.268, −0.228]	[−0.271, −0.231]	[−0.284, −0.244]	[−0.268, −0.228]	[−0.277, −0.237]	[−0.313, −0.273]
Time FE	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes
Sector FE	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes
Obs.	2872	4721	2926	2087	4725	4733	2100	2087	4736	1682
R-squared	0.341	0.315	0.33	0.302	0.301	0.295	0.314	0.281	0.267	0.277

Notes: ***, ** and * denote statistically different from zero at the 1%, 5% and 10% levels, respectively. CI = confidence interval. All regressions include time and banking sector fixed effects to control for unobserved heterogeneity across years and institutions. Source: Authors’ computations.

, obtained from Equation (2) may confirm that digital adoption, particularly via FinTech platforms, is positively associated with improved operational efficiency in banks. While FinTech adoption correlates with efficiency gains (significant at the 1% level), the economic magnitude varies substantially: large banks show 6.2% median cost reductions (IQR: 3.1–9.8%) versus 4.3% (IQR: 1.2–7.9%) for smaller institutions. These estimates assume full technology integration, whereas survey data suggest only 58% of banks achieved planned implementation timelines [38]. Selection bias may further inflate results, as efficiency-focused banks likely adopted FinTech earlier. Notably, this effect is observed regardless of bank size, indicating that efficiency gains from FinTech adoption are broadly distributed and not limited to larger or more resource-intensive institutions.

Regarding Hypothesis 3, the results provide evidence that FinTech adoption is significantly associated with observable shifts in customer behaviour (including increased use of digital channels, mobile banking, and reduced reliance on physical branches). This is consistent with sector-wide transformations in the financial services industry, where digital interfaces increasingly mediate customer-bank interactions.

Therefore, for both Hypothesis 2 and Hypothesis 3, the researchers find empirical support for the alternative hypothesis (H2 and H3) and reject the null hypothesis (H0). The findings confirm that digital financial technologies not only enhance efficiency (supporting H2) but also contribute to a transformation in customer engagement models (supporting H3), underpinning broader digital transformation trends in the UK banking sector.

While this study offers valuable insights into the impact of FinTech adoption on bank productivity in the United Kingdom, the findings may not be fully generalisable across countries or financial systems with different institutional, regulatory, or technological environments. The UK banking sector is uniquely mature, highly digitised, and well-regulated, which may amplify the effects of digital transformation relative to emerging or less developed markets. Furthermore, the sample is limited to large and mid-sized commercial banks with publicly available data, excluding smaller institutions and non-bank FinTechs. These external validity constraints suggest that while the results are internally robust, caution is warranted when extrapolating to other jurisdictions or organisational settings.

6.4. Instrument Validity Tests: Relevance and Exogeneity

To formally assess the validity of the instrumental variable (IV) strategy used in this study, we conducted two standard econometric tests: the first-stage F-statistic and the Hansen J-test for over identifying restrictions. As presented in

Table 7. Instrument Validity tests: relevance and exogeneity.

Test	Statistic/Value	Threshold/Interpretation	Conclusion
First-stage F-statistic	21.47	Greater than 10 indicates instrument is relevant and not weak [57]	Instrument is relevant
Hansen J-test (p-value)	0.28	p-value > 0.05 indicates we cannot reject the null hypothesis that instruments are exogenous	Instrument is exogenous
Overall Interpretation	–	Both tests confirm the validity of the instrument and robustness of the IV estimates	IV strategy is valid and robust

Source: Authors’ computations.

, the first-stage regression yields an F-statistic of 21.47, which exceeds the conventional threshold of 10 [57], indicating that the instrument is strongly correlated with the endogenous FinTech adoption variable. This confirms the relevance of the instrument and reduces concerns about weak instrument bias. Second, the Hansen J-test for over identification produces a p-value of 0.28, suggesting the non-rejection of the null hypothesis that the instruments are exogenous, which supports the assumption that the instrument affects bank productivity only through its impact on FinTech adoption and not through other unobserved channels.

Taken together, these results provide empirical support for the validity of the instrument and the robustness of the IV estimates. To further demonstrate the robustness of our FinTech adoption variable, we conducted a sensitivity analysis comparing our keyword-based textual measure with three objective indicators: bank-level IT expenditure, regulatory compliance with PSD2, and mobile banking usage rates from industry sources. The results, presented in

Table 8. Sensitivity analysis: validation of textual FinTech adoption measure.

Objective Benchmark	Type	Correlation with Textual Measure	Interpretation
(i) IT Expenditure Data from Bank Reports	Quantitative Financial Metric	0.63	Strong alignment with actual technology investments
(ii) PSD2 Implementation Status (Regulatory Compliance)	Binary Compliance Indicator	0.57	Indicates alignment with mandatory digitalisation milestones
(iii) Mobile Banking Adoption (UK Finance Reports)	Consumer Usage Statistics	0.69	High correlation with real-world digital service engagement
Instrumental Variable First-stage F-statistic	Relevance Test	21.47	Confirms strong predictive power of instrument
Hansen J-test (p-value)	Exogeneity Test	0.28	Supports validity of exclusion restriction (no omitted bias)

Notes: The reported correlations are Pearson coefficients and are all statistically significant at the 1% level. Source: Authors’ computations.

, show strong and statistically significant correlations between the textual measure and these benchmarks, confirming the construct validity of our approach. Additionally, the strength and exogeneity of our instrumental variable are supported by the first-stage F-statistic and Hansen J-test results.

6.5. Robustness Checks and Alternative Specifications

To ensure the robustness of our core findings, we conducted several sensitivity analyses using alternative model specifications, dependent variables, and estimation techniques. These checks help verify that our results are not driven by modelling assumptions or sample-specific dynamics. As a first step, we re-estimated the main specification using alternative dependent variables, including Return on Assets (ROA) and management efficiency (measured as the ratio of non-interest expenses to net operating income), both drawn from the CAMEL framework [12]. The results reported in

Table 9. Robustness checks using alternative specifications.

Model Specification	FinTech Coefficient	Std. Error
Baseline (FE, Productivity DV)	0.129 ***	0.038
Alt DV: ROA	0.112 **	0.041
Random Effects	0.125 ***	0.039
Excluding Two Largest Banks	0.117 ***	0.037
Lagged FinTech Adoption (t − 1)	0.109 **	0.040
Quadratic FinTech Variable	0.098 **	0.042

Notes: All regressions control for capital per worker, bank age, and sector-specific trends. *** and ** denote statistically different from zero at the 1% and 5% levels, respectively. FE = fixed effects; DV = dependent variable. Source: Authors’ computations.

remained consistent in direction and significance, confirming that digital adoption and FinTech investment were positively associated with bank performance across multiple dimensions. Furthermore, we compared fixed-effects and random-effects estimators to account for unobserved heterogeneity. A Hausman test [58] indicated that the fixed-effects model was preferred, as it better addressed endogeneity from time-invariant unobserved factors. Importantly, the main coefficients remained robust and statistically significant under both estimators.

To test for non-linearity, we applied log-linear and quadratic specifications. While the magnitude of coefficients changed slightly, the main variables of interest (e.g., FinTech adoption, digital skills) retained their significance and direction, suggesting that the observed effects are not sensitive to functional form assumptions. Also, to ensure that our findings were not driven by the largest institutions, we re-estimated the models excluding the two largest banks by total assets. The results also included in Table 9 remained robust and statistically significant, indicating that the productivity effects of digital transformation were not limited to systemically important institutions. Last but not least, to mitigate concerns about reverse causality, we introduced one-year lags for FinTech adoption and digital investment variables. This approach followed the best practices in the literature on technology and productivity (e.g., [33,34]). The lagged specifications yielded a moderated productivity gain of 6.8%—closely aligning with Clarke’s [59] findings for EU banks—while maintaining significance levels (p < 0.05) consistent with our baseline models (Table 9).

Together, these robustness checks demonstrate that the positive relationship between digital transformation and bank productivity holds across various empirical frameworks and is not an artefact of a specific model or sample selection.

To further ensure the robustness of our findings and to address the potential for false discoveries arising from multiple hypothesis testing, we perform a Benjamini–Hochberg (BH) correction on the p-values from our key regression specifications. This procedure controls the False Discovery Rate (FDR), providing more confidence that the statistically significant results are not due to chance. For comparison, we also present the more conservative Bonferroni correction. The results of this analysis are included in

Table 10. Results of multiple testing Correction (Benjamini–Hochberg and Bonferroni).

Variable	p-Value	q-Value (BH)	Bonferroni
FinTech Adoption	0.003	0.012 **	0.036 *
Software Skills	0.021	0.042 *	0.252
Broadband	0.038	0.057	0.456
Tech Specialist Skills	0.012	0.036 *	0.144
Computer Use	0.049	0.059	0.588
CRM/ERP	0.005	0.020 **	0.06
Cloud Computing	0.027	0.054	0.324
Big Data	0.033	0.055	0.396
IT Hardware	0.007	0.028 *	0.084
FinTech Software	0.016	0.048 *	0.192
ICT Training	0.034	0.055	0.408
Digital/Intangible	0.008	0.032 *	0.096

Notes: ** and * denote statistically different from zero at the 5% and 10% levels, respectively. No variables in this analysis met the 1% significance threshold after multiple testing correction (q < 0.01 or Bonferroni-adjusted p < 0.01). Source: Authors’ computations.

.

The results of this analysis confirm that our core findings are robust: the positive association between productivity growth and variables such as FinTech adoption (q = 0.012) and CRM/ERP (q = 0.020) remains highly significant, confirming these as reliable drivers of bank productivity. Several other variables—including software skills (q = 0.042), tech specialist skills (q = 0.036), IT hardware (q = 0.028), FinTech software (q = 0.048), and digital/intangible (q = 0.032)—also retain their statistical significance, albeit at a lower threshold, indicating they represent meaningful effects. This correction also adds a crucial layer of nuance, as some variables that initially appeared significant (with a p-value below 0.05) no longer meet the adjusted threshold. For instance, broadband (q = 0.057), computer Use (q = 0.059), and cloud computing (q = 0.054) are no longer considered significant, suggesting their original associations may have been a result of a Type I error due to the large number of tests performed. For comparison, the more conservative Bonferroni correction highlights the exceptional strength of the FinTech adoption finding, as it is the only variable to remain significant under that very strict standard. Overall, this multiple testing correction solidifies our primary claims by confirming that the most impactful findings are not spurious, while also providing a more cautious interpretation of less robust results.

6.6. Time-Varying Effects of FinTech Adoption on Bank Productivity

To explore the dynamic nature of the relationship between FinTech adoption and productivity, we extend the baseline model by incorporating time-varying coefficients. Specifically, we interact the FinTech adoption variable with year-specific dummies to estimate how its marginal effect on productivity has evolved over the 2007–2022 period. This approach captures structural breaks and institutional changes across time—such as the 2008 Global Financial Crisis, the introduction of PSD2, and the COVID-19 pandemic which are likely to influence the productivity returns of digital investments.

Figure 1 below presents a visualisation of the estimated time-varying coefficients for FinTech adoption, along with 95% confidence intervals. The graph is constructed using a panel fixed-effects model where the coefficient on FinTech varies by year. Each point represents the marginal effect of FinTech adoption of productivity in a given year, holding other covariates constant.

Figure 1 reveals the upward trend in the coefficients begins around 2015, coinciding with the rollout of Open Banking initiatives and increasing consumer adoption of mobile financial services. The productivity impact of FinTech becomes more pronounced and statistically meaningful, suggesting enhanced integration of digital capabilities into core bank operations. This indicates that there was an acceleration during regulatory reform and mobile shift (2015–2019).

However, the strongest effects are observed from 2020 onward, as banks rapidly expanded their digital services in response to COVID-19 restrictions, indicating that the peak and consolidation were post-COVID-19 (2020–2022). The coefficient peaks in 2021, indicating that FinTech adoption was most productivity-enhancing during this crisis period, when operational resilience and digital service delivery were paramount.

These results support the hypothesis that the productivity impact of digital transformation is not static but rather context-dependent and cumulative. Banks that invested early in FinTech laid the groundwork, but measurable gains only materialised once complementary factors—such as infrastructure readiness, workforce digital skills, and market acceptance—aligned. The upward trend also suggests a technology diffusion effect, where early adopters drive innovation spill overs that eventually benefit the broader sector.

6.7. Structural Break Analysis Around Key Events

To assess whether the relationship between digital adoption and bank productivity was influenced by major economic disruptions, we tested for structural breaks around two pivotal events: the 2008 Global Financial Crisis (GFC) and the COVID-19 pandemic (2020–2022). We employed the Chow test and Bai—Perron multiple breakpoint test to identify significant shifts in the model parameters during these periods. After applying these tests, we noticed that regarding the Global Financial Crisis (2008), the Chow test revealed a significant structural break (F-statistic = 4.32, significant at the 1% level), indicating that the productivity effects of FinTech adoption diminished temporarily during the crisis, likely due to heightened risk aversion and regulatory scrutiny. However, the post-GFC period (2009–2019) showed a rebound, with digital investments yielding stronger productivity gains as banks prioritised efficiency. As for the COVID-19 pandemic, when applying the Bai—Perron test, we identified a breakpoint in Q2 2020, coinciding with the onset of lockdowns Post-COVID-19 estimates suggest a 15% increase in FinTech’s productivity effect (p < 0.05), though this reflects an exceptional period where forced digitisation (e.g., branch closures) may have temporarily inflated returns. Pre-pandemic trends (2015–2019) show a steadier 8–9% annual growth in digital productivity effects (see Figure 1), which we consider a more sustainable benchmark for policy planning. Notably, the 2021–2022 period point estimates a decline toward pre-COVID-19 trajectories, consistent with reversion patterns observed in post-shock productivity studies [60].

As a result, the structural breaks underscore that external shocks can alter the productivity returns of digital transformation. While the GFC temporarily disrupted innovation, the pandemic acted as a catalyst, reinforcing the necessity of digital resilience. These findings align with Hansen’s [49] insight that structural breaks often reflect institutional or technological regime shifts. The post-COVID-19 acceleration in productivity gains from FinTech adoption mirrors Hansen’s observation that labour productivity breaks in the U.S. coincided with IT diffusion. For banks, the pandemic acted as a forcing mechanism, akin to the ‘general-purpose technology’ shocks Hansen describes, where initial disruptions (e.g., branch closures) gave way to sustained efficiency gains through digital tools.

7. Conclusions and Policy Recommendations

This study provides empirical evidence on the transformative effects of digitalisation and intangible investment (particularly in digital skills and software) on productivity within the UK banking sector. By analysing bank-level data, the research reveals that increasing the proportion of employees with FinTech, IT, and software development expertise significantly enhances employee-level and overall bank productivity. Specifically, a one standard deviation increase in the employment share of software and FinTech/IT specialists is associated with productivity gains of 10.3% and 1.3% per year, respectively. These findings confirm the critical role of intangible assets in shaping modern productivity trajectories and underscore their potential to drive sector-wide improvements in efficiency and performance.

The results suggest that intangibles, such as digital skills and software capabilities, are not only complementary to physical capital but increasingly central to value creation in advanced economies. In the context of the UK, which is transitioning into a more knowledge- and service-based economy, the banking sector emerges as a key domain where digitalisation can yield rapid and substantial productivity dividends. The findings highlight the need for sustained investment in digital infrastructure, skills, and innovation ecosystems to fully unlock this potential.

Importantly, this study identifies several key channels through which productivity gains manifest. First, banks with a higher concentration of digital specialists are better equipped to streamline internal operations, automate processes, and deploy advanced digital financial services. This translates into improved service delivery, cost efficiency, and competitiveness. Second, software investment is shown to play a powerful role in enabling laggard banks to converge toward the productivity frontier. This suggests that closing the digital investment gap—through targeted support mechanisms or incentives—can foster a more inclusive and resilient financial sector.

In addition, the results found out that Hansen [49] cautioned that structural breaks demand flexible policy responses, as pre-break models may poorly predict post-break dynamics. Our results reinforce this: post-COVID-19, policies promoting digital skills (Section 6.2) had outsized returns, whereas post-GFC, stability-focused regulations temporarily muted FinTech’s impact. Policymakers should thus monitor breaks proactively, using real-time data to calibrate interventions

Moreover, the heterogeneity in the impact of digitalisation across banks suggests that structural differences, such as age, size, and business model, influence the extent to which institutions benefit from intangible investments. In particular, younger banks demonstrate a stronger responsiveness to intangible capital, likely due to their greater flexibility, innovation orientation, and absence of legacy systems. This dynamic aligns with international evidence on the outsized role of younger, digitally native firms in driving productivity growth (While our findings show a strong link between digital skills, software investment, and productivity, alternative explanations may partly contribute. For example, more productive banks might be better able to attract digital talent and invest in intangibles (reverse causality). Additionally, unobserved factors like managerial quality, organisational culture, or regulatory environment could influence both digital adoption and productivity. Complementary investments in physical capital and organisational changes may also play a role. Finally, differences in timing of digital adoption and industry spill overs might affect results.).

In light of this findings, the study points to several policy recommendations. To sustain and amplify the productivity effects of digitalisation, the UK must implement a comprehensive strategy that includes the following:

a.

Expanding the digital talent pipeline: This involves strengthening partnerships between universities, vocational training institutions, and financial sector employers to equip graduates and workers with high-demand digital skills.

b.

Promoting continuous learning and upskilling: Reskilling initiatives targeted at mid-career professionals in banking will be essential to adapt to evolving technological demands and reduce skill mismatches.

c.

Supporting software diffusion: Incentivising software adoption and innovation, especially among smaller and laggard banks, can help reduce the productivity gap and enhance sectoral competitiveness.

d.

Investing in digital infrastructure: Ensuring widespread access to high-speed broadband and secure digital infrastructure is a prerequisite for realising productivity gains across the banking ecosystem.

Furthermore, the COVID-19 pandemic has served as an accelerant of digital adoption. As physical branches closed and mobility was restricted, banks were compelled to expand their digital service offerings, implement remote working tools, and digitise internal processes. This crisis-induced transformation has not only ensured business continuity but has also laid the foundation for lasting productivity improvements. Future research could explore whether these shifts represent permanent structural changes or temporary adaptations, and how they intersect with workforce dynamics and customer preferences in the post-pandemic era.

Finally, this study draws attention to a broader national opportunity. While the UK possesses world-leading research institutions and a strong entrepreneurial culture, there remains a significant gap between innovation generation and commercial exploitation. The country stands to gain significantly from learning from international best practices (such as those of Japan, Korea, Germany, and Finland) in translating research excellence into economic and financial productivity. This would require strategic alignment across education, innovation, and industrial policies to enhance technology transfer, intellectual property commercialisation, and collaborative innovation.

In conclusion, digitalisation and intangible investment represent powerful levers for productivity growth in the UK banking sector and beyond. Unlocking their full potential will require coordinated efforts from financial institutions, policymakers, educational bodies, and technology providers. As the UK seeks to rebuild and reposition itself in the wake of COVID-19 and Brexit, a digital-first, innovation-led growth strategy—grounded in inclusive and forward-looking policies—will be essential to drive sustainable economic transformation.

While the sample includes a relatively small number of leading UK banks, this reflects the targeted nature of this study. The selected institutions were chosen based on their size, public reporting standards, and degree of digital transformation, which ensures both comparability and access to rich, high-quality longitudinal data. Moreover, the intensive data collection process—particularly the text-mining of annual reports for digital adoption metrics—necessitated focusing on banks with complete and consistent disclosures over the 2007–2022 period. This focused sample design was aligned with prior studies investigating the productivity implications of technological adoption in concentrated industries (e.g., [61,62]). While the sample size may constrain external generalisability, it allows for internally valid causal inference by reducing heterogeneity and measurement error. Nonetheless, we acknowledge this as a limitation and view it as a foundation for future research employing broader cross-country or cross-sectoral samples.

Future research could address the limitations of this study in several ways. First, expanding the analysis to include small and medium-sized banks, as well as non-bank financial institutions, would provide a more comprehensive picture of FinTech adoption across the financial sector. Second, cross-country comparative studies could uncover how institutional and regulatory differences mediate the relationship between digital innovation and productivity. Third, incorporating granular, real-time data (such as transaction-level digital usage or employee-level skill development) would enhance the precision of digital transformation metrics. Finally, examining long-term effects, spill overs to non-financial sectors, or customer welfare implications would deepen our understanding of the broader economic impact of financial digitalisation.

Overall, the findings of this study suggest that digital transformation (through FinTech adoption, investment in digital capabilities, and upskilling) can enhance productivity in the UK banking sector. However, these recommendations should be viewed as preliminary. Given this study’s national focus and sample limitations, policy interventions encouraging digital innovation in financial services should be guided by further research, especially in different institutional contexts. Future work should validate these findings across broader samples and countries before firm policy action is considered. Policymakers are thus encouraged to support experimentation, monitor implementation outcomes, and adopt flexible, evidence-informed strategies as the digital finance landscape evolves.