Digital payments system and market disruption

ABSTRACT The traditional banking functions of lending, deposit-taking and payment intermediation are being unbundled in the new frontiers of money that extend from virtual currencies to crypto-assets and from shadow payments to quasi-money. The possibility for digital-centred change in the financial industry is illustrated by distributed ledger technology, of which ‘blockchain’ is the most prominent example of automated decision-making. Other forms of decentralised supply of money, payment services, and funding processes may allow households and businesses to obtain loans and pool risks without having recourse to financial intermediaries. This article examines the alternative provision of access to low-cost zero-friction payments from the perspective of the underbanked. Promoting innovation through alternatives to credit means integrating vulnerable and excluded customers into mainstream financial systems. Blockchain technology backed by a possible modification of the law on the recognition and transfer of property rights might prove instrumental in unlocking the value of the assets possessed by the underbanked or even the unbanked.


A. Introduction
The digitalisation of payments systems has resulted in a large use of artificial intelligence (AI) applications, e.g. machine learning, neural networks and adaptive algorithms which automate decision-making processes while expediting the delivery of financial services. 1echnology innovation assists firms' operating systems by supporting both risk assessment and peer comparison.But this can lead to problems.For example, where machine learning embeds manual intervention and the software for the algorithms on which the rules rely is not perfectly designed. 2The Financial Conduct Authority (FCA) launched sandbox programmes to enhance oversight of prudential and conducts risks of fintech firms and incentivise collaborative dialogue between regulators and regulated firms. 3This regulatory initiative has been followed by the EU Digital Finance Package, which includes an innovative programme, the distributed ledger technology (DLT) Pilot Regime for blockchain services. 4The EU sandbox aims to establish a harmonised framework in the decision-making process between the DLT market infrastructures, the national competent authorities and European financial regulators. 5The FCA also launched a regulatory 'scale up box', a second generation of digital sandboxes, to improve transparency in developing fintech products and systems of internal control. 6ccess to digital technology supports financial inclusion and data sharing, which means banks provide financial services at affordable costs to disadvantaged customers.Whether technology-based forms of payments create the conditions to improve participation in the credit market among minority groups (i.e.individuals with disabilities and mental health issues, ethnic minorities, low-income households) depends on the design of firms' internal data and their operational and management information process. 7This article examines the alternative provision of access of low-cost zero friction payments from the perspective of the underbanked.Blockchain technology backed by a possible modification of the law on the recognition and transfer of property rights might prove instrumental in unlocking the value of the assets possessed by the unbanked. 8he article proceeds as follows.The next section examines the digitalisation of bank business activities along with the digital transformation of payment platforms, and analyses which elements of regulation are potentially most easily automated.It explores whether automation of know-your-customer (KYC) rules and anti-money-laundering (AML) decisions (e.g.automated blocking of unusually large payments) are 'algorithm ready' and presented in a way which would allow automated software-based compliance.This raises the question of interpretation of data to ensure consistency and consensus in the automation process.
Section C considers regulatory technology (regtech) applications in the decisionmaking process of bank intermediaries and illustrates how they adapt to AI systems for producing information about modelling prudential risk and capital at risk for investors and regulators.Automated systems can incentivise a greater use of principles and judgement in regulation and supervision, even though the need to make the rulebook machine-readable might initially favour a shift away from principles, and towards rigidly applied rules.It also discusses the role of automated supervisory actions and how it will be affected by the use of algorithms to mitigate undesired regulatory outcomes, and whether artificial applications will reduce the role of human judgement in monitoring compliance with principles.Achieving the full benefits of technology in regulatory oversight will therefore require standardised access to institutions' operating systems and the data they contain.The opportunities provided by technology-based products can be used to transform regulatory oversight in ways that will yield far wider benefits than are sometimes envisaged.
Section D addresses the main issues around the use of digital payments systems for vulnerable consumers and questions about potential discriminatory outcomes of algorithmic machines which can hinder financial inclusion.Using technology to improve the information available to investors and customers allows them to better observe and anticipate business decisions, and therefore ensures that these are desired outcomes, consistent with regulatory principles.The last section sets out some concluding remarks.

B. Digital financial intermediaries
The use of technology in the banking sector generally refers to algorithms in business decision-making both in investment contracts and in the business strategy.9Notably, the application of algorithms by financial institutions is used for automated risk assessment.Most algorithms operate within set variables, but self-learning algorithms operate beyond the control of their programmers, which makes the role of traditional mutual forms of traditional financial intermediary (e.g.building societies and credit unions) increasingly virtual rather than face-to-face. 10Machine learning introduces automated agents such as robo-advisors and chatbots, although it is debatable whether they could ever have the same rights as a natural person. 11egulatory technologies can increase the speed of client on-boarding and reduce risk as a distributed shared ledger acts as an immutable assured audit trail of all KYC processes and the automation of account opening. 12Digitalisation of lending transactions through automated reading of data can enhance traceability of customers (e.g.verification of a customer's identity) and disclosure of information (e.g.KYC due diligence). 13This could be particularly useful in the area of anti-money laundering (AML) and could mitigate the cyber risk of crypto assets being used for criminal purposes. 14AML rules are inherently affected by technological platforms (e.g.data-mining techniques) which shape the regulatory framework into a data-based regime. 15The EU Fifth Money Laundering Directive (5AMLD) provides for centralised automated mechanisms for bank and payment accounts to protect and verify personal data when carrying out AML investigations. 16Specifically, the 5AMLD regulates central registries and central electronic data retrieval systems so as to allow timely identification of any natural and legal persons involved in suspicious activities. 17The digital environment of financial transactions is exposed to cybercrimes, which makes the risk-based approach of AML obsolete and calls in question the need to develop automated regulation. 18In this context, the use of technological applications can reduce the costs of intermediation by automating the collection, verification and transmission of required information to the regulatory authorities. 19At the same time, it supports regulatory objectives by improving the accuracy and comparability of information, which would also enhance the ability of supervisors to monitor regulatory compliance.
Technology replaces traditional forms of financial intermediation by digital intermediary channels which aim to include underbanked and vulnerable customers in mainstream credit systems. 20Access to financial services through sophisticated software has led to the automation of consumer platforms, notably crowdfunding (donation-based, rewardbased and equity) 21 and peer-to-peer lending (P2P). 22As a result, digital payments have reduced the role of intermediaries in evaluating the customer's profile and the suitability of products.Payment systems are closely linked to the broader impact of technology on the financial industry and public services and to wider issues of identity and data infrastructure (Paypal, M-Pesa, AliPay, WeChat Pay). 23While there is a very large diffusion of data through digital investment schemes (e.g.biometric identification), there is relatively scarce understanding of the policy and practice applied to cyber security.
The growth of the P2P market, mainly driven by crowdfunding platforms, has provided investors with automated access to loan portfolios and created alternative forms of funds for consumers, although policy and regulatory issues arise with respect to potential systemic risk in this new business model. 24In P2P lending mechanisms, users lend capital to borrowers and investors receive a credit claim to document the principal's commitment without recourse to bank intermediaries. 25Lenders and borrowers interact across automated investment tools such as artificial machines (LendingRobot) which support the lender's search cost and allow private investors to compete with institutional investors. 26utomated customer decision-making allocates lenders' funds automatically and assists investors to design loan portfolios: well-developed P2P platforms (Lending Club, Funding Circle and Prosper)27 offer automated lending options where a lender can auto-select lending criteria (interest rate, risk profile, market segment). 28These options improve screening and monitoring services, replacing the chain of intermediation in the assessment of borrowers' creditworthiness and loan requests. 29This enables the loan application and loan approval processes to be expedited, facilitating timely credit decisions for applicants. 30Automated lending processes support the human assessment in providing a loan and substantially lower underwriting and compliance costs for lenders; the resulting data can therefore be leveraged to improve their risk management. 31As a result, software-driven automated decision-making customises the origination and distribution of loans to consumers.However, this is accompanied by the risk of inflating the price of debt securities whilst the lack of transparency and the absence of prudential supervision can affect the quality of the lending market and increase financial instability. 32

C. Regtech and algorithmic systems
Algorithmic systems expedite the operation of bank intermediaries through sophisticated software which affords the opportunity to analyse legal texts without manual intervention. 33By employing digital solutions, i.e. a distributed shared ledger, an intermediary can rapidly verify the identity of its clients and assess the potential risks of illegal intentions for the business relationship. 34Automated decision-making involves information gathering and the communication of prudential risk to investors and regulators, particularly in relation to bank capital rules (e.g.pillar 1 and pillar 2 of Basel II). 35The use of judgement on the basis of principles rather than mechanical rules is limited in the case of bank capital and applied only for validation of capital models and to capital requirements under pillar 2, reflecting operational and other risks outside the standard categories of market and credit risk (and more recently through adjustments to the conservation buffer of Basel III). 36Automated systems to calculate risk weights have the potential to make transparent the internal models for credit scoring and loss-absorbency capacity.
Computerised analytical models and centralised standard-setting (e.g.shared data repositories) can create the conditions to aggregate the flows of information and coordinate more localised regulatory engagements. 37This centralised approach of risk modelling could redefine the regulatory burden between regulators and banks: advanced predictive analytics can improve the assessment of banks' credit exposure and probability of default. 38These are applications for AI models which generate predictions with respect to desired outcomes: the decision-making process is based on trained machine learning and underlying computer programmes which give rise to opacity in the data patterns. 39he development of new data technologies such as blockchain and APIs (Application Programming Interfaces) in the banking industry offers the opportunity, through dialogue between regulators and industry to address this opacity. 40utomated methodologies for the modelling and communication of capital at risk are integrated into regulatory frameworks, making the distinction between internal models and standardised approaches meaningless.Automated decision-making could reduce reliance on banks' sources of information about the creditworthiness of borrowers and firms' lending decisions through a shared data platform. 41This would limit the discretionary review of banks in granting loans while, in parallel, enhancing timely monitoring of risk and the predictability of unexpected losses.
The use of automated methods involves initial cost, risk of error in the system, risk of over-reliance and increased systemic risk if all firms follow similar artificial intelligence solutions that lead to highly homogeneous market behaviour. 42In this context, recourse to machine learning by bank intermediaries may give rise to certain harms for the credit market.Intelligent systems which make decisions impact humans, and what recourse humans have to take back control when algorithms fail or demonstrate prejudice constitutes a challenge for regulators and policymakers. 43In such a scenario, most or nearly all financial intermediaries would end up following similar strategies as the algorithms used would be likely to generate herd behaviour effects. 44Another risk involves the implications of machine disruption for privacy protection and data security.This is evident in the MiFID II algorithmic trading requirements 45 and the widespread industry interest in using a shared third-party for KYC regulations to disclose identity information and hence achieve lower costs of on-boarding.AI models such as deep learning can assist public authorities to expedite the supervisory functions of risk assessment, crisis management and investor protection.For instance, smart contracts and DLT can be used to automate the execution of financial transactions.Code governing a variety of applications could be designed so as to reflect regulatory standards and thus guarantee full compliance. 46Regtech can support regulators in formulating substantive rules and allow supervisors to assess the impact of a firm's risk models by transferring the underlying data sets into the supervisor's risk model systems. 47

I. The potential of regtech for firms' operating processes
Regtech is often viewed as a technological response to the vastly increased burden of compliance on financial firms since the global financial crisis. 48A major part of this increase in regulation has been introduced to enhance the resilience of the financial system, by strengthening prudential rules at the level of both the individual institution and the wider system.There has also been a marked growth in requirements of other forms of regulation, notably conduct of business and also KYC and AML reporting requirements. 49he opportunity for making revenues from reducing compliance burdens is an attractive one and has led to a substantial number of regtech start-ups. 50This is a natural application, with the emergence of new financial technologies running in tandem with the substantial rise in regulatory burden since the crisis. 51There are however clear limitations to what can be achieved from such automation of compliance.The use of technology in this manner to reduce compliance costs and strengthen supervision can conflict with the exercise of manual or judgmental intervention of regulators, for example in authorisation processes, detection of financial crime, or the identification of mis-selling practices. 52The technology is often applied to regulatory requirements in the design of financial products in a way that ensures compliance with rules, without the need for human input. 53In this sense, regtech might be seen as a more 'natural' partner for the rule-based variety of financial regulation, and it is likely that only a relatively small part of the existing financial rulebook can be easily translated into code for use in this manner.Since principles and outcomes remain widely used, this appears to place a fairly substantial limitation on the potential of regtech.It is also conceivable that following on from this, the expanded use of computer code in financial regulation might itself introduce undesirable pressure to shift back to a more command and control-focused regulatory style, raising broader questions about the shift to technology. 54The potential of regtech is not limited to reduce compliance costs and expedite decision-making processes: it extends to reshape the way customers engage with financial services.Specifically, the regulatory transformations of market infrastructures to include digital identities and blockchainenabled technologies lead to the use of regtech solutions to foster financial inclusion. 55hus, regtech has the potential to support regulators in the supervision of digital payments system in order to maintain financial stability and integrity.In this context, regtech promotes financial inclusion by providing real-time information and data to map financial access and usage to identify gaps in provision. 56As a result, the employment of regtech to design better financial and regulatory systems to achieve policy objectives facilitates inclusionary services for the underbanked through open access to digital data and wide offer of credit products. 57

II. The transition to data-driven finance
Modern data technologies have the potential to provide regulatory authorities with nearcomplete oversight of prudential and conduct risks, both for individual institutions and at the systemic level.One particularly vivid expression of this point of view envisages using technology for the real-time tracking of the global flow of funds. 58The principal motivation for this broad interpretation of regtech is macroprudential.The global financial crisis strongly reinforced the need for a sufficient degree of system-level oversight to work in addition to the 'microprudential oversight' carried out at the individual-firm level. 59This suggests a new regime of 'smart' regulation that would harness the transition to data-driven finance in order to allow for something closer to full, real-time oversight, and thereby answer the call of analysts who expressed the desire for such capability in their visions for a safer post-crisis financial system. 60Hildebrandt argues that smart regulation should identify a code-driven system of cryptographic law capable of taking decisions which affect legal subjects. 61There are other advocates of this broader interpretation of regtech.Zetzsche and others emphasise the transformative potential of regtech as being where its real 'prize' may lie. 62Kavassallis and others evidence the potential improvements of risk monitoring where digital standardised documents are made available to all relevant parties, including the supervisory and regulatory authorities. 63Butler and O'Brien note the transformative potential of such practices for the supervision of the financial system, although greater international harmonisation of regulatory regimes is likely to be required in order fully to harness this wider potential. 64 comprehensive market-wide information system in which the regulator sees and responds to every undesired development is not something that is imminently achievable.Rather, the opportunity lies in using technology to improve institutional operations and governance.Technology should be applied to strengthen both management information (making it available, understandable and actionable) and the governance of systems and data within institutions, in order better to achieve both business and regulatory outcomes.This approach to regtech, emphasising its use to facilitate improved data and information availability and hence improve governance and controls, will allow it to be applied to the full range of regulatory operations, not just to the enforcement of reporting and other rules but also to the principle-based approaches that cannot be directly translated into code.
As well as maintaining the present balance in the regulatory regime between rules and principles, this can help avoid other unwanted consequences such as the possibility that increased levels of automated and mechanistic compliance in turn will reduce emphasis on culture and values within financial institutions. 65This is above all a challenge of governance, for individual firms, for the financial services industry, and for the regulatory authorities.Technology is increasingly central to financial firms, so it is a board responsibility to oversee and ensure effective application of technology in operational and business processes.One challenge is overcoming the divides within firms, between specialists such as those in information technology and data science, and other staff with client facing and operational roles, by 'orchestrating' the adoption of technology so that business understanding and control are maintained.Another challenge is ensuring that senior management and board members have an adequate grasp of both technological opportunities and technological risks.
Achieving the potential of technology requires an unfamiliar degree of co-operation between financial services firms, on data and technology standards, on the sharing of data and on exploring opportunities for shared processing, all pursued with a view to achieving improved long-term outcomes for clients and other stakeholders. 66This may however meet with resistance from managements that have been routinely focused on short-term profit performance.Regulators have a central role going beyond their traditional mandates of oversight from a distance, intervening only when there is an imminent prudential or conduct threat.They will need to take some responsibility for co-ordinating the technological developments and engaging in ongoing dialogue with regulated firms about the most effective means of employing technology to achieve both business and regulatory outcomes.

D. Financial inclusion and digital payments systems
The use of digital payments platforms has exploited the potential of inclusionary services for market participants while increasing the employment of artificial systems to automate decision-making processes.Promises of financial technologies have been lauded in various quarters as advancing both opportunities for regulatory innovation and transaction costs savings. 67Technology-based payments have become instrumental in promoting small business lending, access to financial services at fair pricing for customers with disabilities and credit for low-income households.The innovation brought about by M-Pesa in Kenya first and then also in Uganda is often cited as integrating unbanked customers to mainstream financial systems. 68The possibility for technology-centred change in the financial industry is illustrated, for example, by the case of China which has seen rapid shifts to both mobile payments (Alipay, WeChat Pay) largely displacing notes and coins in urban areas; and to non-bank loan intermediation through the dramatic growth in the Chinese version of P2P lending. 69he adoption of automated procedures to support manual decisions embeds sophisticated computational techniques driven by algorithms that elaborate the flow of information received from autonomous predictive models.Despite the positive aspects of digital technologies, a growing debate about the data-gathering and the datasets used to elaborate inputs into computerised programmes raises questions about the accuracy of outcomes for final users. 70The algorithmic codes formulate a series of statements which might reflect the behaviour and routine habits of consumers.However, algorithms may contain biased methodologies inherent because of human error which are apt to amplify undesirable practices, such as the marginalisation and exclusion of protected customers. 71Further, algorithms are not objective: they do not provide accurate predictions of desired outcomes because they cannot possess the necessary accurate data.As Sunstein argued, 'algorithms may lack information that human beings have, and for that reason, some human beings might be able to outperform algorithms'. 72Poor training data and discriminatory proxies taint decision-making systems and exploit consumers' behaviour, although it is debated whether the discrimination is intentional or it is a result of mere defective software. 73Intelligent machines are trained by gathering data which may not necessarily be representative of all classes of consumers considered by the software so as to provide the full range of services.The training data cannot offer a complete reflection of individuals' profiles simply because it is unlikely to incorporate information about inscrutable factual criteria (e.g.financial condition, sexual orientation, level of disability). 74The reliance on training data has been questioned in terms of poor disclosure and the opaque methods by which the 'black-box' of predictive judgement design is embedded in the internal models. 75Machine learning analysis can produce wrong assessments of protected categories which replicate manual decisions and lead to the redlining of applicants belonging to the category of disadvantaged customers.Therefore, the quality of financial services available to vulnerable persons is limited in terms of what is on offer on account of constraints affecting the sources of information and unfamiliarity with the operating procedures of artificial systems. 76 range of new applications, often based on the transactions data made available through the experience of Open Banking, offers support for savings and decisions 77 , along with programmes making automated investment decisions and robo-advisors proffering investment advice. 78Technology underpins the growth of comparison sites and their use as an alternative to advice from brokers or the media in making saving and investment decisions and is increasingly used in both passive and active investment vehicles.The massification of data and advent of Open Finance create a new environment of inclusionary services for underserved customers. 79Automated market making and trade execution maintains the net asset value of exchange traded funds against their benchmark and the index tracking of passive mutual funds.A range of active investment funds are increasingly using machine learning and other technology-based decision making in their portfolio allocations alongside management judgements. 80In this context, new technology-supported asset classes are emerging, including loan-based crowdfunding and crypto asset-based payment systems, such as the 'Libra' initiative launched by Facebook, which operate as crypto stablecoin in smart contract platforms. 81Cryptoassets have the potential to overcome the current barriers of decentralised blockchain networks but constitute a challenge to policymakers, regulators and stakeholders because of the risk of altering the money market. 82They operate outside the central banks' arena, bringing an innovative mechanism of lending and free money transfers: this would certainly be a new frontier for cross-border payment systems. 83In this context, open banking and regtech relate to decentralised provision of financial services which employs technology to provide market participants access and control of the data.As Zetzsche and others argued, 'financial inclusion […] comes from the decentralization of finance enabling the embedding of local compliance standards and customs which tend to reduce costs of access to financial services'. 84Decentralized Finance (DeFI) has the potential to reduce intermediary costs, increase transparency through blockchain-based records, provide round-the-clock access to financial markets, expedite settlement transactions, and increase financial inclusion by allowing anyone globally with an internet connection to access DeFi platforms. 85For example, cryptoassets could have significant potential inclusionary services for remittances by foreign workers which are subject to excessive transfer fees and use obsolete technology. 86Using technology to enable speedy transfer of remittances at nearly zero costs (including as a major innovation FX conversion) in an environment that is safe from external threats can have an appreciable impact on financial inclusion, especially where remittances are an important part of a family's annual income. 87Through the use of digital ID and other identification techniques it can secure access to the unbanked giving the opportunity to keep the bulk of the remittances in safe storage.Currently, remittances by migrant workers are subject to high transfer fees.Whether a social media company such as Facebookwhose leadership has identified the need for efficiency in cross-border transfers and retail remittancescan improve the lives of migrant workers and the families that those workers support back in their home countries is an open question. 88afe storage of savings emanating from the remittances, once a part of them has gone into consumption, gives poor households and the previously unbanked the possibility to receive stable and predictable returns on savings, which would allow for better planning of the households' consumption and investment needs.Then, the transfer of some of the balances into a savings account would allow the very poor and the unbanked to use some of the funds to buy insurance to cover the impact on earnings of health and other contingencies (e.g. a bad harvest).Moreover, turning part of the remittances into savings in a seamless process constrains consumption for instant gratification and can boost the long-term investment plans of low-income households. 89Further, carefully planned savings balances may eventually be used for the purpose of human capital development including private investment on education.Finally, cash balances can be used as collateral to enable very poor households and the previously unbanked to acquire capital assets such as machinery, which can boost the productivity and income of a small business.However, a number of issues related to financial stability, privacy considerations and compliance with money laundering and countering the financing of terrorism rules remains a potential stumbling block for DeFI platforms. 90inancial services firms are increasingly automating credit and insurance risk assessment, fraud detection, and other financial services processes.There is a shift from more mechanical rules to algorithms that utilise a wide range of data sources ('big data') and software that can update itself and learn from its own performance; this is one definition of artificial intelligence, software that learns from data rather than having all rules pre-programmed. 91Automated tools are increasingly used in credit scoring and in loan and insurance origination, substantially lowering costs and offering potential improvements in risk measurement and management. 92hese technological innovations prompt questions about oversight and transparency. 93otable episodes of undesired outcomes of algorithmic systems are found in the biased pricing of credit or insurance products, based not on actual risk of loss but other customer characteristics such as age, religion or ethnicity. 94Discriminatory results of sophisticated algorithms employed to assess the creditworthiness of customers led to evident bias against 'people of colour' in the widely used Fair Isaac Corporation (FICO) credit scoring in the United States, perpetuating a dual credit system. 95The key question is therefore not whether a new technology is biased or not (all processes contain some bias), but rather ensuring it is adopted in a way that is consistent with desired regulatory outcomes.The adoption of any new technologies should therefore be expected to reduce bias, alongside improvements in process efficiency. 96This should not only be articulated as a regulatory principle, but also supported by dialogue between regulators and industry; in the case of FICO, it should explore the development of a more sophisticated credit scoring assessment more accurately reflecting default risk, with less bias. 97nother significant example of undesired outcomes of automated decision-making process is found in the UK Post Office scandal, a case of faulty accounting software which caused reporting shortfalls and resulted in postmasters being wrongly convicted as a result of computer errors. 98The Post Office's sophisticated IT system, the Horizon software, was at the epicentre of false statements about incorrect missing money from branch accounts. 99Defective data used to feed cloud-based machines produced inaccurate information about the financial profile of employees.Poor management decisions and failures of investigation and disclosure were involved, calling in question the reliability of technology products and a big tech company supplier. 100This also evidenced a lack of monitoring of the dataset and predictive models, which employed flawed variables elaborated in the software. 101The transparency and the fairness of the computer methodologies are the crux of the matter for the quality of the output.Trustworthy in mechanical instruments may hide practices that are designed (intentionally or in self-interested unfamiliarity) to perpetuate biased procedures: the experience of the British Post Office revealed an inherent dependency on the vulnerability of automated programmes. 102he potential benefits of digital payments system should be weighed carefully against the risks. 103In the long term, there is scope for AI to be used as a service-provider for network interconnectedness and as a tool to monitor the business conduct of financial institutions. 104Although machines can be more reliable than humans, new risks can be built into systems.The question arises as to whether the use of AI automatically excludes certain classes of consumers. 105The risk of discrimination is also associated with the disruption of computer programmes, which cause lack of transparency and unfair practices. 106Weaknesses in managing and interpreting the amount of data provided by machines constitute a barrier to the application of AI in the financial sector.

E. Conclusion
The range of digital technologies used in financial services is very broad, including household and small business lending, online and mobile payments, insurance, capital market transactions, wealth management and regulatory reporting and compliance.Likewise, a wide range of digital initiatives seek to promote financial inclusion.With the advent of digital payments systems, lenders have the possibility to access more information in order to assess the credit quality of borrowers and to make decisions on whether (and how much) to lend more quickly.Promoting technological innovation through alternatives to credit means integrating vulnerable and excluded customers into mainstream financial systems.
The use of algorithms and computer systems to coordinate supervisory authorities has become a key component in facilitating the delivery of transactions and improving firm's operational processes, although it can pose risks of undesired biases. 107In this context, regtech can facilitate compliance processes, while ensuring information disclosure and contractual certainty and predictability of enforcement actions in order to avoid poor outcomes for customers. 108The 'narrow' interpretation of regtech seeks to automate compliance and hence reduce its costs; while the broad interpretation presents an ambitious vision of comprehensive regulatory oversight that anticipates and prevents undesired outcomes.The central challenge of regtech is not simply 'automated regulation', but rather establishing the appropriate governance of technology in financial services, involving regulators, regulated firms, and technology suppliers.
Regulators employ technology to monitor compliance requirements and prevent suspicious activities (e.g.cybercrime, money laundering, fraudulent transactions).However, the adoption of intelligent machines raises concerns about the appropriateness of supervisory authorities' policies to ensure that making data machine-readable entails accurate control and permissioning for access to and use of data which can reduce regulatory burdens and transaction costs. 109n automated decision-making process can enhance the governance of data: it encompasses risk management and minimises reputational risk, legal risk and operational risk. 110Furthermore, automated procedures can expedite real-time information, particularly with respect to the asset quality of banks' balance sheets (automatic credit scoring) and for compliance with rules on conduct (e.g.mis-selling or rogue trading). 111Automated practices such as sandboxes experiment with fintech products through natural language processing and cognitive computing in order to secure compliance with regulatory process and supervision. 112echnological applications in banking services have attracted much attention from stakeholders and regulators, both because of the perception that they should support domestic capacity in what is a nascent and rapidly growing new industry with potential global impact, and because digital technology can address some of the perceived shortcomings of the traditional financial services industry (e.g.lack of consumer protection, weaknesses in governance, gaps in compliance and improved provision to previously underserved regions).The UK Government has launched a Fintech Sector strategy, which includes the formation of a 'Cryptoassets Taskforce' with the aim of positioning the UK at the forefront of harnessing the potential benefits of the underlying technology, while guarding against potential risks.113Supervisory authorities are also taking steps to support innovation with a leading role played by 'Project Innovate' and regulatory sandboxes programmes at the FCA, which allow automated machines to reduce the manual intervention of regulators. 114Banks are adapting to this innovation and technology offers opportunities and risks to payment intermediation.
The alternative provision of payments promotes financial inclusion globally, widening access to banking and insurance services both for vulnerable households and small businesses.However, financial services innovation (cryptographic security, massive data processing, distributed computing, artificial intelligence) and new software solutions for delivering financial services (crowdfunding platforms, cryptocurrencies, blockchain) require constant dialogue between regulators and regulated institutions on the appropriate design of regulation and its technological implementation.This dialogue needs to align itself with the ongoing widespread digital transformation of the operational processes, business organisation and market structure across technology products.The increasing mainstreaming of the cryptoasset industry provides customers access to decentralised forms of finance, particularly in relation to remittances, although it has the potential to alter the global payments system while, in turn, potentially leading to market disruption.