FinTech hypes abound. In the news, FinTech is “disruptive”, “revolutionary” and armed with “digital weapons”, that will “tear down” barriers and traditional financial institutions (World Economic Forum, 2017). Although FinTech (see the Box below for a definition) has been expanding very rapidly in financial markets, as documented in the Numbers section of this Journal, the jury is still out, and their potential impact on banks and financial institutions, as we know them today, is far from clear yet. The tension between stability and competition underlies the whole debate on FinTech and on how to regulate it. The crucial question is whether and how far FinTechs are replacing banks and other incumbent financial institutions. And whether, in doing so, they will induce a healthy competitive process, enhancing efficiency in a market with high entry barriers, or rather cause disruption and financial instability. This issue of European Economy deals especially with the relationship between FinTechs and banks. Our bottom line is that FinTechs enhance competition in financial markets, provide services that traditional financial institutions do less efficiently or do not do at all, and widen the pool of users of such services. But they will not replace banks in most of their key functions. In most cases, FinTechs provide a more efficient way to do the same old things. Yet banks are well placed to adopt technological innovations, and do the old things in the new way themselves.
FinTechs provide indeed the same services as banks, possibly more efficiently because of technologies, but in a different and unbundled way. For example, like banks, crowdfunding platforms transform savings into loans and investments. Yet, differently from banks, the information they use is based on big data not on long term relationships; access to services is only decentralized through internet platforms; risk and maturity transformation is not carried out; lenders and borrowers or investors and investment opportunities are matched directly. There is disintermediation in this case. These are pure FinTech activities. However, these pure FinTech unbundled activities have limited scope. For example, it is difficult for platforms to offer to their clients diversified investment opportunities without keeping part of the risk on their books, or otherwise securitizing loan portfolios. Other functions carried out through FinTechs instead of banks, like payment systems (e.g. Apple pay instead of credit cards) are still supported by banks. Banks loose part of their margins, but still keep the final interface with their clients, and because of the efficiency of these new systems, they expand their range of activities. Hence, in this case, there may be strong complementarities between banks and FinTechs. The value chain of banks includes many bundled services and activities. FinTech generally carry out one or few of these activities in an unbundled way. Yet, bundling provides powerful economies of scope. The economics of banking is precisely based on, and has a strong rationale in the ability of banks to bundle services like deposits, payments, lending etc. For this reason, FinTechs will also have to bundle several services if they wish to expand their activities (as for the crowdfunding example above) or integrate their services with those of banks (as for the payment systems above). The business model of FinTechs, therefore, is highly likely to gradually converges towards that of banks, as we will discuss in this editorial. As this happens, it is no longer clear that FinTechs have a neat competitive advantage on banks, besides for the legacy costs that banks face in reorganizing their business. Moreover, as FinTechs expand their range of activities, the scope for regulatory arbitrage will decline. We share Ferrarini’s view in this issue (see also the section on regulation of this editorial below), that a case by case regulatory approach should be implemented, essentially applying existing regulations on FinTechs, on the basis of the service they carry out. Regulation should be applied when services are carried out (of course with an element of proportionality), independently from which institution is carrying them out. For example, if we consider again loan based crowdfundig, different regulatory frameworks could be relevant, depending on what these platforms actually do. Banking regulation could be unnecessary, if platforms do not have the opaqueness of banks in transforming risks and maturities and do not keep such risks on their balance sheets, for example by collecting deposits and lending outside a peer-to-peer (P2P) framework. But it should be enforced if platforms carry out such activities. Once regulatory arbitrage is ruled out, and the same regulatory framework is imposed on all institutions on the basis of the functions they perform, the playing field is levelled. Then the only competitive advantage is the one granted by technology and the organization of activities. The framework becomes one of pure competition with technological innovation. Vives in this issue discusses several competitive options banks and FinTechs face. We also discuss this at length in section 3 of this editorial. Cases of entrants with digital innovations and their disruptive effects abound in sectors affected by digital technologies. Netflix caused the “bust” of Blockbuster and Amazon that of many retailers and booksellers. Skype took 40% of the international calls markets in less than ten years. For the incumbents the deadly mix of the newcomers was lower-costs higherefficiency with better or new products and services, “sprinkled” with incumbents’ inability to swiſtly account and adapt to the changing landscape. Although evocative, these examples do not fit precisely the financial industry. This is significantly different because banking is multiproduct, with largely heterogeneous customers, intrinsically plagued by asymmetric information and heavily regulated. Competition will enhance efficiency, bring in new players, but also strengthen the resilient incumbents, able to play the new game. Intermediation will keep being a crucial function of financial markets. Intermediation will partly be carried out in a different way than today: much more internet and internet platforms; much more processing of hard information through big data. But banks will note disappear. If some do, they will be replaced by other, more efficient ones. The real casualties will not be banking activities, but mostly small banks and banking jobs.
We will develop further our arguments in the rest of this editorial. In the next section, we will first discuss the key economic ingredients characterizing banking activities, and how they might be affected by FinTech, in particular with respect to maturity and risk transformation, payment systems and the management of information. In the following section, we will compare the revenue models and the incentive frameworks characterizing the activities of banks and FinTechs and how these may affect their competitive prospects. We will finally conclude with a discussion on regulation.
Fintech refers to the novel processes and products that become available for financial services thanks to digital technological advancements. More precisely, the Financial Stability Board defines fintech as “technologically enabled financial innovation that could result in new business models, applications, processes or products with an associated material effect on financial markets and institutions and the provision of financial services”. 2 The areas of actual and potential expansion of Fintech are: a) transactions execution (payments, clearing and settlement); b) funds management (deposit, lending, capital raising and investment management); and c) insurance. The ability to impact on essentially all the services typically offered by traditional financial institutions, such as banks, comes from cost reductions implied by digital technology advancements, improved and novel products for consumers and limited regulatory burden. More specifically, with technological advancements Fintech operators benefit from: i) lower costs of search that enable matching in financial markets more effectively, ii) economies of scale in collecting and manipulating large bunches of data, iii) cheaper and more secure transmission of information, iv) lower costs of verification.
A bank is an institution whose current operations consist in granting loans and receiving deposits from the public (Freixas and Rochet, 2008). This entails performing simultaneously three sets of activities: transmuting the characteristics of financial assets and liabilities, providing payment services, collecting and processing information (see Dermine, in this issue, for a thorough analysis of the impact of Fintech on many different financial services). Transmuting the characteristics of financial assets and liabilities is mainly realized through maturity transformation, that is the use of short-term funding to grant long-term loans. This function is crucial to any economic system, since it allows to fund long-term investments, and therefore foster productivity, while at the same time insuring depositors from idiosyncratic liquidity shocks. Because of their function in providing liquidity to their customers, banks are also well placed in offering payment services. Information processing includes all the activities related with the ex-ante screening of potential borrowers, with the ex-post monitoring of their behavior, and with the construction and management of a diversified portfolio that maximizes the return to risk ratio. Economies of scope justify the internalization and bundling of all these key services within the boundaries of one institutions. They provide a core rationale for banks’ existence. 3 We therefore need the perspective of each of these three key services to understand the impact of FinTech on the banking industry. We will discuss them in turns.
The first key task of banks is to transmute the characteristics of financial assets and liabilities, in particular through maturity transformation. Banks can exploit diversified large pools of small size depositors to cope with the impact of idiosyncratic liquidity shocks. Since it is unlikely that depositors unexpectedly need to withdraw their funds all at the same time, banks can set aside a limited buffer of liquid assets to grant longer-term loans. In synthesis, they can transform short-term sight deposits into long-term loans. This is the essence of banks’ ability to provide liquidity services (Bryant, 1980, and Diamond and Dybvig, 1983). Unless technological innovation will change significantly the way in which maturity transformation is performed, which seems to be unlikely, the interesting question from our perspective is whether FinTech companies can also provide liquidity management services. The answer is yes and no. Yes, because any FinTech company can raise funds and put them in a pool, from which its clients can make withdrawals when needed. No, because if they use these funds to grant illiquid loans or acquire less liquid assets, they would need a specific authorization. In particular, if FinTech companies raise deposit-like funds to grant illiquid loans, they would be acting by definition as banks, and as such they would be under a bank charter. Therefore, as long as banks are subject to fractional reserve requirements, they have an advantage in providing liquidity services to their depositors, because they can perform maturity transformation and earn the interest margin. In a way, FinTech companies providing liquidity services are like 100% reserve or “narrow” banks, with the possible additional weakness that their accounts are not as trustable as those of banks, since the latter are subject to much stricter regulations. 4 Moreover, bank depositors are also partly shielded by credit risk by capital buffers and deposit insurance. Also, the structure of bank liabilities gives a priority to the holders of some asset classes, such as depositors and bondholders. Clearly, all this is not the case with FinTech companies, that typically act as brokers, leaving on investors the credit risk of the loans that they grant (unless it is covered externally by an insurance company). While the new discipline on bank resolution increased significantly the share of credit risk on banks’ depositors and other creditors, this has been partly matched by the surge in capital requirements. Because of maturity transformation, banks also disentangle any loan that they grant from the funding of each single depositor, obtaining in this way a much better portfolio diversification than what a single depositor could achieve. In principle, any financial intermediary could perform these same activities. But while venture capital and private equity firms are typically much better than banks in screening and monitoring opaque entrepreneurs,they normally do not obtain the same degree of portfolio diversification. And while investment funds have very diversified portfolios, they typically invest only in listed shares. If they are based on a pure peer-to-peer (P2P) principle, crowdfunding lending and investment platforms cannot offer any diversification to investors. If they do, then platforms will have to take up part of their risk on their books (hence falling under banking regulation) or act as issuers of securities (then falling under security regulation). In fact, most of the match-makers currently adopt the so called “agency model”, where they do not retain the risk of the loan that they originate, do not interfere with its price, and receive compensation through commission fees, that may be paid by both sides of the transaction. It is therefore very likely that, for a given portfolio of assets, the riskiness of FinTech liabilities is higher than that of bank deposits and plain bonds. And that, due to their different incentives, FinTech companies are likely to have riskier asset portfolios than banks. Hence, as far as banks also adopt new information management technologies and regulatory arbitrage is ruled out, the threat to their business coming from this channel of credit and liquidity risk seems limited.
Banks’ ability to provide instruments for liquidity and risk management is very much related to their ability to supply payment services. There are obvious and strong economies of scope in providing at the same time liquidity and payment services: customers facing a liquidity need are much better off if they can make payments directly from their deposit account. This is the very reason why checks were originally introduced, and why ATMs and POS followed. Many payments can be made by transferring value across accounts, with no need of recurring to cash (or other central-bank liabilities, such as bank reserves). In most of cases the transfer is among banks. In fact, even many services that appear to be extremely innovative (e.g. Apple pay) are in fact technological devices that make it easier to transfer value across bank accounts. But a number of non-bank financial intermediaries are indeed emerging, from payment institutions to electronic money providers. In some countries,also non-financial corporation, such as telecommunication companies, are entering the market of payment services, exploiting their large base of customers. In China, for example, telecommunication companies have been offering for years payment services linked to deposit-like accounts, that oſten paid higher interest rates than deposits at commercial banks. However, to a large extent, this expansion was possible and profitable because of the limited diffusion of bank accounts in China (as in all other emerging economies, see Vives in this issue) and because bank deposits were subject to interest rate ceilings. In most countries, non-financial corporations are not allowed to pay interest rates on their accounts, and there are no interest rate ceilings on bank deposits. In general, if non-financial corporations or non-bank financial corporations can afford to compete with banks in payment services by paying higher interest rates on their accounts, there must be some kind of benefit that compensates for their inability to exploit the fractional reserve requirements and earn the interest margin granted by maturity transformation. We can think of at least three reasons. First, lighter regulatory requirements than banks. The role of telecommunication companies in China is the case in point. Second, better technologies, allowing non-banks to provide similar payment services at lower costs. Third, different and more effective economies of scope than those available to banks. The case of regulatory arbitrage is not particularly interesting, since it depends on the decisions of the authorities. As we argue below, the more FinTechs carry out activities similar to banks, and the more they become systemically relevant, the lower there should be room leſt to regulatory arbitrage. Technological advantage is relevant in the short-run, when new players can enter the market and exploit their better technologies. But there are no clear reasons why in the long-run banks should not be able to adopt the same technology as non-banks. Of course, banks face huge legacy costs that slow down their transition to new technologies. But as far as technological adoption is fast enough for banks not to lose their network economies, we should expect technological convergence. The presence of economies of scope between payment services and other business activities that banks cannot replicate is no doubt the most interesting issue. These economies of scope arise especially between providers of payment services and of other services typically affected by network externalities, like Amazon and Apple in consumption and Google in internet services. To generate these transactions not only the matching must be efficient, but it is necessary that both sides of the market, sellers and buyers, lenders and borrowers, are willing to be “on board”. In particular, there must be a sufficiently large pool of lenders ready to offer funding to borrowers. This is partly the reason why these IT giants are starting to offer payments and other banking services like (indirectly) consumer credit, where economies of scope are huge in information processing: purchases, payments and internet searches alike allow these companies to evaluate the credit risk of their clients possibly better than banks. Other economies of scope relate to the possibility of extending consumer credit to customers, thereby using interlinked pricing strategies. This is an old game, think for example at the consumer-credit banks owned by the sellers of durable goods like cars. With a customer base such that of Amazon, the potential of this activity expands immensely. Yet this connection is strictly restricted to consumer credit. And if companies start to do consumer credit, they will also face standard regulatory restrictions. Interestingly, Amazon’s store cards, that provide consumer credit, are issued by a consumer financial service bank, Synchrony, which also manages credit scoring and payments. There might be technological economies of scope to be exploited by linking electronic purchase platforms with electronic payment platforms. Paypal, the largest world supplier of electronic payment services, started its business as a linked service to E-bay, the electronic auction and sale site. However, Paypal offers today services far and above payments on E-bay, most of which are linked to bank or to credit card accounts (also linked to bank accounts). On top of that, Paypal has nowadays a bank charter. Finally, the digital ledger or blockchain technology would require a separate discussion, beyond the scope of this editorial. Yet we should at least mention digital payments with virtual currencies such as the Bitcoin. Blockchain technologies record any type of digitalized information permanently and virtually with no possibility of manipulation by anyone. This helps in building trust with no need for the State or notaries guaranteeing the actual legitimacy of the transaction and the ownership of the assets transferred. This certainty and certification of property rights greatly facilitates the commerce and exchanges, building on “automatic, machine based trust”. This is of course an option also to banks, not necessarily only to new comers.
The entire financial sector builds on information and information management. Recent developments in ICT have radically changed the way information is processed by financial institutions. As discussed by Bofondi and Gobbi in this issue, these developments have impacts in three different dimensions. First, data storage and processing, because cloud computing allows for the implementation of on demand high level computational capacity at fairly low cost. Second, data transfer through the internet at low cost. Third, data availability, because of the increasing digitalization of society and the economy. The type of information that financial institutions have and the way they use it to take their decisions is a crucial element to consider in discussing the potential impact of FinTech on banks. FinTechs function on big data and on the standardization of information. Banks, most of the time, on soſt and relationship based information. It is also clear that any advancement on information technology profoundly affects the financial sector: FinTech operators are modifying both production and distribution of financial services. As for the production of financial services, a huge mass of personal information is collected and analyzed nowadays. Large client bases and their past behavior allow to predict preferences, needs and trends and to offer the right financial product at the right moment and with the right price. Big data and machine learning are two key ingredients that are dramatically changing the landscape of financial services. Applications and effects on the industry are growing and will be pervasive, from screening potential borrowers to pricing risk at the individual level. Amazon Store Cards boast that they can take instant credit decisions in 15 seconds, something unthinkable through standard means of credit processing. Distribution of financial services is deeply affected as well, with new channels, personalization, flexibility and better matching. Here, the driver of change is the possibility to match different sides of the market easily and effectively. Online platforms allow sellers and buyers to screen for the best deals in a process that is more efficient the more populated is the other side of the market, a network externality. At the same time, large information bases allow FinTech provider to assess the characteristics of their customer to implement price discrimination policies. What matters most for financial applications of tools such as machine learning, big data and matching is the ability to recognize patterns quickly by digging in vast data set, an activity that would be virtually impossible for humans. The idea is not new as even standard regression econometric models can be seen as tools for pattern recognition. The novelty lies in the possibility to rely on extremely large set of data that are explored with more and more powerful computers and algorithms that explore, learn and identify patterns. There are however four key questions concerning the transition to hard information and big data. A first question is who owns and has access to the information. The screening capacity is based on hard information derived from huge data sets. As stated, the gist of digital innovations is the large amount of data for pattern recognition and the network externalities that are needed for matching and that non-linearly increase with the size of the network, for example in peerto-peer platforms. If the information is private, only platforms with large client base have a sufficient scale to have such data. And certainly, giant internet companies like Amazon have huge amount of data on which to base their analyses. Incumbent large traditional operators, such as banks, will also have an informational advantage. It is not clear how far these institutions make use of this information, how far the information itself is already digitalized and how far their ICT facilities allow banks to process this information. Whatever the case, incumbent banks can certainly build up large data bases at a lower cost than new entrants. Hence, new FinTech entrants will initially suffer from small scale. Naturally, also publicly available information can be used. And regulation may force private owner of information to disclose it to entrants. For example, the new Payments System Directive (PSD2) imposes to banks to release information on their clients’ accounts to other financial institutions, on request of clients themselves. Even though these prospects presently refer to deposits and current account conditions, they might be extended to credit performances and the assessment of borrowers. However, second question, the processing of hard information has huge legal and social implications, in terms of privacy, in terms of the mechanisms of reciprocal assessment in society, and of cyber risk. The more information and data on the financial behavior of individuals become public, the lower barriers to entry, but also the more we move away from a society where screening occurs through direct economic and social interactions. Understanding the implications of this pattern is beyond the scope of this editorial, but it is very likely that regularly restraints will be set up, specifying to what extent private information may become public and shared. In this respect, who has legitimate channels to be the holder of large volumes of private data (banks on their clients), will also keep being in an advantageous competitive position. A third question is how far hard data can fully replace soſt information. It is clear, for example, that relationship lending is hard to replace in granting credit to small-medium-enterprises (SMEs), with still fairly high degrees of opacity in their accounts and future business prospects. Or in evaluating large investments or loans, involving a large concentration of risk. This claim could partly be mitigated by the fact that one of the interesting applications of machine learning is the area of natural language recognition and interpretation. For example, the huge amount of lines of texts in social networks could be investigated to identify preferences, desires and attitudes. In the future, this may have very deep consequences in the banking sector as well. Relationship banking is built on human interactions between a loan officer and a prospective borrower. The former is meant to be able to interpret the behavior of the latter and give a meaning and a judgment to the borrower’s trustworthiness and other subjective matters. Improving pattern recognition with machine learning, for example in text and verbal communication, could complement (or perhaps replace) this human activity. Finally, the fourth question concerns the different incentives that banks and FinTechs have in processing information. Banks act as delegated monitors for their clients: they screen ex-ante the applications, by evaluating in detail the prospects of the potential borrowers and the value of the collateral that they may be posting; they monitor ex-post their performance along the whole duration of the lending relationship, possibly enforcing covenants capable of limiting losses in case of default (Diamond, 1984). Thus, they manage the credit risk of the investors, partly holding a share of it in their balance sheets. Moreover, the risk and maturity transformation function carried out by banks, and the inherent structural instability of their balance sheets (risk of bankruns), provide very strong incentives for better information collection and management than for non-bank institutions, that do not carry out such functions (Diamond and Rajan, 2001). Directly managing credit risk and incentives to collect information seem less strong for many FinTechs, where platforms have an originate and distribute function and do not keep risks on their balance sheet. This different structure may well involve high moral hazard and lower incentives for actively screen investments and monitor ex post performances.
Watch This : https://www.youtube.com/watch?v=Z5vxRC8dMvs