Journal of Financial Economics

REF 135 Technologies and Changes in the Financial Sector

In this article, we discuss the contribution of machine learning techniques and new data sources (new data) to credit-risk modelling. Credit scoring was historically one of the first fields of application of machine learning techniques. Today, these techniques permit to exploit new sources of data made available by the digitalization of customer relationships and social networks. The combination of the emergence of new methodologies and new data has structurally changed the credit industry and favored the emergence of new players. First, we analyse the incremental contribution of machine learning techniques per se. We show that they lead to significant productivity gains but that the forecasting improvement remains modest. Second, we quantify the contribution of the "datadiversity", whether or not these new data are exploited through machine learning. It appears that some of these data contain weak signals that significantly improve the quality of the assessment of borrowers' creditworthiness. At the microeconomic level, these new approaches promote financial inclusion and access to credit for the most vulnerable borrowers. However, machine learning applied to these data can also lead to severe biases and discrimination.

Financial services, including asset management, like all economic sectors, will be profoundly transformed by the adoption of digital technologies by 2025 for the benefit of customers and stakeholders. The French asset management industry is a key industry of the economy. Alongside banks, it contributes to finance the different stages of entreprises. This industry is composed by around 650 companies which manage more than 4,000 billion euros and represent nearly 100,000 jobs. They face important strategic issues that have a negative impact on their revenue growth and on their margins. How to answer it? Invest in digital technologies. Artificial intelligence tools make it possible to improve the funds performance, to better know customers or to capture new ones, to improve business processes, to reduce the number of intermediaries... However, their penetration will require structural changes on data management and on the management of human resources.

Despite the recent successes of artificial intelligence (AI), it is not a new field. Tools provided by data sciences have been at the heart of a series of progresses allowing machines to "solve complex problems, without being intelligent". Secondary innovations coming for AI will transform the financial industry in three directions: towards client experience and construction of bespoke products on the fly, towards real economy and nowcasting, and towards risk management. These innovations are already impacting market participants that are modularising their services and re-engineering themselves into platforms. Since this sector has good reasons to be highly regulated, some specificities of AI, i.e. the prominent role of data and the use of external software libraries, can appear as new sources of uncertainty.

With 965 billion USD of revenues in 2018, Global Transaction Banking is a keystone of the banking industry. Three trends, which reinforce each other, are at play in today's market landscape: the rise of non-traditional players and business models, technological innovation that is unprecedent in terms of rhythm, and a shift towards finer and more sophisticated client needs-based segmentation. In the next five to ten years, technology will continue to be a key force and challenge, with levels of innovation and investments – particularly with respect to digital channels, APIs, artificial intelligence and blockchain – that should have a significant and relatively fast impact on the industry boundaries and structure.

Digitalization of payment has rapidly generalized in China from its emergence in the early 2010 decade. Its uses applications available on smartphones, provided for the most by two digital services providers, Alipay and WeChat Pay. This article relates the emergence of these services in China and tries to elucidate the reasons of their rapid adoption. It uses the analytical tools of industrial economics of finance to understand why the two main companies involved in this success did not compete so much to acquire a definitive leadership. It draws the lines of the resulting consequences of their emergence on the financial intermediation in China and presents the attitude and actions of control and regulatory institutions. It considers at end the possible extensions of the model outside China.

This paper draws a distinction between wholesale CBDC (WCBDC), accessible only to financial intermediaries, and retail CBDC (RCBDC), accessible to the general public. The issuance of one could be dissociated from the other, implying the possibility of one, two, or no CBDC(s). The paper considers the possible motives, modalities, and consequences of central bank digital currency (CBDC) issuance in both forms for the economy, the financial system, monetary policy and financial stability. The issuance of a CBDC would represent a supply shock, which would support economic growth in the medium to long run and could transitorily weigh on prices. Furthermore, the issuance of a RCBDC could put a floor to bank deposit rates and, if it is remunerated, raise them. If the RCBDC were not remunerated, the effective lower bound would be raised to zero and the effectiveness of asset purchases by the central bank could be diminished. If it were, the interest and exchange rate channels should be strengthened. The remuneration of the RCBDC would thus seem to create a trade-off between the effectiveness of monetary policy and the cost of bank intermediation.

A recent phenomenon, initial coin offerings (ICOs) are a fundraising method that is still emerging but is beginning to take shape, allowing the emergence and the financing of new companies in technological and innovative sectors. The analysis of the global market highlights the still marginal nature of this method of financing, but also the creation of a new form of commitment in a business project that attracts investors whose number now exceeds the closed circle of specialists. The analysis of French ICO projects, using an original database, highlights the growing interest of this fundraising mechanism by some project initiators, who seem to welcome the opportunity to call on a community of international investors and introduce this method along with other, more traditional methods of financing. In view of the risks associated with this type of investment, the article finally presents the French incentive framework, the optional visa for token issues issued by the Autorité des marchés financiers (AMF).

In the field of insurance, the substitution of Smart Contracts for traditional contracts offers obvious advantages: transparency of benefits and premiums, speed of compensation, reduction of management costs, security of operations. However, insurance companies face three types of challenges for their implementation: constructing the covered event and verifying the attractiveness of the insured for this mode of coverage in an automated framework, finding and managing the underlying data essential to the Smart Contract, and organizing a profound change in the management of their activity and organization. Insurers will find it difficult not to apply this technology, at least to parts of their portfolio of contracts.

NeoBanks bring together a diverse mix of new financial intermediaries, offering online or mobile banking applications. Their characteristics have evolved since their appearance in the 2000s as well as the nature of their relationship with traditional banks. However, their profitability targets have, so far, hardly been met partly because of the difficulty and the cost of acquiring new customers. Is the new generation of neobanks, 100% digital and mainly accessible by mobile applications, likely to disrupt the banking sector?

The paper investigates how Fintech can influence the structure of banking markets. Fintech have several competitive advantages relative to traditional banks. They have lower costs thanks to the use of information technologies at the core of their business model. They can supply services with higher quality like a greater reactivity facing the demand of a credit. There are however limits to the impact of Fintech. Traditional banks can have an advantage for financial costs because of their deposit-collection activity. The technological advantage of Fintech can be more or less important based on the fact that private information on the borrower plays a major role on the credit market.

Lately, the longevity of the traditional banking model has been more acutely put into question. Because of the joint context of, on the one hand, the major technological transformations which foster the emergence of new models, such as “neobanks” exclusively on line, fintechs or GAFAs which are coming into the banking sector and, on the other hand, very unfavorable financial conditions. Thus there is currently a strong belief that the only way to get out of this difficult situation would be a defensive retract. While returning to basics of retail banking activity, the aim of this article is to demonstrate that actually, banks have major assets that can be highlighted in regards to the new emerging models. Within this context, in an offensive vision of their activity, banks have to give more and more added value to their clients in a comprehensive strengthened relationship model. This implies major investments in training as well as digital, but also building alliances with new partners of the banking sector, where appropriate.

The goal of this article is to analyze the consequences of the emergence of the Fintech on the forms of financial intermediation. The idea that Fintech is likely to favor decentralized forms of financing (like those characteristic of the financial markets) is discussed. Financing platforms can be analyzed as forms of re-intermediation since the role of banks in this market is very important. Balance sheet intermediaries have important features that will not be challenged by the Fintech and traditional forms of banking will not disappear. On the other hand, the Fintech are likely to restructure the banking market because of the Bigtech's entry on this market. Bigtech are indeed able to exploit new economies of scope generated by their electronic commerce activity. Finally, the banks' capacities to cope with these evolutions are studied.

Digital, environmental and social transitions upset financial institutions financial risks in a Schumpeterian process. These upheavals involve exogenous factors, stemming from institutions' counterparts, from systemic sources, responses from national and supranational public authorities to the effects of transitions, and endogenous, linked to competitors less hampered by risk inheritance and the systemic importance of their decisions.

These transformations call into question current risk management methods fundamental principles, and their reliance on stationarity of economic cycles. They advocate for the creation of new tools, approaches, and algorithms, including the introduction of new methodologies to collect, structure, analyze and predict from large amounts of data. The results of generalist's tests applied to natural language understanding algorithms, computer vision and deep reasoning, are reaching sufficiently satisfactory levels to test their employability in the financial sector. However, this transition must follow a progressive approach factoring the economic stakes of the large financial institutions and their risk inheritance.

This publication identifies the key issues for financial risks arising from transitions. It explores three application areas of machine learning and deep learning technologies, to meet the new challenges of financial risks. Finally, it concludes on the research axes to adapt these new technological tools to the specificities of financial institutions to make the transitions successful and inclusive.

New risks arise from the digital transformation. With the increase of cyber threats, insurance contracts appear as fundamental tools to improve the resilience of society. While the cyber insurance market is expanding, quantification of the economic impact of this risk is still blurred. It requires sophisticated stochastic models - to capture the complexity of the involved phenomena and their high volatility, for example the accumulation risk - while taking the weakness of available data into account. In this article, we emphasize the difficulties to collect accurate data, and the impact of behavior of the agents on risk evaluation. We also mention potential methodological paths to overcome these issues.

The structure of national accounts clearly indicates that the annual flow of household savings has three allocations: loan repayments, personal contribution for real estate investments and financial investments. Unfortunately, this decomposition of the flow of savings is not now documented. The main reason for this surprising ignorance is that, for the financing of these different operations, households do not only resort to credit and current savings, but also to prior savings. This last resource corresponds to one of these wealth management operations, the knowledge of which can only result from specific surveys, which are now non-existent.

For a country like France, based on several conjectures, we observe that this allocation of the flow of savings could be variable according to the characteristics of the year considered (level of the savings rate, importance of the recourse to credit ). However, the share of financial investments appears to be high and stable (around half of the total flow). Refunds would be large, but unstable (between 25 and 40%). The share of personal contributions would also be unstable, but at a much more modest level (less than 20%). For countries such as the United States or the United Kingdom, the share of repayments in current savings could be much lower, due to late repayments, largely relying on pre-savings. This could be one of the reasons for the permanent weakness of their household savings rate compared with countries like France.

Using a sample of 26 countries over the period 1990 to 2013, we empirically examine the correlation between the size of the shadow banking system and the size of traditional financial institutions, namely banks and institutional investors. First of all, we conduct a survey of the entire sample as well as the countries of the euro area and then a review country-by-country. The question is whether shadow finance and conventional finance should be seen as complements or substitutes. Our results tend to show that the shadow banking system complements rather than replaces the activities of traditional financial institutions, including banks. Bank sponsorship of shadow banking activities has clearly reinforced this link.