2020's FinTech policy cycle started relatively quietly today with the release of a research report from The first full week of 2020 brings an interesting set of FinTech data from De Nederlandsche Bank (the Dutch Central Bank): Working Paper No. 663 "The Economic Forces Driving Fintech Adoption Across Countries."
The data in the report provide a solid overview of the global scale of the FinTech industry spoiler alert: adoption rates are lumpy and not as economically significant as industry promoters would want). Second, the data are used to provide the foundation for arguing that increased regulation may be needed in the sector. In the end, the report raises more questions than it answers.
Before we start, it is important to note that research working papers do NOT represent the policy view of any official sector entity. Observations and policy conclusions are NOT actual proposals. Research papers nonetheless provide insight into the direction of thinking.
Data – Significant Scale, but still small
Drawing from a diverse set of data sets and research, the research paper generates some unsurprising payments statistics: The majority of market penetration by “Big Tech” is in the payments space (16% of GDP in China and other emerging markets)…..but less than one percent in the United States, India and Brazil.
In the credit intermediation space, however, the United States and the UK lead with adoption rates. In 2016, online marketplace and peer-to-peer lending accounting for 8% of all new mortgage lending in the United States. In 2019, they account for an impressive 38% of unsecured personal lending in the United States. 15.1% of all U.S. small businesses (and 6.3% of all UK small businesses) received credit from FinTech lenders in 2018.
The Dutch central bank indicates that high transaction costs and low competition (not regulatory policy) are the main drivers for these high adoption rates. In fairness, researchers attempted to find direct correlations between either monteary policy (low interest rates) or high macroprudential regulatory requirements but in the end indicated that insuffficient evidence yet exists to draw a clear conclusion.
Within emerging markets, age also plays a significant role with high usage rates prevalent among younger members of the population. However, in the United States and the UK, ironically over 50% of all borrowers using FinTech platforms were older (GenX, baby boomers).
Sadly, the paper missed as many data points as it provided. Particular missing puzzle pieces include:
Any data from the European Central Bank, much less insight into whether the ECB’s early experience with instant payments is similar to, or differs from, payments patterns in other high-adoption countries like Kenya and China.
Most problematically, the paper indicates that in 2017 total FinTech credit only amounted to 0.14% of all global financial system assets. If the sector is that small, can it possibly pose any systemic risk at the global level? More importantly, what was the growth rate? How much larger was the sector in 2019? And did the distribution of activity remain concentrated in specific national and municipal pockets? The paper inspires more questions than it answers.
The Dutch central bank research report draws conclusions concerning regulatory policy in one paragraph at the end of the paper….without showing data or analysis to support the conclusion. The rest of this blog raises the most important questions for discussion.
Consider first the paragraph in its entirety:
“Third, while FinTech innovations can sometimes overcome specific market failures (e.g. by reducing information asymmetries, transaction costs, etc.) FinTech activities will remain subject to the same well-known risks traditionally present in finance. For instance, deposit-like activities remain subject to liquidity mismatch and the potential for bank runs, even when they are offered by non-banks. New financial assets can still be subject to speculative bubbles, as was the case with Bitcoin in 2017-8.26 If specific FinTech or BigTech firms achieve a large enough scale, there is the potential for them to become systemically important (“too-big-to-fail”), resulting in moral hazard and excessive risk taking.27 Finally, new forms of interconnectedness, including operational dependencies (such as reliance on third-party services such as cloud computing) could transmit market shocks across institutions and markets. Managing these risks will remain the remit of public sector authorities. Supervisors must continue to adapt regulatory frameworks and crisis management tools accordingly.”
Few would contest that Bitcoin in 2017 exhibited classic asset bubble characteristics. But where was the moral hazard? The classic definition of moral hazard involves incentives that encourage risky behavior because the actor knows that the costs of risky behavior will be borne by a second or a third party. Moral hazard helps explain why elaborate disclosure rules exist in securities markets because issuers always have an informational advantage relative to asset purchasers. But in the BitCoin example the asset itself is created through automated problem-solving (called “mining”). Intermediaries buying and selling the asset – and initial coin offering (ICO) issuers) remain subject to securities disclosure rules as various securities regulators made clear from 2016 to the present.
Are FinTech deposits really subject to the same liquidity mismatch risks as classic bank deposits? This is a topic worth its own stand-alone research paper. Depositors at banks are subject to liquidity mismatch risks because their deposits are lent out by banks. It is far from clear that most FinTech lenders perform this intermediation service. Most FinTech lenders do not perform credit allocation services on behalf of depositors. They merely provide a platform for buyers and sellers of credit to meet and agree terms for exchange. While some FinTech payment providers may see users retain cash/assets in their accounts for long periods of time, those payment providers typically do not have the authority or the business model that would enable them to mobilize those customer accounts for lending.
If every payment company customer sought return of their cash balances on the same day, it would not generate a “bank run” in the classic sense because all the cash should still be in the custodian accounts. Of course, widespread redemption could generate significant technical stress on the system. But that is called “operational risk” and is an entirely different kind of regulatory concern.
Then we have the chain reaction risks (forgive the pun) that dominated the attention of international financial regulators during the last two quarters of 2019. “Interconnectedness” and “systemic risk” are synonyms in the regulatory lexicon. Usually they are also the foundation for regulatory capital and other oversight requirements. Do systemic risks operate in the same manner within FinTech firms as within traditional financial institutions? Again, this is worth an entire separate paper.
Many in the industry will argue that blockchain-based finance eliminates many systemic risks by automatically insulating each component of the chain from the other. Perhaps. But it also possible that insufficient funds in one part of the chain triggers separate automated actions in other parts of the chain that could deliver unexpected or different kinds of stress events. High levels of reliance on a small number of key infrastructure providers (e.g., cloud computing companies) does create exposures to operational risks associated with infrastructure breakdowns, compromise, and abuse of market power….but these familiar risks take on new forms as finance shifts towards increased automation.
The research paper released today by the Dutch Central Bank starts strong with a review of data points which, in the end, indicate the industry has a great deal of growth ahead of it. We are just at the front end of understanding why and how FinTech adoption rates are so lumpy. This report does a good job of pointing out some key peaks and valleys that require further examination. But it indirectly also illustrates the stunning lack of solid cross-border and cross-sectoral data that have to provide the foundation for policymaking.
Last year, we saw the Financial Stability Board trigger a sustained pivot towards Big Tech regulation starting in February (long before the Libra proposal). Those efforts gained momentum throughout the year and across multiple major meetings including, impressively, both the G7 and G20 summits. If this research report provides any indication, policymakers face a daunting task in identifying which specific risks they want to address as they explore how (not whether) to expand the regulatory perimeter to non-bank technology companies.