Executive summary

Finance is undergoing a profound transformation. Digital technologies are

reshaping payments, lending, insurance and wealth management – a process that
the COVID-19 pandemic has accelerated. While this is making financial services in
many economies more diverse, competitive, efficient, and inclusive, it may also
increase concentration in markets. Moreover, new risks may arise to a range of key
public policy goals. This paper draws on the underlying economics of financial
services and their industrial organization to examine – with recent empirical
evidence – the implications of digital innovation for market structure and attendant
policies, including financial and competition regulation.
The key organizing framework for the discussion is economic frictions such as
information asymmetries and economic forces such as economies of scale and
scope. These frictions and forces give rise to financial intermediaries and shape
market structure. We show that while technological advances are not new to
finance, digital innovation has brought major improvements in the connectivity of
systems and in computing power and cost, which have resulted in large volumes of
newly created and usable data. For example, mobile phone usage has surged
globally, social and economic activity has shifted online (often to platform-based
businesses), and new technologies like cloud computing have become widely
adopted.
These improvements have alleviated frictions, blurred firm and industry
boundaries, and given rise to new business models. New, often smaller and
specialized financial technology (fintech) players have unbundled services (see
definitions below). However, classic economic forces remain relevant. Economies of
scale and network effects are strong in digital platforms and cloud computing.
These scale effects, alongside economies of scope encourage re-bundling, and
allow large technology (big tech) firms and other new players to deepen their
inroads into core financial products. Available evidence shows that big tech firms in
particular are rapidly expanding their footprint in finance, and can use big data in
ways that reduce the need for collateral. Meanwhile, incumbent financial institutions
have adapted by adopting new technologies and disaggregating their production of
financial services to improve efficiency.
Digital innovation could drive a range of industrial organization outcomes. On
the one hand, digital technology enables niche providers to reach a target customer
base and be economically viable. On the other hand, customer acquisition, funding,
“assembly,” and switching costs tend to favour larger providers of digital financial
services. One possibility is a “barbell” outcome composed of a few large players and
many niche players. The large, multi-product players could include traditional
financial institutions, fintechs and big techs – thus both incumbents and new
entrants. Small players may include fintechs as well as geographically or sector
focused incumbents.
While a ‘‘barbell’’ is not the only potential outcome, it is a central case given the
economic forces at work. It is a potential steady-state market structure as some
participants leverage scale economies and network effects to grow larger, while
innovation continues to result in new entrants. There will be a tendency for players
to either hyper-focus or to aim for the large, multi-product space. However,
continued atomization, stalled re-aggregation, or limits on entry could result in a
different configuration.

v
 This analysis gives rise to important policy issues regarding competition,
regulatory perimeters, and ensuring a level playing field. Concentration risks may
increase in the provision of financial services to end-users, and in the provision of
infrastructure to financial institutions. Market structures that concentrate data and
supercharge network effects could reduce intermediation costs and broaden
inclusion. In many markets, however, the resulting market power might be seen as
detrimental. Competition regulators will have to strike a balance appropriate to the
needs of their markets, since different societies will attach different preferences to
market structure outcomes.
At the same time, financial regulatory authorities are working to manage policy
trade-offs among (i) stability and integrity, (ii) competition and efficiency, and (iii)
consumer protection and privacy. The barbell outcome, for example, could present
challenges in terms of stability with respect to both large and small payers.
Widespread access to data raises privacy concerns. Regulators need to balance the
innovation and efficiency brought by new entrants with the potential challenges for
oversight, enforcement and consumer protection. Emerging policy approaches –
such as new anti-trust rules for the digital era, data mobility requirements and data
protection laws – may help to mitigate the policy trade-offs. Yet the responsibility
for these changes generally lies with different public authorities, and with
legislatures. To navigate this new territory effectively, and to balance the necessary
policy goals, authorities will need to collaborate. This will need to occur both
domestically – with cooperation between central banks, financial sector regulators,
other industry regulators, and competition and data protection authorities – and
across borders. Such collaboration can help to ensure regulatory consistency and
peer learning within and between countries, and ultimately better development
outcomes for the country.

Definitions of key terms used in this paper

Digital financial services (DFS) are financial services which rely on digital technologies for their delivery and use by
consumers.

Fintech refers to digital technologies that have the potential to transform the provision of financial services
spurring the development of new – or modify existing – business models, applications, processes, and products. In
practice, the term “fintech” is also broadly used to denote the ongoing wave of new DFS. Examples of these
technologies include web, mobile, cloud services, machine learning, digital ID, and application programming
interfaces (APIs).

A fintech firm is one that specializes in offering DFS to consumers, or enables other providers to offer DFS.
While many of these companies are relatively new to the financial sector, others are by now well-established public
companies. Examples of fintechs include digital payment providers (eg PayPal), financial infrastructure/connectivity
providers (eg Plaid), digital insurers (eg BIMA, Policy Bazaar), peer-to-peer lending platforms (eg Afluenta, Funding
Circle, Investree).

A big tech firm is a large company whose primary activity is digital services. Examples of big techs include
online search engines, social media platforms, e-commerce platforms, ride-hailing platforms, and mobile network
operators. Numerous big techs have started to offer DFS, leveraging their large customer bases and the data they
have on transactions and activities that give rise to payments or a need for credit, insurance or other financial
services.

vi
1. Introduction

Digital innovation is transforming financial services. Innovations in financial

technology such as mobile money, peer-to-peer (P2P) or marketplace lending,
robo- advice, insurance technology (insurtech) and crypto-assets have emerged
around the world.2 In the past decade, fintech has already driven greater access to
and convenience of financial services for retail users. Meanwhile, artificial
intelligence (AI), cloud services, and distributed ledger technology (DLT) are
transforming wholesale markets in areas as diverse as financial market trading and
regulatory and supervisory technology (regtech and suptech). A host of new firms
have sprung up to apply new technologies to meet customer demand and most
incumbents indicate that digital transformation is a strategic priority (Feyen et al
2021). Indeed, leading banks are rapidly closing gaps in digitization of internal
processes and customer offerings, to compete with fintechs and the large
technology (big tech) firms that have also entered the fray (BIS 2019; Frost et al
2019).
These developments have the potential to make markets more diverse,
competitive, efficient, and inclusive, but could also increase concentration.
Innovation has introduced competition and increased inclusion, particularly in
emerging markets and developing economies (Pazarbasioglu et al 2020; Frost et al
2021). Fintech seems to have thrived particularly in markets where the financial
system had been less developed (FSB 2020; Didier et al 2021). However, the
underlying economics of intermediation combined with new technology may lead
to concentration among both traditional and new financial services providers.
Monopolistic or anti- competitive behaviours by big technology platforms are
already being scrutinized. As financial services move towards similar technology-
driven configurations, regulators are grappling with questions of how best to
regulate and supervise a landscape that is increasingly characterized by new players
and business models; and to address potential challenges to financial stability,
financial integrity, fair competition, and consumer protection (including data
privacy).
The COVID-19 pandemic has accelerated the digital transformation. In
particular, the need for digital connectivity to replace physical interactions between
consumers and providers, and in the processes that produce financial services, will
be even more important as economies, financial services providers, businesses and
individuals navigate the pandemic and the eventual post-COVID-19 world. For
instance, the pandemic has already accelerated the shift to digital payments (Auer et
al 2020a). It has also intensified e-commerce (BIS 2020; Alfonso et al 2021), which
may benefit big tech firms and their activities in finance. Countries with more
stringent COVID-19 policies and lower community mobility experienced a larger
increase in financial app downloads in the wake of the outbreak (Didier et al 2021).
Finally, it may be speeding up work on central bank digital currencies (CBDCs) (Auer
et al 2020b).
This paper examines the implications of digital innovation for market structure
and attendant policies, including financial and competition regulation. There have
been a number of surveys of regulatory responses. This paper takes a step back, to
look at what the economic theory of banking and financial intermediation can tell us
about how technology may drive industrial organization in the sector, and how that
might inform further policy responses. The paper roots the impact of the digital

1
2
transformation of finance in how innovation has enabled providers to address long-
standing challenges of financial intermediation – including asymmetric information,
uncertainty, incomplete markets, and fixed and variable costs of production. The
paper describes how digital innovation affects these key economic frictions in
finance, and alters the financial services value chain and industrial organization. The
forces driving these changes, and potential outcomes in terms of industry structure,
lead to insights for policy makers on how to harness the benefits of fintech, while
mitigating some of the risks, particularly around competition and market structure.
The focus is on economic and technological forces that apply broadly across
financial services. It recognizes that the sector encompasses a wide range of
different products and services, and is composed of numerous sub-markets that
might use different technologies or have different economic structures. These may
thus diverge in market structure and competition outcomes.

2. Economic frictions and forces in financial services

Fundamentally, financial firms, like other firms, owe their existence to transaction
costs (Baltensperger 1980). In the absence of complete trust between parties,
market interactions on the production side and the customer side are characterized
by risks, due eg to principal-agent challenges and incomplete or asymmetric
information. Solving these to reduce risk and create trust imposes costs on both
institutions and consumers, including contracting, search, and verification costs. For
example, lending is characterized by information asymmetries ex ante, as lenders
need to determine the risk profile of potential borrowers, and ex post, as they need
to monitor the repayment capacity of borrowers (Dewatripoint and Tirole 1994). A
fundamental feature of payment markets is the need to keep track of payment
obligations, and to verify the identity of account holders or the veracity of payment
tokens (Kahn and Roberds 2009). Different actors in the payment processing chain
must trust that the other links will not expose them to fraud or liability, and
customers require trustworthy counterparties with which to lodge funds and reliable
processes for their delivery. Financial market investment and insurance are subject
to uncertainty around future outcomes, adverse selection, and moral hazard. Those
creating investment products rely on sound underwriting and execution services to
be able to offer a quality product to their customers. Customers in turn must be
able to trust the soundness of the investments and of the operations that underlie
their ability to buy and sell.
As in other industries, internalization of activities within a single financial
services firm overcomes principal-agent and asymmetric information challenges, to
align interests and monitor actions. This ensures trusted interactions across teams.
Linking deposit taking to lending enables close coordination of asset and liability
management. Combining payments execution with account management allows the
provider to confirm availability of funds before transfer instructions are carried out.
Linking underwriting, trading, and sales allows firms to design new investment
products in line with market conditions and investor preferences.
The presence of uncertainty about future outcomes, for example whether a
borrower will go bankrupt, adds further frictions. Because it is difficult to define a
contract for all potential future states of the world – and the resulting solvency
status of the borrower – markets are not complete in the Arrow-Debreu (1954)
sense. As a

3
 result, resource allocations may be inefficient: some borrowers will have to pay
higher interest rates to compensate the lender for expected (but not necessarily
realized) losses, costly insurance may be required, or some loans will not be made at
all. More generally, the difficulty of tailoring products to the precise circumstances
of different customers, due either to asymmetric information or uncertainty about
outcomes, means that pricing, maturity, or other terms will inevitably be less than
perfectly suited to the circumstances of some clients. They may decline the offer, or
the intermediary may deem certain segments unviable commercially. Either way,
some customers will not be served.
Financial services companies are structured to address particular information
gaps and frictions related to intermediation. Banks developed to address the
maturity transformation challenge resulting from incomplete information about
future liquidity needs of depositors. They also address the transactional costs and
risk management needs of intermediating investments across individuals who lack
direct knowledge of other counterparties or information about economic activities
taking place elsewhere. Banks diversify across large numbers of borrowers to better
manage uncertainty of outcomes. Exchanges and brokers address the search and
transaction costs of individual issuers and investors, reducing information
asymmetries through listing requirements and publishing prices, and providing
infrastructure and services to match and enable transactions between buyers and
sellers who don’t know each other. Because so much of the intermediation process
is not readily visible to customers, and risks may only be realized after an extended
period of time, the consumer need for trusted providers manifests more than in
other industries. A provider that gains customer trust with one product can leverage
that trust to offer other services.
Providing these services requires not only information and financial resources,
but also real resources. These include the labour, equipment, and premises to
produce financial contracts, manage accounts, and process customer transactions.
The mix of labour, physical capital, financial capital, and trust capital will vary by
business type, and has been rapidly changing. Nevertheless, financial services
production is subject, in varying degrees, to the same frictions that affect
production in other industries, including real resource indivisibility and fixed costs.
Thus, while financial intermediation is subject to unique information and transaction
frictions, as an industry it is also subject to familiar economic forces observed across
sectors.
Basic economic forces are at play across both the financial and real resources
deployed by financial firms. These include:
 Economies of scale. On the supply side, traditional financial firms have had large
fixed cost investment needs for the creation and maintenance of back-office
systems and physical distribution networks to connect to the consumer. There
may also be fixed costs of minimum capital requirements and regulatory
compliance operations. As in any industry with fixed costs, economies of scale
emerge when a larger producer can amortize those costs over a larger
customer base. Scale also permits the development of a diversified balance
sheet to better manage liquidity and credit risk. Scale can reduce the marginal
cost of risk-taking (Mester 2010) and allow better pricing and/or ability to serve
a wider range of customers.
 Economies of scope. Also on the supply side, financial intermediaries enjoy
economies of scope by bundling interdependent financial services that can be
delivered through the same physical customer interfaces and leverage the same

4
 balance sheet. Cross-selling loan products and insurance, and offering both
asset and liability products, can create synergies and reduce costs. Economies
of scope are further enforced by the demand side as many customers prefer a
conveniently offered suite of products. This also reinforces the institution’s role
of gatekeeper to the customer.
 Network effects. On the demand side, network effects (or “externalities”) are
significant in financial services such as payments, where the value of the
network to all users (both payers and payees) increases when the number of
connected users increases. A bank serving a business, suppliers and as well as
its customers, could more efficiently connect counterparties to quickly transfer
payments and provide working capital.
Given the available technology and prevailing regulation, these forces have
historically conferred advantages to “first movers” and large, vertically and
horizontally integrated players. This has been particularly true for capital-intensive
products such as lending, for institutions that put consumers’ savings at risk such as
deposit taking, and for capital and network-intensive areas such as payments.

2.1 The Impact of digital innovation on key economic frictions

The adoption of technology is not new in the financial sector, but a number of
constraints had defined the operating environment until recently. In the late 20th
century, the industry was already characterized by a relatively high degree of
computerization since most financial services were dematerialized. 3 Only payments
frequently required physical cash or a check, and onboarding for new products and
services often required in-person or paper-based processes. Still, reaching and
connecting to customers routinely required physical infrastructure such as branches
and automated teller machines (ATMs). Customers wishing to transact with
counterparties using other banks had to use expensive and sometimes slow or risky
processes such as wire transfers. Even after the advent of digital payment systems
and the dematerialization of securities, connectivity remained a barrier to entry – an
institution typically had to be licensed and part of the consortium of banks or
brokerage houses to participate in a transactional network. Furthermore, data
processing and storage were expensive, requiring the operation of bespoke
mainframes and data centres. This limited the volume of information that could be
gathered, stored, analysed, and exchanged to improve efficiency, better price risk,
and tailor products to customer needs.

Technology advances in connectivity, data processing, and storage

Significant technology advances have taken place in two key areas that have
contributed to the current wave of technology-based finance:
 Increased connectivity. Internet and mobile technology have rapidly increased
the ability to transfer information and interact remotely, both between
businesses and directly to the consumer. Through mobile and smartphones,
which are near-ubiquitous, technology has increased access to, and the
efficiency of, direct delivery channels and promises lower-cost, tailored
financial services.

3
This included data centres, front and back-office connectivity to core banking systems, branch
automation, and interoperable payments networks connecting financial firms, including wire
services, automated clearing houses (ACH), and ATM networks.

5
 As of late 2019, GSMA estimated that there were more than 5 billion mobile
subscriptions worldwide. Building on this user base, there were nearly one
billion registered mobile money accounts (Graph 1). In principle, most financial
services can be now delivered directly and digitally, vastly increasing access to
finance. An emerging class of services and assets could in principle even be
delivered without the need for an intermediary. At the same time, the rapid
increase in connectivity has enabled large network effects and strengthened the
position of established intermediaries offering mobile networks and
subscriptions, such as telecom companies, particularly in some emerging
market and developing economies (EMDEs). Furthermore, the development of
widely used applications and services like social media, search and social
communication have enabled more peer-to-peer casual interactions. These are
increasingly being tapped for economic interactions, as well. This has
strengthened the position of the companies providing these services.

Mobile and mobile wallet growth worldwide

In billions Graph 1

0
2010 2011 2012 2013 2014 2015 2016 2017 2018 2019
Number of registered mobile money accounts Unique mobile subscribers1
1
Defined by GSMA Intelligence as individual person that can account for multiple ‘mobile connections’ (ie SIM cards).

Source: GSMA Intelligence.

 Low-cost computing and data storage. Computing efficiency has risen

exponentially, as the cost of data storage has rapidly declined. The Cray-1 was
the world’s fastest “supercomputer” from 1976-1982, operating at 80 MHz,
weighing 5.5 tons, and costing $10 million. 4 The A11 chip powering an iPhone X
runs at a speed of 2.39 GHz – 30 times faster – and at a small fraction of the
size and cost.5 Data that required a room full of tape or hard drives in 1980 can
fit on a micro secure digital (SD) card today. As the cost of data storage has
fallen, from USD 0.11 per gigabyte in 2009 to 0.02 in 2020, the volume of data
being generated globally has increased exponentially – to an estimated 48
zettabytes (48 trillion gigabytes) in 2020 (Graph 2). The ability to process such
data rapidly has also risen with advances in artificial intelligence (AI) and
machine learning (FSB 2017). Digital technologies like smartphones, and
activities like social media and e-commerce generate a wealth of new data.
Much of the new data relates to individuals (“personal data”) or firms, and can
be processed automatically to

4
See https://www.computerhistory.org/revolution/supercomputers/10/7 .
5
See https://igotoffer.com/blog/iphone-x-sum-technologies .

6
 identify the characteristics of existing or potential clients and offer the financial
services that best fit them.

Costs of storage have declined as global data volumes have surged Graph 2
US dollars Zettabytes

0.09 45

0.06 30

0.03 15

0.00
0
2010 2012 2014 2016 2018 2020
Lhs: Average cost per GB for hard drives1 Rhs: Annual size of global datasphere
From Sep 2017, data extrapolated using the growth rate in price per MB from http://www.jcmit.net/diskprice.htm. The increase in 2012 is
explained by flooding in Thailand, where one-third of hard drives were produced globally. One zettabyte is one trillion gigabytes.

Sources: BACKBLAZE; jcmit.net/diskprice; SEAGATE (2018), The digitization of the world from edge to core, November.

Data, networks, and new business models

These advances have enabled the creation of huge amounts of capturable data, new
tools to analyse those data, and new business models leveraging insights from the
analysis. Connectivity generates and captures a variety of data, and digitalization of
more activities generates higher volumes of data at an even greater velocity.
Storage and processing capacity allows data to be organized, validated, and
analysed, including through computationally intensive techniques such as AI,
including machine learning applications on vast troves of data. Digitally “native”
data with high volume, variety, velocity and veracity are referred to as “Big Data.”6
These characteristics mean that not only did big data emerge due to increased
digitization of activities (including those facilitated by connectivity), but big data
also requires low cost storage and high capacity computing power to be useful. Big
data emerges from a variety of sources, including the location and usage data from
mobile phones, the contact information from social networks, the delivery
information from logistics companies, and the sales data from retail outlets and
payments networks. Big data is being used in a wide range of traditional financial
services and new types of businesses to improve credit analysis, process efficiency,
risk management, product design, customer service and other areas. Examples
include Trusting Social, which uses call records to develop credit scores, and Tenda
Pago, which uses retailers’ consumer goods order information as a basis for working
capital loans. Data can be an important resource for, and driver of, economic
development (World Bank 2021).

6
Earlier versions used the first three V’s, with veracity added as a characteristic necessary to make the
data truly useful (Lukoianova and Rubin, 2014). Some add a fifth V, value (BBVA 2020).

7
Enterprises using cloud storage in Europe rose sharply over 2014–18
In per cent of enterprises Graph 3

60

40

20

0
FI DK NL GB MT HR CZ EA PT LT DE SK HU LV GR RO BG
SE NO IE BE EE CY SI LU IT AT ES FR ME RS PL TR BA
2014 2018
Source: Eurostat.

These advances also enable a number of new business models for the provision
of technology, one of which is cloud-based computing. This combines the ability of
an enterprise or individual to connect to externally operated and managed data
centres with low-cost computing power and storage. The result is an ability to
obtain infrastructure on demand and reduce the fixed cost barrier to entry to offer
financial services. Surveys show that cloud adoption has proceeded rapidly. For
instance, in Europe, Eurostat estimates that over half of enterprises in the Nordic
countries used cloud storage in 2018, while firms in the Netherlands, Ireland, UK
and Malta have also seen rapid adoption (Graph 3). Recent evidence suggests that
cloud adoption has even accelerated during the COVID-19 pandemic (McKinsey
2020). For instance, a recent survey of IT leaders from 250 mid-sized companies
around the world found that 82% of respondents increased cloud usage as a result
of the COVID-19 pandemic and 91% are planning a more strategic use of cloud in
the near future (Snow 2020). Innovation in cloud services is proceeding apace, as
growing volumes of venture capital and private equity funding have flown into new
applications of cloud technology and development operations (DevOps) over the
past three years (Graph 4).
Another result of the technology advances described above has been the wide
emergence of platform-based business models. Fintechs, big techs and even some
incumbents have moved increasingly to a role as “matchmakers” between different
users and providers on their platforms (Croxson et al 2021). These businesses
leverage the connectivity of individuals and businesses, and the ability to quickly
and easily collaborate, discover counterparties, and package and deliver a range of
digital and physical goods and services. Platforms are two-sided or multi-sided
markets, and benefit from network effects that create more value for each
participant with increased numbers of other participants.

8
Investment in CloudTech and DevOps companies1 Graph 4
USD bn Number
45.6 38.9

//
USD USD bn

//
bn 300
15

200
10

100
5

0
0
200 2020
200 200 200 200 201 201 201 201 201
0 YTD
2 4 6 8 0 2 4 6 8

Lhs:
Total capital Rhs: Number of deals
invested
1
As classified by PitchBook. CloudTech and development operations (DevOps) include companies that help organizations develop,
operationalize and monitor software applications at scale. Data accessed on 18 November 2020.
Source: PitchBook Data Inc.

The platform provider also benefits from the network effects that attract more
users, transaction volumes from which to earn revenues, and data that in turn
enables the provider to target users with more goods or services. The ability to
connect large numbers of users to a wide array of services has led to the rise of so-
called ‘super apps’ such as WeChat, AliPay, and (perhaps) Facebook. These and
other technology companies operating platforms with large customer bases are
often referred to as big techs (Frost et al 2019). For this analysis, almost any large
company that has or can leverage big data against a large customer base could
become a big tech. Examples include telecom providers like Vodafone M-Pesa, and
retailers such as Walmart, which is already a substantial e-commerce player and in
early 2021 announced intentions to create a fintech operation. It could also include
business-to-business (B2B) players such as consumer goods manufacturers and
distributors. Among other features, platform operators can leverage digital delivery
channels and application programming interfaces (APIs) to seamlessly “embed”
financial services such as payments or loans into non-financial services such as e-
commerce or social media.
Network effects in these business models can lead to market concentration and
competition concerns. The effect of adding users on one side of a platform market
(ride-hailing drivers, for example), creating more value to users on the other side
(riders), and therefore becoming more attractive to users on the first side (drivers),
can result in a positive growth spiral and a “winner-takes-all” or “winner-takes-most”
outcome, where all market participants want to be on the same platform. There is
the risk that a platform that builds a dominant market position would extract rents
from its massive data and networks (Croxson et al 2021). The multi-sided nature of
the market, however, makes it difficult to define the relevant market interactions,
such as product tying, for competition analysis. Furthermore, multi-sided markets
are characterized by complex pricing structures, including asymmetric pricing that
more finely divides the consumer surplus amongst participants. Pricing practices
that could be considered anti-competitive in one-sided markets might be pro-
competitive in a multi-sided market by attracting more participants and increasing

9
welfare-enhancing network effects (White et al 2018). The implications of these
aspects for financial

1
services delivered through fintech or big tech platforms will be explored further
below.

Impact on transaction costs, information asymmetries, and market gaps

Through the advances in connectivity and computing, digital innovation helps to
alleviate transaction costs and enable a wide range of new financial services
business models. Adoption cuts across traditional and new providers. Digital
technologies can help reduce the costs of collecting, storing, processing, and
exchanging information
– including search costs, replication costs, tracking costs, and verification costs
(Goldfarb and Tucker 2019). In finance, this can help borrowers to more quickly find
a suitable offer on a loan, or help savers to find an investment product that suits
their specific needs. Technologies can also help parties to automate verification of
events, enabling the creation of a wider range of state-contingent products. For
instance, an insurance client may use a phone camera or other remote device to
document damage that results in an insurance claim, or a drone may verify that a
field has been planted to enable disbursement of an agricultural loan. Blockchain
technologies take these a level further by providing a set of underlying capabilities
making it easier for individuals and businesses to interact on a peer-to-peer basis
even when they do not know each other beforehand (trustless context). At the limit,
such advances may obviate the need for intermediation, by enabling individuals or
firms – such as the lenders and borrowers in P2P and private credit markets – to
interact directly without the need for a bank to aggregate funds, assess credit, and
provide ongoing monitoring and servicing functions.
Incumbents, fintech entrants, and big techs are leveraging technology to
address economic frictions. They can close information gaps (asymmetric
information, uncertainty), and cut costs by:
 Reducing information asymmetries. Classic theories of intermediation describe
how banks address information asymmetries between borrowers and lenders
(Aghion and Bolton 1997; Holmström and Tirole 1997). They also discuss how
residual gaps increase credit costs or restrict the supply of credit, both in terms
of interest rates and the costs of ancillary activities such as attaching collateral.
Better information, for example leveraging consumer data and analytics, could
improve risk assessment and reduce the need for collateral as an indicator of
creditworthiness in lending (Gambacorta et al 2020; see Box A). The use of vast
troves of data enables the provision of financial services to individuals and
small businesses with no or limited traditional credit history, promoting
financial inclusion. More information and greater transparency tend to increase
trust, enabling providers and users of funds to work with a wider range of
counterparties. With enough transparency between users of finance and
providers of funds, intermediation itself may not be necessary. For example,
fintech credit models such as P2P or marketplace lending seek to directly link
investors and borrowers (Claessens et al 2018).

1
 Enabling customization of financial services to construct more complete
markets. Data and automation facilitate the execution and monitoring of
complex contracts and creation of more complete markets. This could bring
financial services closer to the stylized general equilibrium of the classic Arrow-
Debreu (1954) model. Traditional core banking systems and marketing channels
were built around standardized products, and did not facilitate a fully
consumer- centric approach to product development and marketing. More
tailored services, such as loans, investment advice, or retirement planning
structured to take account of the individual circumstances of the borrower
across different outcomes and states of the world, required highly trained and
expensive experts. In contrast, fintech-enabled automation of processes
reduces the set-up costs for customized products and can track different
contingencies across a wide range of outcomes. Digital marketing systems
enable specialized products to find a sufficient customer base. The increased
availability of data and computing power makes it possible to better price risk,
tailor a product or service to the needs of the consumer, and potentially
construct a range of state-contingent products, executed and monitored
through smart contracts or other emerging technologies.
 Reducing fixed and marginal costs of producing financial services. Technology
has reduced the costs of, and need for, much of the traditional physical
infrastructure that drove fixed costs for the direct financial services provider. An
early example that has been particularly important in EMDEs is mobile money.
Mobile money has reduced the need for traditional bank branches and
payments acceptance infrastructure, eg point of sale (POS) terminals, and has
become a widely used alternative to bank accounts in some countries. Cloud-
based infrastructure, including Banking-as-a-Service (BaaS), provides
computing power, data storage, and even compliance services. A niche financial
services provider can connect to these services and purchase as much capacity
as it requires on demand without the need to build data centres and other
back-office infrastructure (or, in the case of BaaS, a license). 7 Financial
intermediaries can reduce marginal costs through technology-enabled
automation and “straight- through” processing, which are accelerating with the
expanded use of data and AI-based processes. Digital innovation can also help
to overcome spatial (geographical) barriers, and even to bridge differences
across legal jurisdictions (akin to processes discussed by Bodie and Merton
2005). The increased use of digital tools and platforms for a range of economic
activities makes it possible to embed tailored financial products into non-
financial activities, drastically reducing customer acquisition costs and risks.
Contextualized finance will be discussed further below.

7
While branch fixed costs may be eliminated by direct connectivity to customers, technology fixed
cost has been variabilized at the level of the financial services provider but remains a high fixed cost
offering from the technology provider(s). The relevant market for tech infrastructure shifts from
financial services to IT/cloud services. The resulting market concentration issues are addressed
below.

1
 Box A

Data and collateral

Banks usually require borrowers to pledge tangible assets, such as real estate, to help offset asymmetric information
problems. Collateral performs several functions: it can indicate creditworthiness, to the extent that having
accumulated the means to possess a tangible asset signals something about ability to repay a loan; it provides an
alternative means to recover the loan in case the borrower becomes unable to repay; and it aligns interests because
the potential loss of the collateral would be costly for a borrower who becomes unwilling to repay. By contrast, big
techs can use big data to better assess firms’ creditworthiness. They can link repayment performance to other
sources of value, such as access to new inventory or to an e-commerce platform to sell goods. These capabilities
could help to reduce the importance of collateral in solving asymmetric information problems and maintaining
alignment of interests.

Gambacorta et al (2020) compare how credit from a big tech firm (Ant Group) and traditional bank lending
correlate with local economic activity, house prices, and firm-specific characteristics. Using a unique random sample
of more than 2 million Chinese firms, the paper finds that big tech credit does not correlate with local business
conditions and house prices when controlling for demand factors.

Graph A1 indicates the unconditional elasticity between the different credit forms and house price. The dots in
the figures indicate the average logarithm credit use (y-axis) and the average logarithm of housing price (x-axis) at
the city-year level. The left-hand panel plots big tech credit, the middle panel plots bank secured credit (mortgages)
and the right-hand panel plots bank unsecured credit. Linear trend lines are reported in each graph, together with
95% confidence bands. The (unconditional) elasticity of big tech credit with respect to house prices is 0.09, while that
of unsecured bank credit is twice as high (0.184). That of secured bank credit is five times higher (0.488). A greater
use of big tech credit could therefore have macroeconomic implications, weakening the relevance of the collateral
channel.

Elasticity of big tech credit to asset prices Graph A1

Based on a 100,000 random sample of firms served by both MYbank (Ant Group) and traditional Chinese banks. The dots in the figures
indicate the average logarithm credit use (y-axis) and the average logarithm of housing price (x-axis) at the city-year level. Growth rates are
approximated using first differences of log values. The left-hand panel plots big tech credit, the middle panel plots bank secured credit and
the right hand panel plots bank unsecured credit. Linear trend lines are reported in each graph, together with 95% degree confidence
bands. Standard errors in brackets.
Source: Gambacorta et al (2020).

 Reducing search and switching costs for consumers. Advances in connectivity

and online search mean finding and using a financial services provider in the
next state or province – or across the globe – can be as easy as finding one

1
 in the

1
 same city. On the consumer side, the marginal cost of using multiple financial
service providers is also reduced. Technologies have become increasingly plug-
and-play, from the advent of common internet communications protocols and
mark-up languages, to “open banking” and standardized APIs that facilitate
data access and exchange. Increased access to information about remote
providers and their products and prices reduces search costs for users. The
reduction in search and switching costs thus applies both to providers and
individual customers. For example, a customer can now choose a start-up
remittance service provider offering lower fees and ready connectivity to the
customer’s bank account, or an established social media company that already
has the customer’s trust and connectivity to friends and family to whom
remittances are sent. Reduced search costs could eventually enable those who
need finance to connect directly to those with excess funds to invest, with smart
loan contracts replacing the monitoring provided by intermediaries. Lower
search, on-boarding, and switching costs could make markets more
competitive.
Despite reduced costs, increased efficiency, and all the new data and
computing power available, fintech remains subject to the same risks traditionally
present in finance. Credit, liquidity, market, and operational risks can be reduced or
transferred, but not eliminated completely. Market failures can result in financial
instability, speculative bubbles, domino effects and, potentially, systemic risks (Frost
2020). In most markets digital finance has not reached levels at which it presents
systemic risk, but as fintech and big tech finance grow, and traditional banks use
similar technologies to a greater degree, that will shift. P2P lending in China reached
significant volumes before the regulatory crackdown began in 2017. This culminated
in the November 2019 requirement that all platforms close or convert to regulated
small loan providers within two years. P2P lending in China and the more recent
Wirecard failure in Germany offer lessons on the need for appropriate regulation
and supervision of fintech. Big tech firms may already have reached the level of
systemic importance (too-big-to-fail) in some economies, notably in China.
Authorities are actively formulating appropriate regulatory responses to these risks
(see Section 4).
The digital transformation of finance also introduces – or heightens – cyber
risks. Cybersecurity is an issue for all sectors and for traditional financial services
providers as well as fintechs and big techs. The attack surface is broadening,
however, as interconnectivity increases and the disaggregation of services
introduces more links to each product chain and user interface. These factors can
also increase operational risks as a greater number of distinct entities may be
involved in the provision of a single product or service, creating complex webs of
operational dependency. These risks may have increased further during the COVID-
19 pandemic (Aldasoro et al 2021). At the same time, certain core services have
become more concentrated (eg provision of cloud services), creating the potential
that single points of failure could result in systemic disruption.
Finally, there are new risks to privacy and consumer protection. The
proliferation of providers and tailored services, reaching new customers and
embedding financial products into other activities, increase the risks of
compromising privacy, abusive practices or errors by unaware consumers.
Consumer protection and regulation of risks arising from fintech and digital
transformation are addressed in other papers in the forthcoming Future of Finance
series.

1
 2.2 Consequences for financial services firms

The financial services value chain consists of four broad components: customer
interface, back-office functions, infrastructure, and balance sheet. The operations
that ‘manufacture’ financial services involve the first three of these. The theory of
the firm in economics highlights transactions costs as a key determinant of whether
a given process or interaction will be mediated through the market or within a firm
(Coase 1937; Williamson 1971). Prior to the advent of fintech, the combination of
transactions costs and economies of scale and scope resulted in large financial
intermediaries that tended to be vertically and horizontally integrated, providing all
four components (and multitudes of subcomponents) internally.
i. Customer interface. This consists of physical networks such as branches, ATMs,
and points-of-sale to distribute core financial products and services. In the past,
customers needed to be physically present to conduct activities such as
opening an account, buying a product, making a transaction, and getting
advice. In many cases, these activities were paper-based and cumbersome.
Although the advent of the internet gave rise to online and mobile banking,
these were again owned and operated by a particular financial institution. As a
result, financial institutions largely remained as gatekeepers to the consumer
with a basic bank account being the pre-condition to participate in the
economy and acquire additional financial services. As a result, financial
institutions were also able to tap into a highly regulated, but sticky and cheap
source of funding: retail deposits.
ii. Back and middle office. This consists of all the departments, processes, and
internal systems to operate the bank’s products and balance sheets. These
functions include risk management, regulatory compliance, credit decisions,
fraud detection, call centres, trade reconciliation, technology operations, and
record keeping. Prior to the advent of fintech, these processes involved manual
and paper-based intervention and required cross-departmental coordination.
Organizational complexity and siloed legacy IT systems further increased
transaction costs, left customer data unexploited, impeded the ability to
innovate, and degraded customer experiences.
iii. Infrastructure. This connects individual financial sector participants (including
the central bank) to collectively comprise the financial system. This includes
payment and settlement systems (retail and wholesale) as well as other
financial infrastructures such as credit information bureaus and arrangements
that facilitate transactions across cross-borders. Many of these infrastructures
still face inefficiencies due to, for example, limited operating hours and
interoperability constraints. Further, until recently, non-banks typically could
not access these infrastructures directly and were therefore reliant on banks.
iv. Balance sheet. Deploying customer deposits (or other sources of funds) into
productive loans (and other uses of funds) through balance sheet
intermediation is one of the oldest functions in finance. A modern balance
sheet serves more complex functions and is optimized to attract a mix of retail
and wholesale funds, to match assets and liabilities across multiple dimensions
(tenor, currency, etc.) and to maintain adequate capital and liquidity levels.
Balance sheets of regulated financial institutions are subject to a variety of
requirements and constraints. Certain types of financial institutions – eg
payment service providers, mutual fund managers, and more recently P2P or
marketplace lenders – do not create customer liabilities and assets on their own
balance sheets. Nevertheless, balance

1
 sheet finance remains dominant in most EMDEs. Even in advanced economies,
assets that are ultimately funded through capital markets are often originated
on-balance sheet before being securitized or sold to a ‘shadow bank.’
As technology has reduced transaction costs, the boundaries of the firm have
changed and the production chain for financial services could be more
disaggregated (vertical disintegration). This has been taking place for decades; an
example is the use of third-party ATM and card networks to process banking
withdrawals or payments transactions. First, the ATM moved the transaction outside
the bank branch. Then, connectivity and information transfer technologies enabled
different companies to provide elements of the transaction value chain – an issuing
bank, a merchant processor operating a POS terminal, and a network operator. As a
result, the customer-facing bank could focus on its core competencies and
customers. With advances in technology, this process has accelerated. There is an
increasing array of products and services that can be sourced from multiple
different providers, combined, and offered to the customer as an integrated whole.
The customer-facing provider might incorporate features and functions from
external vendors, outsourcing providers, or partners. Examples include using third-
party sales agents or internet marketplace originators, external credit scoring
services, outsourced card issuer processing, and white-label integration of fintech
services.
Advances in information and connectivity have also led to an unbundling of
financial services consumption (horizontal disintegration). Customers now have
more information about different providers and an increased ability to interact
across providers, so they can choose different providers for the different financial
products they use. Similarly, providers have more access to information about new
or potential customers, can market to them better, and can offer services directly.
The availability of data also makes it easier for providers to underwrite a customer
without having had a prior relationship, so the inherent competitive advantage of a
single provider in offering additional products is diminished. Services that were
once consumed from one financial institution can be offered by a wide range of
specialized firms. Examples in both personal and business banking underscore this
unbundling (Graph 5). However, even as the consumer’s need for a horizontally
integrated provider has been reduced, providers can now integrate more products
from different providers. The outcome in terms of horizontal disintegration depends
on individual preferences, views on convenience, and costs to search and assemble
product sets. All of these are affected by changing access modes, eg from branch to
internet, agent, or mobile app.

Specialist providers challenging the services of large banks Graph 5

Unbundling of personal banking Unbundling of business banking

Source: World Bank; authors’ elaboration after Barrett (2018).

1
 Consumer interfaces are changing from physical branches to convenient digital
access from anywhere. Digital consumer interfaces (eg web, mobile) enable new
players to directly reach consumers both near and far. By using apps from different
competitors, consumers can build their own complete bank and gain more
personalized services. This has likely reduced the value of legacy physical networks
and the incumbent’s grip on the customer, who can easily shift funds to different
providers to access preferred services (Shevlin 2019). For example, the innovation of
mobile money was a network-based account ledger provided by non-bank players
and mobile network operators, leveraging already-ubiquitous mobile network
connectivity infrastructure. This was combined with a new ecosystem of low-cost
agents for the customer interface, and a high degree of automation to eliminate
most back-office processes. Mobile money ecosystems deliver low value/high
volume transactions cost effectively, making this a viable offering for low income
inclusive finance segments. Similarly, fintech payment aggregators (eg Square and
Stripe) offer convenient payment services to merchants without the need for them
to set up a merchant account with a processing bank. The aggregators act as the
‘merchant of record’ taking responsibility for compliance vis-à-vis network and bank
requirements. A related development is that the customer interface could be with
institutions outside the jurisdictional boundaries of the customer. On the internet, it
does not matter to the consumer if the provider is in Kansas or Kathmandu
(although it may matter to the regulator!) Services have adapted accordingly.
Non-financial companies can also leverage customer connectivity and data to
offer financial services, altering the competitive landscape and blurring industry
boundaries. As mobile applications and the internet have become a dominant way
through which consumers live their lives and businesses operate, e-commerce,
telecom, online search, logistics/transportation, and social media platforms have
become the new gatekeepers to the customer. Digital platforms and big techs in
particular are well placed to directly tap into their vast customer networks and data
(BIS 2019). They are able to offer contextually relevant financial services to their user
bases to enhance the customer’s experience or increase sales of their core product.
For example, e-commerce, ride-hailing, and social media platforms have integrated
seamless payment services to strengthen the value proposition in their core
markets. Micro-insurance has been embedded into mobile money and the sale of
consumer products. Users do not desire most financial services in and of
themselves, but see them as a means to another activity (eg pay a driver to reach a
destination, take out a working capital loan to obtain inventory). As such,
embedding the financial service directly into that core activity can be attractive to
the customer. It can even make the customer more aware of services about which
they otherwise might not have known, or to which they might not have had access.
Banks have used customer relationships to cross-sell financial services since the
beginning of banking; now an increasing range of non-bank actors are doing the
same thing.
The back and middle office are being revamped – or eliminated entirely.
Technology has reduced the costs of, and need for, much of the traditional back-
office infrastructure, from paper processing to data centres. The back office is being
revamped to lower costs and improve communication. Process automation and
upgrades to software and IT systems are causing a restructuring of financial
institutions and a reduction of full-time employees. Middle-office functions such as
reconciliations are increasingly unnecessary. Entire processes, and many of the skills
that previously had to be hired, can be replaced with automation or expert systems.
This has also given room for third parties to step in, such as cloud service providers

1
and fintechs with specialized capabilities that sell their software-as-a-service (SaaS)
to banks (eg data processing, credit scoring, electronic know-your-customer (e-
KYC)). For example, a bank can work with a credit-scoring company that leverages
unique data and scoring expertise to provide a seamless customer experience.
Unencumbered by legacy systems, some new entrants have built new software
platforms to handle their core banking activities, while others rent. Cloud-based
infrastructure and BaaS providers allow smaller banks to outsource technology
operations and leverage the latest systems to compete alongside larger banks. They
enable small start-ups to purchase not only data processing and storage capacity,
but also entry points into regulated infrastructure and regulatory compliance. Thus,
smaller players can grow capacity in tandem with their customer base, without the
initial setup costs and step function cost curves previously required.
Technology has also accelerated the internationalization of back-office
functions. For several decades there has been a trend of offshoring back-office
functions. Many financial firms first shifted call centres and customer service to
lower-cost environments; they later shifted application processing and IT
operations. This trend is accelerated by the advent of cloud computing. Entire IT
systems, including core banking systems, can now be hosted anywhere in the world.
In regions with common regulatory frameworks, even the entire provision of some
products and services could be outside the boundaries of the customer’s country.
For example, a digital bank domiciled in Lithuania can bundle products and services
from elsewhere in the European Union (EU) and offer them to customers over the
entire EU. On the other hand, in jurisdictions where the use of cloud-based
infrastructure and/or overseas data processing and storage have been restricted
(due to data localization or other policies), efficiencies and entry may be limited.
This is particularly true if the market is too small to support its own localized
infrastructure. The shift of back-office functions
– and associated processes and customer data – to other jurisdictions can raise a
new set of supervisory challenges.
Supporting infrastructures are changing. Much of the traditional infrastructure,
which was both a cost and source of competitive advantage for incumbents, can be
reduced or replaced, or must be shared. Digital identity (ID) and know your
customer (KYC) registries are implemented in many countries and can help redefine
the onboarding and authentication process that has long required in-person
verification at a branch (D’Silva et al 2019). New infrastructures such as API
platforms for payments or lending, and distributed ledger systems that can support
crypto-assets or stablecoins, could produce large network effects and change the
role of, or even disintermediate, certain players (Arner et al 2020). Open banking
initiatives require banks to share customer data with fintechs. In some jurisdictions
banks must initiate transactions introduced by fintechs upon customer request,
further eroding the traditional incumbent-customer relationship. At the same time,
public payment system innovations have allowed access to non-banks. On the one
hand, open banking APIs enable non-banks to offer payment initiation services,
without them having to participate in any of the payment systems themselves. On
the other hand central banks are requiring transparent and risk-based access criteria
for payment and settlement systems, replacing earlier restrictions of access to only
banks.

1
 3. Implications for the industrial organization of the
financial sector

3.1 Implications for market entry

Digital innovation has reduced cost barriers, allowing new and smaller players to
enter. The elimination of many fixed costs and a reduction in variable and switching
costs makes it possible for low-cost providers to enter the market, subject to local
regulation. Although a trusted reputation must still be developed, small providers
are more likely than in the past to be economically viable. Such new providers may
strip away particular customer segments and revenue bases from traditional
providers, or broaden access to finance for previously underserved segments,
without needing to achieve large scale and scope. Apps and cloud-based
computing and software platforms have enabled entrepreneurs to quickly bootstrap
without the need to raise huge amounts of capital to finance massive upfront
investments. APIs and open banking initiatives have the potential to further
accelerate this trend, since a new service need not wrench the customer completely
away from the incumbent, and can therefore build trust by layering a service on top
of the safety net provided by legacy institutions.
Infrastructure connectivity and a reduced need for physical branches allow
established companies from other sectors to offer financial services as well. These
entrants can deploy automated processes in lieu of hiring a specialized workforce,
connect to the financial transactions infrastructure, and leverage cloud-based
infrastructures to reduce the cost of cross-market entry. For many of these new
providers, such as digital platform companies and some telecoms, consumer trust
and a customer base are already established in their core markets and potentially
transferable to the financial sector (Oliver Wyman 2019). Some are able to combine
financial services with other products or core capabilities as part of a platform
offering. This is particularly relevant in EMDEs where the financial system is less
developed and access to financial services is more limited. This affords fintechs and
big techs more room to expand their financial activities and compete with
incumbents (FSB 2020).
Entry into financial services is accelerated by the ability to turn regulatory
compliance into a technology integration process. Several banks have built BaaS
platforms to serve fintechs and even other banks. They can connect these clients to
the payment system or a bank balance sheet, thus reducing the need to deal with
the complexities of licensing, regulation, and developing their own core banking
systems. Leveraging BaaS, new entrants can tap into existing – and already
regulated – financial transactions and balance sheet providers. Where open banking
is in force, they can also leverage the data, and sometimes functionality, of existing
financial accounts. This allows a wider range of use cases into which finance can be
embedded; in particular, a range of commercial platforms or transactional situations
can layer on a financial service from a BaaS provider without much of the systems
and compliance overhead previously required. While increasing competition, this
multi-faceted and multi-tiered access to the regulated banking sector introduces
new challenges for regulators and supervisors, as discussed below.

2
3.2 Shifting economies of scale and scope

However, basic economic forces remain relevant. Economies of scale and scope
remain, even as the minimum efficient scale for service delivery is lower for the
individual user and for most financial services providers. That has been
accomplished in part by shifting the scale effects to the infrastructure providers;
scale remains highly relevant in areas of cloud computing and data processing and
software platforms. In fact, new forms of scale have emerged in connectivity and
computing, along with previously present economies of scale in capital, including
reputation, or ”trust capital.”
In the digital finance era, economic frictions have not disappeared, but take on
new forms. The costs of services have been reduced in many cases, but certainly not
eliminated. Financial services providers, particularly new non-bank providers, face
two important costs in the unbundled marketplace:
 Customer acquisition costs. Customer acquisition costs, comprising marketing,
onboarding, KYC, and initial credit assessment, remain significant relative to
revenues, especially for retail financial services. This is partly due to regulatory
requirements, but marketing and other onboarding costs remain pertinent even
when remote interactions are possible (eg leveraging e-KYC). While technology
has made it easy to directly reach users digitally, the cost to acquire new
customers remains high because of user inertia, which is particularly present in
retail and SME business lines. Customer acquisition costs are also significant for
wholesale customers, as systems integration and business process changes may
be cumbersome, and initial credit assessment is more complex. Amortizing
these and other fixed costs across more customers and products allows for
economies of scale and scope. This naturally puts larger incumbents and big
techs at an advantage.
 Funding costs. As noted above, some products have natural complementarities
that reduce the cost of providing one or the other product if offered in tandem.
Most clearly, offering loans is more economical if there is a cheap source of
funding in the form of deposits. There are other complementarities in
derivatives, stock lending, insurance, etc. Size, and the diversification it brings,
can provide a funding advantage due to lower risk, and greater liquidity of
traded funding instruments (bonds, shares) issued at scale. In general, big tech
firms have a higher overall cost of funding than global systemically important
financial institutions (G-SIFIs) because they use more equity and have no
deposits. However, when analysing similar funding components (ie bonds) big
techs tend to have a lower cost at issuance (Graph 6).

2
Average spread of active bonds to benchmark government bonds at issuance
In basis points Graph 6

300

200

100

Aegon

ING Groep

Charles Schwab
China Construction Bank

Deutsche Bank

Wells Fargo
Morgan Stanley
Prudential

Prudential Financial
Amazon
Vodafone Group
Tencent

Microsoft

Ind'l & Com'l Bank of

Standard Chartered

American Int'l Group

State Street
BNP Paribas

Goldman Sachs
Citigroup

Mizuho Financial Group

Fin'l Group Royal Bank of

Orange

Societe Generale

Credit Suisse

MetLife
Banco Santander

JPMorgan Chase

BlackRock
Aviva

China

Bank of China

Agricultural Bank of China

Bank of America

Canada
Group Mitsubishi UFJ
Sumitomo Mitsui Fin'l
Allianz
Alibaba Group

Barclays
HSBC Holdings
Credit Agricole

UBS
Baidu

UniCredit
Apple

AXA
Alphabet

Big tech
G-SIFIs
The dashed horizontal lines indicate the simple average.
Average spread of active bonds over the respective benchmark government bond at issuance as collected by the Bloomberg SRCH function.
Terminal accessed on 3 Sep 2020. Filters used: Corporates, Active Bonds, Issue Date >01/01/2013 and Issuer Name as listed in the graph.
Sources: Cornelli et al (2020b); Bloomberg.

Consumers and other users also experience frictions in the unbundled financial
services marketplace. As consumers balance choice and convenience, they may
favour firms offering broader scope:
 Assembly costs. Simplicity and convenience have significant value, often
resulting in preferences for a single provider that offers an integrated suite of
financial products and services, even if each individual product may be less
convenient or well-designed, or marginally more expensive, than those of
alternative niche providers. For many consumers, there is a cost to the time,
effort, and potential confusion of assembling fragmented services from
unbundled providers and moving funds between them. Thus, most savers still
use intermediaries or platforms to deploy their savings instead of searching out
individual borrowers. Similarly, “contextualized finance” embeds the financial
service into the actual economic activity the customer is seeking to perform (eg
pay for a trip on a hail riding platform, take out a loan on an e-commerce
platform). This can provide the service that is needed exactly when it is needed.
Where the customer is already using a platform provider or big tech,
onboarding costs are reduced, and convenience for the user creates economies
of scope for the provider. The rise of “super apps” such as WeChat, AliPay,
Yandex, and Grab, and contenders such as Rappi and Revolut which add
multiple services, is testament to the value consumers place on being able to
go to a single source for multiple services.
 Switching costs. For consumers, business users and financial services providers
considering a change of vendor, there are barriers in the form of effort, inertia,
inconvenience, and the difficulty of disentangling a service that has been linked

2
 to multiple activities and systems. Switching costs can also result from
intentional technical design. Consumers may flock to digital platforms with
large customer bases to benefit from network effects. Yet once they become
large, platforms may seek to increase switching costs. If a service has been
integrated into multiple activities or workflows, or has accumulated significant
user data, it may also able to provide better or more tailored services that a
new provider would not be able to match. Where proprietary protocols are in
place, or historical data are difficult to share, connectivity across different
providers might be hindered.

3.3 Impact on financial services providers

Reduced economic frictions, the ability to reconfigure the value chain, new
opportunities for entry, and shifting economies of scale and scope will affect
different financial services providers differently. The impact will depend on their
current market positions and ability to leverage technology. Incumbents start with
advantages in trust capital and regulatory position. Fintechs are proving nimble at
leveraging data, connectivity, and improved processing capacity, and at converting
regulatory barriers into solvable technology challenges. Big techs compound the
advantages of fintechs with large scale existing customer bases for non-financial
business lines, associated customer data, and high levels of brand trust (Stulz 2020;
OECD 2020).8 Individuals play a limited role as financial service providers today,
interacting largely via intermediaries. In a future of increased connectivity and
decentralized finance, individuals may be further empowered as direct providers of
financial services.

Incumbents
Although there is increasing competition, the continued relevance of basic
economic forces enables incumbent financial institutions with integrated product
offerings to maintain – at least for now – significant market share. Incumbents have
an advantage in having already overcome some of the information asymmetries
that form a barrier to trust: they are regulated and have been in the market and
demonstrated their reliability over time. Multi-product incumbents ameliorate
customer switching and assembly costs. Incumbents benefit from economies of
scale in financial capital, and the diversification benefits inherent in larger scale and
broader scope. Incumbents also have economies of scale in customer data, which
they are beginning to better leverage, and customers often have higher trust in
incumbents to safeguard their data (Armantier et al 2021).
Unbundling could reduce the cross-subsidies that are inherent in the integrated
banking model. Free use of a current account is often funded by the use of the
balances to lend to other customers, or by the value created in providing other
products to the account holder. While product tying is often considered an anti-
competitive practice, in most markets traditional banking products have explicit
exemption from anti-tying rules (see, eg OCC 1995). Disaggregation can disrupt this
business model, potentially stripping away the more profitable products and
services that have stand-alone profitability (eg remittances), and leaving traditional
providers

8
Surveys of the most trusted consumer brands globally and in the US continue to place big techs
well above financial institutions. See Morning Consult (2021). On the other hand, Armantier et al.
(2021) find that US consumers trust big techs the least to safeguard their personal data, and have

2
 more trust in banks, followed by fintechs and the government. The difference may be explained in
part by the difference in overall brand trust versus trust to safeguard personal data.

2
with an embedded cost base and products with low margins or that the market
expects to get for free. Open banking regulations that require banks to offer access
to their customers’ accounts could accelerate this process.
Incumbents are innovating and upgrading across each component of the value
chain. Indeed, a recent World Bank survey finds that digital transformation is a
strategic priority for the overwhelming majority of incumbents. They aim to shed
overhead costs and employees, improve products, and compete for the consumer
relationship (Feyen et al 2021). Many expect further growth in digital customer
interactions and transactions, which will intensify competition. The largest
institutions can leverage their scale, customer base, regulatory expertise, and
capacity to update their technology stacks and manage large balance sheets in
order to effectively compete with big tech entrants. Smaller institutions can take
advantage of on- demand infrastructure and external service providers to undertake
a rapid and less costly digital transformation. Incumbents that are not able to adapt
in time may perish or be acquired.
Some incumbents have built distribution platforms to enrich their own product
offering by allowing fintechs to offer their products directly to the incumbent’s
users. In essence, these incumbents have become marketplaces and leverage
network effects.9 Innovations in infrastructure such as APIs can facilitate these
partnerships, which benefit from both the fintechs’ need to achieve scale, and
banks’ continued centrality in the financial services relationship for many customers.
As long as customers remain tethered to bank accounts directly or through fintechs,
banks can continue to benefit from their own economies of scale in capital,
connectivity, and high-fidelity consumer financial data.
Banks continue to have deep expertise in regulation and compliance and have
already passed key regulatory hurdles ahead of new entrants. Many fintechs and big
techs offer their financial services in close direct or indirect collaboration with banks
to access the regulated financial system at the back office or infrastructure layers.
B2B fintechs and big techs offer their services to incumbents in the form of
partnerships or SaaS engagements. This is also the main level where banks engage
with big techs to procure state-of-the-art cloud computing and data processing
services which require deep expertise and economies of scale. Business-to-
consumer (B2C) fintechs and big techs that compete directly at the customer
interface level often still rely on banks themselves to store value (eg the so-called
float, or deposited funds, of mobile money operators). They also use banks to
access the payment system, or use other core banking system functions that are
difficult to build in-house (eg by using BaaS). Moreover, depending on the
regulatory environment and the business model choices fintechs make, their
customers often need to have an account with a financial institution to use the
fintechs’ services, driving continued demand for at least some incumbents’ services.
Incumbents also have a comparative advantage in managing complex, large
balance sheets and navigating compliance with evolving regulations. This is critical
since finance is one of the most highly regulated sectors. As such, many fintechs and
big techs may be reluctant to become full-fledged regulated financial institutions,
which can be a slow, costly, and uncertain process. However, regtech solutions could
erode part of this comparative advantage of incumbents (Auer 2019). As another

9
See EY (2020) and Croxson et al. (2021).

2
example, lending platforms also rely on incumbents and institutional investors to
purchase the loans originated on the platform.
The main challenge for incumbents is that they are encumbered by legacy
systems, networks, and cultures; not all will be able to adapt and survive. Legacy,
fixed-cost infrastructure cannot be easily scaled back at the same rate as it becomes
obsolete. Financial institutions (and regulators) are also mindful of the need to
maintain minimum services for non-digitized customer segments, especially in more
remote locations where alternatives are limited. Incumbents bear regulatory and
compliance requirements that may not apply to new competitors or that can be
evaded via regulatory arbitrage. Incumbents may also have service obligations not
imposed on fintech or big tech players. These regulatory and social obligations can
present competitive disadvantages if steps are not taken to level the playing field.

Fintechs
After driving the unbundling of financial services, many fintechs have begun to re-
bundle. Fintechs have been adding new products to their offerings, by themselves
or via partnerships and platform approaches. This allows them to cross-sell to
existing customers and make themselves more attractive to new ones. Some
examples are shown in Graph 7. For some, this has been the result of opportunities
to provide a better customer experience and increase revenue. For others, it has
been a matter of necessity. Since scale and scope economies still matter, fintechs
that focused on a limited range of products may not be able to achieve funding
synergies or amortize customer acquisition costs, regulatory compliance costs, and
investments in brand recognition across a sufficiently large revenue base to be able
to generate profits or fully compete on price with incumbents. It will be interesting
to observe how these newly re-bundled product sets evolve. Newer companies
building up a set of services for a particular customer base in today’s environment,
leveraging high quality-data on market needs and customer profitability, will
doubtless arrive at a different product set than was developed by universal banks
through agglomeration of products over a long period, from the 18th to the 20th
centuries.
New entrants initially stayed away from financial activities that carry large
licensing, capital and regulatory burdens. Most new players focused on activities
such as payments, cards and financial advice, and avoided highly regulated, capital
intensive activities such as on-balance sheet lending and securities underwriting.
Indeed, the rapid rise of fintechs was facilitated in many markets by regulatory gaps
that allowed new service providers to enter and operate with minimal regulatory
burden. As regulators caught up, fintech strategies shifted; in some markets they
were forced to seek licenses.
Some fintechs have chosen a regulated status as part of their evolving product
strategies. For example, alternative finance platforms initially focused on matching
supply and demand of capital, avoiding the regulatory burden of intermediation.
However, in the process of re-bundling, some fintechs have sought full banking
licenses. In particular, lending platforms have faced challenges to fund their
operations in wholesale markets or using a pure P2P funding model. As a result,
they have become more reliant on incumbents for funding, or have sought their
own banking licenses. For example, Grab and SoFi have sought banking licenses,
while GoJek and MoneyTap partner with banks (the latter also has a non-bank
financial company license). Other fintechs prefer to avoid a higher regulatory
burden by expanding across borders to reach scale with their existing product
offering. The

2
 COVID-19 pandemic and resulting economic dislocations have exacerbated the
challenges of relying on crowdfunding, capital markets, or bank partners for balance
sheet capacity. As a result, two prominent early fintech lenders, OnDeck and
Kabbage, were sold to other companies in July and August 2020.

Fintech product offerings have expanded over time Graph 7

Big techs
Basic economic forces such as scale and scope economies, along with network
effects, also play to the fundamental strengths of big tech companies. Big techs
already enjoy an active customer base, creating economies of scale and scope and
network effects in their core markets; they are increasingly able to leverage these to
exert market power in the financial sector (BIS 2019; Frost et al 2019). Agility in
systems development and creating easy-to-use customer interfaces are relative
strengths of tech companies compared to traditional financial institutions, and big
techs’ access to customer data and understanding of customer behaviours may
equal or exceed those of financial institutions. Big techs can bundle financial
services to reinforce the value proposition of the ecosystem in their core markets,
such as e-commerce and social media, or offer financial services as a new product
line cross-sold to the existing customer base. This phenomenon is manifest in
developed and emerging markets, often led by the latter. Alibaba was a leader in
layering payments onto e-commerce with AliPay. It then developed a broad range
of financial services. (By contrast, eBay’s integration of the original incarnation of
PayPal did not go as far). Marketplaces from Amazon to Gojek, Grab, Jumia, Lazada,
and Mercado Libre are integrating merchant loans into their offerings. Because
financial services are layered over other cash generating activities, these companies
can subsidize the financial services. For instance, WeChat Pay, WhatsApp Pay and
other providers allow users to send money to each other for free – though fees may
be higher for merchants and business users. In some cases, the value a big tech can
generate from customer data by driving traffic to its core business may be higher
than the financial service fee income.

2
 Big techs have already amassed vast troves of relevant alternative customer
data which traditional financial institutions lack. Combined with their ability to
process these data for insights, this has enabled big techs to offer tailored financial
services such as insurance or working capital loans to merchants on their platforms
who traditionally lack collateral or strong credit histories. This has uprooted the
notion of traditional “relationship banking” in which banks have superior proprietary
hard and soft information about their clients thus keeping them captive. Big techs
also often choose to partner with incumbent financial institutions to offer these
services so they can focus on their core markets. Regardless, bundling of services
sets in motion a feedback loop, ie the data-network-activities (DNA) loop (see Box
B) where big tech ecosystems become more valuable to users. This begets new
users, higher user engagement, and more user data. This in turn boosts network
effects and economies of scale and scope, which enable big techs to increase the
value of their ecosystems.
Big techs can leverage unique market power in providing contextual finance –
the bundling of financial services with core activities. In addition to the data-
network- activity feedback loop described above, “contextualized finance” may
result in improved operational efficiency and portfolio performance relative to
traditional financial institutions. A ride hailing service that is financing the driver’s
automobile can monitor usage, maintenance patterns, and cash flows. A distribution
platform offering working capital credit to retailers purchasing inventory could cut
off the supply of goods if the retailer is late in repaying; similarly, an e-commerce
platform providing logistics and financing to merchants could seize inventory. Given
network effects and high switching costs, big techs could enforce loan repayments
by the simple threat of a downgrade or exclusion from their ecosystem in case of
default. Their position as customer gatekeepers in their core businesses helps
establish roles as gatekeepers and potentially enforcers for financial services, as well.
Although research has shown that fintech and big tech credit tends to complement,
rather than replace, other forms of credit (see Box B), some big techs have explored
becoming a full-fledged deposit-taking financial institution. Big techs active in
EMDEs, with lower levels of financial development and inclusion and more
permissive regulatory environments, have had more scope to venture on this path.
Other jurisdictions impose limits to the span of activities of regulated financial
institutions, with the result that big techs have not taken on a regulatory status that
might restrict their ability to continue their other activities.

2
 Box B

The data-network-activity feedback loop may explain the rapid growth of big tech
credit

The data-network-activity (DNA) feedback loop helps big tech firms to correctly predict clients’ creditworthiness,
future income and individual demands (BIS 2019). As a result, big techs can have a competitive advantage over
banks and serve firms and households that otherwise would remain unbanked. Recent work suggests that big techs’
credit scoring applied to small vendors outperforms models based on credit bureau ratings and traditional borrower
characteristics (Frost et al 2019; Gambacorta et al 2019). All this could represent a significant advance in financial
inclusion and help improve firms’ performance (see Luohan Academy, 2018).

Cornelli et al (2020a) find that in 2019, fintech and big tech credit (together “total alternative credit”) reached
nearly USD 800 bn globally (see Figure B1). Big tech credit has shown particularly rapid growth in Asia (China, Japan,
Korea and Southeast Asia), and some countries in Africa and Latin America. By contrast, fintech credit volumes have
declined in 2018-19 due to market and regulatory developments in China. Outside China, fintech credit is still
growing. These alternative forms of credit are more developed in countries with higher GDP per capita (at a
declining rate), where banking sector mark-ups are higher and where banking regulation is less stringent. Fintech
credit is also more developed where there are fewer bank branches per capita. Moreover, these alternative forms of
credit are more developed where the ease of doing business is higher, investor protection disclosure and the
efficiency of the judicial system are higher, the bank credit to deposit ratio is lower, and where bond and equity
markets are more developed. Overall, both fintech and big tech credit seems to complement other forms of credit,
not to substitute for them, and may increase overall credit access.

Global big tech credit is booming, overtaking fintech credit Graph B1

Big tech credit is overtaking fintech credit1 These alternative forms of lending are becoming a
significant portion of total credit in a few economies
USD bn USD mn, logarithmic scale Per cent

100,00 2.5
0
750 2.0
10,000
600 1.5
1,000
450 1.0
100
300
0.5
10
150 0.0
2013 2014 2015 2016 2017 2018 2019
1 CN US JP KR GB ID NL RU KE DE
Lending volume: Fintech Big tech
Lending volume (lhs): 2 Ratio to total stock of credit (rhs):3
Fintech Total alternative credit4
Big tech
1
2019 fintech lending volume figures are estimated on AU, CN, EU, GB, NZ and US. 2 Data for 2019. 3 Domestic credit provided by the
financial sector. Data for 2018. 4 Total alternative credit is defined as the sum of fintech and big tech credit.
Sources: IMF World Economic Outlook; World Bank; Brismo.com; Cambridge Centre for Alternative Finance and research partners;
WDZJ.com; companies’ reports. Cornelli et al (2020a).

Individuals as financial services providers

Technology will enable individuals to move beyond providing finance via
intermediaries. Today, individual savers provide the deposits that banks use to lend,
and equity capital via investment and pension funds. There is the occasional direct

2
family loan, or high net worth angel investor providing direct finance. Technology
could allow this to be much more pervasive. P2P was an initial step in that direction.
Reduction of the frictions that give rise to intermediation raises the prospect of
radical disintermediation such that individuals become financial services providers
directly. The core of financial services is the transfer of resources between
individuals directly and over time. New technologies could decentralize decision-
making, risk- taking and record keeping (FSB 2019a). Distributed ledger
technologies (DLT) have already demonstrated the potential for transfers of value
without intermediaries. P2P lending was designed as a way to directly transfer
resources and risk, with limited centralized matching services by the P2P platform.
Technology obviates the need for a bank balance sheet by synthesizing a set of
contributing liabilities against each asset (loan) as it is originated. Even these
platforms could be eliminated. Drastic reductions in search costs could allow
individuals at the retail level to connect directly to borrowers. Publicly available data
and analytics applications could be applied to investments decisions and portfolio
construction at the individual level, as robo-advisors are already doing. Smart
contracts could automate monitoring and collection processes and reduce the
coordination costs for a borrower working with multiple small lenders.
To date, consumers have not broadly embraced the role of direct payments
provider in decentralized models. DLT-based alternative payments via
cryptocurrencies are growing but the main use of cryptocurrencies has been
speculation rather than payments. Most individuals using crypto-currency still prefer
to go through crypto-currency wallet providers or exchanges. While decentralized
finance (“DeFi”) and decentralized apps (“DApps”) are growing, these are still a niche
market (Schär 2021). Other venues that might have fostered direct peer-to-peer
exchange such as social media platforms and messaging apps have either offered
payments intermediation themselves or linked to banks.
Similarly, individuals have been reluctant to act as direct lenders assuming full
responsibility for credit screening and lending decisions. P2P marketplaces initially
offered savers the ability to choose individual borrowers. The option for micro-
assembly of loan portfolios is not much used, however; most consumers chose to let
the platforms algorithmically allocate their funds. Behaviour on equity crowdfunding
platforms is somewhat different, though increasingly these platforms are also
offering diversified fund structures to consumers who would prefer not to
individually vet investments. OurCrowd, for example, began with a direct individual
selection model but now operates a full menu of funds to allow investors to
programmatically deploy and diversify.
The direction of technology and developments in decentralized finance,
however, suggests that these trends bear monitoring. While savings and investment
activities still bear a high cost of assembly, part of the promise of technology is the
empowerment of individuals. Technology is reducing the need for intermediaries.
Un- intermediated finance would render any individual with surplus funds a financial
services provider. While the emergence of individuals as direct providers of financial
services is nascent at present, it cannot be discounted as a potential influence on
market structure in the future.

3
3.4 Impact on sector composition, competition, concentration

The diversity of the financial sector has increased in both developed and emerging
markets. Alongside incumbents, there are many different types and sizes of fintechs
and big techs, offering a wide range of financial services. Some remain focused on a
single product or service, while others have leveraged their initial successes to
broaden their service offerings (eg Square and PayPal moving from payments to
lending). Some fintechs are converting to banks, while others have become service
providers to, or value chain partners with banks. Depending on licensing
approaches in different jurisdictions, a range of digital-only or digital-mainly neo-
banks emerged. They compete largely in the same regulatory space as incumbents,
but with modernized business models and streamlined infrastructure. Big techs with
core markets ranging from telecommunications to logistics, transportation, e-
commerce marketplaces, and online search add another set of diverse players to the
industry. Incumbents are also adopting new technology, partnering with fintechs,
setting up new digital units and otherwise adapting to the new environment. This
hyper- diversification of financial services has implications for competition, as well
as for regulation (see section 4).
Fintechs have significantly increased competitive pressure by directly engaging
with the consumer. The battle to ‘own the customer’ will be closely fought, since in
most industries the links in the value chain closest to the customer earn the highest
margins (eg European Banking Authority (2018), OECD (2020), Petralia (2020)). Some
fintechs have skilfully targeted high-margin products or services and stripped those
away from incumbents, leaving the incumbents with a lower-margin product mix.
For example, companies like Wise (previously Transferwise) and Remitly targeted the
relatively high-margin international transfer business. However, many fintechs are
struggling to compete at scale. Through re-bundling, some digital-only banks may
be able to broaden scope and compete at scale, but at the cost of having to
maintain a broader product mix in which some products or customers will generate
lower margins.
Disruption may be contained in areas where fintechs themselves relying on
incumbents for their operations. As described above, many fintechs still rely on
banks to store value and process payments, and users often originate fintech
transactions from an account with a financial institution. Products like Wise rely on
existing bank or card accounts not only for the cash-in/cash-out (CI/CO) functions,
but also for a layer of KYC compliance. In some markets, digital financial services
providers have built their own infrastructure for some of these functions (eg mobile
agent networks for CI/CO). The introduction of stablecoins and CBDCs might reduce
the dependence of fintechs on incumbents, particularly for payment services.
For fintechs that target back-office activities, the relationship with incumbents
will be collaborative. For example, B2B fintechs offer services to incumbents in the
form of partnerships or SaaS engagements. BaaS providers present a distinct
competitive challenge. These regulated banks enable a range of businesses,
including fintechs, to compete more effectively with other banks. The ability to
embed a tailored payment, loan, insurance or other financial service into any
economic, business, or social activity may be the most powerful disruptor of
traditional financial services, and this is an area in which big techs excel.
Big techs have the potential to compete at scale and become dominant players.
They already enjoy economies of scale and scope and network effects in their core

3
markets. By adding financial services to their ecosystems, on their own or using
BaaS, they can leverage these economies to compete with traditional providers and
fintechs. They also have the clout and deep pockets to navigate complex regulations
and build parallel infrastructures and closed-loop systems, which could solidify their
role as gatekeepers to the consumer (and their relevant data) in the financial sector.
The resulting “winner-takes-all” or “winner-takes-most” dynamics may reduce
competition and contestability of the sector and trigger monopolistic behaviour as
customers become locked-in. Some big techs (eg AliPay) fortify customer lock-in by
allowing funds to flow into their systems for free but charging a fee for withdrawals.
They may tie the availability of a loan to consumption of other platform services, or
cross-subsidize financial services with revenues from their core business. This could
result in consumer benefits but destructive competition for the rest of the sector
and potentially in other sectors. While there are ways regulators can attempt to re-
level the playing field,10 the intermingling of financial and non-financial services
presents a number of challenging financial regulation and competition policy issues
(see below).
Incumbents large and small are embracing digital transformation across the
value chain to compete with fintechs and big techs. Competitive pressure on
traditional financial institutions may force even those that are lagging to transform
or risk erosion of their customer base, income, and margins. In addition to
partnering with fintechs, some have set up their own digital-only banks. Where
regulation permits, as in Thailand, some banks have countered the big tech
incursion into banking with bank-centric e-commerce marketplace offerings (for
example, Kasikorn Bank’s K+ market). Formerly captive local markets have become
highly contestable by international players (eg DBS entry to India via Digibank).
Local banks or those serving specific industries may still be competitive in their
niche markets, if they can manage the transition to digital delivery and maintain
customer trust and loyalty, and the other competitors don’t turn their focus to those
niches.
The largest markets might be expected to receive more attention from
innovators, but small markets are affected as well. Overall, fintech adoption seems
to be higher in higher-income markets (Didier et al 2021), and these will often be
more attractive for Venture Capital investment. However, some mid-size countries
have been ahead of the digital finance curve, perhaps because they had not
received sufficient investment in traditional financial services and there were more
underserved segments or larger margins to capture. Kenya’s ecosystem of digital
payments, savings, and lending is a prominent example. Innovators in small
countries can leverage the cost reductions discussed above to enter markets that
may not have been viable with a higher cost structure. For some services, however,
small markets may lag in terms of home-grown innovation. Innovations centred on
data use may advance faster in larger markets that generate more data (World Bank
2021). Emerging and smaller markets could nonetheless benefit from incumbent
transformation or geographic extension. The local incumbents can adopt new
technologies to improve reach and efficiency, and large regional telecom and e-
commerce players could extend financial services into these smaller markets. The

10
In Thailand, for example, regulators responded to e-commerce entry into financial services by
permitting banks to operate e-commerce marketplaces and offer certain IT services to their clients.
See Bank of Thailand (2019).

3
ability to serve remotely could make a small market part of a larger market,
provided the regulatory frameworks permit.
More intense competition among a more diverse set of industry participants
could result in a range of possible market outcomes. As different institutions
compete, cooperate or do both, industry structure will evolve, depending on which
combination of three forces prevail in a given market.
1. Economies of scale and scope: where production costs remain significant relative
to revenues (or cross-subsidization is possible), and network effects and
switching costs or convenience effects are significant, we may expect
consolidation into a set of large players who can leverage scale in capital,
connectivity, and data to offer a broad range of products at high volume and
low cost. This could potentially also result in displacement of incumbent
financial institutions by big tech firms, in particular in the retail banking domain.
The conditions for “natural monopolies” could result from strong scale
economies combined with low marginal costs of service provision. These are
countered by the reduction in fixed costs that allow for smaller scale entry.
2. Preferences for choice or convenience: As we have seen, alongside the
economies of scale achievable by big players, technology has enabled
unbundling of products and services. These may vary along price, quality, and
service dimensions. Customer can choose a custom mix of separate products
and self- aggregate. This is more likely when customers are extremely price-
sensitive or value more tailored services, and search and switching costs are
low. Niche providers can then focus on product-market fit for a target customer
base, and that customer base can easily find the product and migrate to it. The
niche providers can thrive, taking advantage of their ability to manage
infrastructure costs while reaching and retaining their target customers
wherever they may be. The unbundling effects could dominate, with a resulting
atomization of the industry. In this scenario, there would be space for new
players that enable aggregation of the services and those that provide common
services to the entire market. On the other hand, too much choice can
overwhelm individuals and undermine the benefits that choice can provide
(Schwartz, 2005). Search and switching costs may reduce the convenience of
self-aggregation. Customers may then prefer banks or platforms that integrate
services for them, even if they end up paying more or accepting inferior service
for one or more products.
3. Regulation: The above two forces will not play out independently from the
actions of financial services and competition regulators, who will continue to
shape outcomes with a variety of policy instruments.
Given the diversity of providers and consumers, the corner solutions of large
multi-product players and small focused players may co-exist. As shown in Graph 8,
the tendency of providers to seek scale (leverage scale, scope, and network
economies) alongside consumer preferences for choice could result in four potential
configurations, of which two appear to be dominant outcomes. A small player (left
bottom corner) attempting to provide a broad range of products will find it difficult
to deliver fully across that product range, and will be pressured to either merge with
others to grow large (left upper corner) or to focus on niche services (right lower
corner). A large company with an initially narrow scope (right upper corner) will find
that its size enables it to broaden, and will tend to add products (left upper corner).
For instance, a payments firm may choose to add lending or insurance services,
while an international money services provider may start offering checking
accounts.
3
Corner solutions of scale vs. atomization Graph 8

Degree of
Scale, Scope Large Multi-Product
and Network Large
Economies Narrow Scope
Size enables broader product scope; tendency to shift left
(advantages
to size)

Small Narrow Scope

Small
Multi-Product
Difficult for small player to deliver across range of products: grow or focus

Degree of Choice Preference

(customer tendency to
unbundle)
Source: authors’ elaborations.

Thus the financial services market may coalesce around one set of providers
adopting large/multi-product strategies, and another set offering focused products
and services. This might be termed a “barbell” distribution, with a small set of large
multi-product players on one side, and a large set of small niche service players on
the other.11 While this market structure reflects in part the “winner-takes-all” or
“winner-takes-most” character of some tech-driven markets, the impact of friction
and cost-reducing innovation in finance leaves the door open for smaller players to
thrive. If regulation permits, this may mean continued market fluidity from new
entrants and cross-overs from one sub-sector to another.
The large, multi-product players could include traditional players, fintechs and
big techs – ie both incumbents and new entrants (Graph 9). These may include
banks that master the digital transformation, along with new financial services
entrants that achieve scale, and big techs or other firms that have customer data
and can link to BaaS. A recent survey indicated that large banks are ahead of smaller
banks in terms of digital transformation, and are overall more positive about the
impact of technology on their businesses (Feyen et al 2021). Banks that pull ahead
in terms of digitization will be able to leverage the DNA feedback loop to grow
market share relative to other banks. Some technology-savvy and agile smaller or
newer banks and bank-like players could also ride this wave to grow. 12 Big techs, on
the other hand, enter the large size of the barbell on the strength of the customer
base they bring

11
So-called ‘barbell’ distributions, with clusters on two ends of the scale, have been observed in many
markets, including labour markets (Autor 2019), product pricing and quality (Britton, 2018), and
portfolio strategy (Rutherford 2013). Such distributions have been observed in banking markets in
previous periods or disruption, for example in the wake of mega-mergers in the US in the 1990s
(Deogun 1996) and the financial crisis of 2008.
12
For example, NuBank (a payment services provider) in Brazil went rapidly from challenger to 30
million customers, placing it in the top ranks in that market. Cross-River Bank in the US has just one
branch, but leveraged fintech to be one of the biggest lenders nationwide for the US Government’s

3
pandemic response program (Cowley 2020).

3
 from their core businesses, and the ability to convert those relationships into
financial service product offerings at scale.

Potential barbell financial services market (illustrative)13 Graph 9

Source: authors’ elaborations.

The small, niche players could also include some traditional providers alongside
fintechs (Graph 9). Niche players may directly serve customers with a focused
product set, or may serve large players with B2B services. Small local institutions
such as community banks and credit cooperative banks, for example, occupy
geographic niches and fill gaps not served by larger institutions, and may have
significant trust capital as a competitive advantage (Mester 2018). A recent study of
regional banks in the US found that many have been able to leverage technology
and geographic focus to thrive against large scale players (Alexopoulos 2020). Many
fintechs enter the market as single product providers or with a product tailored to a
specific market segment, such as small businesses in a given industry or set of
markets. Unbundled fintech service providers with a strong value proposition can
viably serve a narrow market given the ability to reduce fixed costs using flexible
cloud-based infrastructure, digital channels, and other service delivery innovations,
and build customer affinity with tailored products. De-centralization through DLT
and other innovations will enable service providers to build on these new products
and infrastructures to provide additional unbundled services. Small niche players
could thus continue to thrive as they tailor their products and services to a specific
market segment or customer base. As noted above, there are scenarios where
individuals could act as direct providers of certain products or services; these would
be extreme instances of small niche players.
There is likely to be a degree of connectivity between the two ends of the
barbell. Some narrow, tailored services will partner with large institutions to provide
their specialized services to the customers of the larger institutions, while big techs
and other platforms may provide some of the connectivity that allows small fintechs
to reach customers. The size of both ends of the barbell are also influenced by
regulations that might limit the size of a particular firm.

13
Graph 9 is purely illustrative and does not constitute endorsement of particular companies’
potential roles in financial services or projections as to their future size or success.

3
 A barbell is a potential steady state market structure, as some participants grow
larger but innovation drives continued entry. During the current transition phase,
small multi-product providers and mid-size players continue to operate, increasingly
challenged to compete against the scale players or to lock in a focused niche. Re-
bundling by the larger formerly single-product fintechs is well underway in the more
advanced markets, and entry by big techs is adding to the large player end of the
distribution directly. Entry can continually replenish the small player end of the
distribution. Large players will inevitably leave some markets less well-served and
new technologies will continue to reduce costs or offer the prospect of new
products. As long as regulatory barriers are not raised, entrepreneurs will continue
to bring innovative new products and services to the market; these may not
compete across the board with the large players but provide a reservoir of
contestability to counterbalance the concentrated end of the barbell. As illustrated
in Graph 8, there will be a tendency for players to either focus or to move to the
large/multi-product space. Left to market forces, the dominance of the two sides of
the barbell may very well squeeze out the middle (except insofar as a company
passes through on its growth path from one side to the other). Those providers that
cannot achieve the scale to compete with the larger providers but fail to focus on a
sustainable niche may be forced to exit. A recent survey by FIS on new banking
relationships established in the US in the last 12 months showed that 37% of new
banking relationships went to large banks and 48% went to niche and neo banks
(community banks, credit unions, online banks and non-bank entities); the in-
between regional banks got only 10% (see Marous 2021).
A barbell is not the only possible outcome, but is a central case given the
economic forces at work. Continued atomization, stalled re-aggregation, or limits on
entry could result in a different configuration. Factors that will contribute to
determining this outcome include core market forces such as the economies of
scale and scope discussed above, regulatory limits on entry or on re-aggregation by
non- banks, and consumer preferences for trusted incumbents or for convenience
and internal product linkage vs. tailored solutions and more direct control. Societal
preferences will matter and will be expressed not only by consumers but also in the
degree to which policies for, and regulation of, financial services, competition, and
consumer protection shape both market forces and consumer behaviour.

4. Policy implications and new trade-offs

National authorities aim to foster the benefits of digital transformation in finance.

These include reduced costs and frictions, increased efficiency and competition,
narrowing information asymmetry, and broadened access to financial services. At
the same time, authorities are aware of the potential risks to the financial system
and to customers (FSB 2017). A policy agenda (“Bali Fintech Agenda”) outlined by
the World Bank and IMF advocates embracing the promise of fintech while
managing risks to consumers and to the stability and integrity of the financial
system (World Bank/IMF 2018). In particular, reinforcing competition and a
commitment to open, free, and contestable markets were seen as a foundation for
innovation and consumer choice while maintaining a level playing field. The Agenda
also recognized the need to adapt regulatory frameworks and supervisory practices,
including monitoring of new market participants, to facilitate the safe entry of new
products, activities, and intermediaries while maintaining financial stability and
sustaining trust, confidence, and the ability

3
 to respond to risks. The importance of robust data infrastructure, both from a
cybersecurity perspective and with respect to concentration risks, data ownership,
protection, and privacy, was also noted.
Reinforcing competition requires taking both horizontal and vertical views of
the financial services landscape, and cooperation between financial and competition
authorities. Horizontally, how a market is defined is fundamental to the assessment
of competition. The multiplicity of financial products, the tendency to bundle them,
and the role of multi-sided market dynamics in providing and pricing all make this a
more complex landscape. Vertically, a competitive market at the consumer level may
mask concentration at different levels of the reconfigured financial services value
chains. With finance increasingly embedded as an adjunct in other businesses, it
becomes critical for financial regulators to coordinate with other regulators,
particularly where the adjacent businesses are themselves highly concentrated or
subject to network effects. Consideration of ex ante vs ex post remedies may need
to recognize the particular features of data concentration, the power to exclude, and
the gateway functions inherent in emerging digital business models.
Adapting regulatory frameworks will require recognizing new and sometimes
heightened trade-offs between specific policy goals. Changes in provider diversity
and the competitive landscape may imply a need to revisit the regulatory perimeter
and supervisory practices. Meanwhile, authorities may need to make tough choices
between: (i) financial stability and market integrity, (ii) efficiency and competition,
and
(iii) data privacy and consumer protection. It is important to find policies that can
mitigate these trade-offs and allow priority goals to be achieved simultaneously to
the degree possible. In the next sections we explore the implications for the
regulatory perimeter, supervisory practices, and competition, and the relevant
trade- offs between the policies that authorities are adopting in practice.

4.1 Implications for the regulatory perimeter and supervisory

practices

The growing diversity of financial services providers and business models often
requires expanding the regulatory perimeter. Payments, loans, and deposit taking
services may be provided by specialized payment service providers (fintechs), e-
commerce platforms (big techs), and other non-banks. It is therefore important that
regulators develop approaches to ensure a level playing field and provide clear
requirements for licensing. Similar activities and similar risks should in principle be
treated similarly, regardless of the market participant, underlying technology, or
method by which the service is provided. Yet in practice, a purely activities-based
approach to regulation may not be sufficient. Particularly the entry of big techs may
require more, not less, entity-based regulation, for instance to address risks around
competition and operational resilience (Restoy 2021). Expansion of the regulatory
perimeter can ensure that activities are appropriately and comparably regulated and
supervised (World Bank and IMF 2018; Alonso et al 2021; Ehrentraud et al 2020).
This may extend to the perimeter of financial safety nets. Data reporting perimeters
may need to be expanded even further, given the disaggregation of finance and the
embedding of financial services into non-financial activities. The adoption of
targeted regtech and suptech solutions may merit consideration by regulators as
part of the policy response to fintech developments (Broeders and Prenio 2018).

3
 Expanding the regulatory perimeter may be challenging in practice. Bringing
new entities into the fold of financial regulation may require legislative changes,
which could be considered controversial and may be resisted by powerful interests
with deep pockets, such as big techs. There may be resistance to granting more
discretionary powers to supervisors, even though such powers may be needed to
continually adapt policy approaches to a changing sector. The challenges may be
even more acute in EMDEs, where resource constraints make it more difficult to stay
abreast of changes and to regulate powerful (foreign) service providers. Yet
emerging policy approaches in several jurisdictions show that applying an entity-
based approach to new players such as big techs may be possible (see section 4.2).
The disaggregation and reconfiguration of finance value chains introduces new
challenges in day-to-day supervision. For a traditional, vertically integrated bank, it
is clear where responsibility lies for the financial soundness, cybersecurity, and
consumer impacts of a product. When the financial services value chain is spread
across different players with, for example, one holding the customer relationship,
another holding the customer funds, a third providing data analytics and deciding
which customers get services, and a fourth providing technology infrastructure, it is
more difficult to pin down responsibility for mishaps or misdeeds and to ensure that
consumers’ interests are protected. Moreover, when customer interactions are
handled by entities that are not directly subject to extant consumer protection
regulations, any shortcomings such as unfair practices may not surface through
traditional supervision focused on the regulated entity. This can lead to undetected
consumer protection risks. This becomes even more challenging when services are
provided across borders.
Complex processes and interactions across different players and systems can
create new points of failure. Multiple entities and interlinkages create a wider attack
surface for cybercriminals, which requires a strong regulatory approach to promote
cyber security. The balance of power in these partnership relationships is very
different than in a traditional outsourcing relationship, so the ability of the regulated
financial institution to enforce its own policies might be challenging. Multi-tiered
access to regulated systems, for example through BaaS models, requires that
regulators ensure that the BaaS provider has sufficient visibility through the value
chain to ensure compliance. In some circumstances, regulators may need to extend
their supervision past the first-tier regulated entity, downstream to the customer
interface entity or upstream to infrastructure providers.
Reconfiguration of value chains is creating concentration risks at the
technology services level that need to be monitored from both competition and
systemic stability perspectives. The cloud services market is highly concentrated.
While provision of services to financial institutions has long been subject to
concentration (for example, limited numbers of mainframe, ATM, cash transport,
and payments network providers), growing reliance by a large swath of the financial
sector on a small number of cloud services providers has been flagged as carrying
the potential for new concentration or single-point-of-failure risks (FSB 2019b).
Notably, four players control around two thirds of the global market for cloud
services (Graph 10). While cloud providers generally have deep expertise in systems
architecture and cyber security, an operational or cyber incident at one major cloud
provider could have systemic implications for the financial system (Danielsson and
Macrae 2020). Open infrastructure, including API hubs, KYC utilities, and changing
access policy for existing payment systems and credit reporting infrastructures, can
mitigate concentration risks, increase contestability, and dilute data concentrations.
However,

3
 there are indicators that a number of B2B fintech services may be equally prone to
concentration; the example of API middleware providers cited earlier is a case in point.

Market share of leading firms in the cloud infrastructure-as-a-service (IaaS)

market, across all industries (Q4 2019)
In per cent Graph 10

23
33
2
2
3
5
6 18
8

Amazon Web Services Microsoft Azure Google Cloud IBM Cloud Alibaba Cloud
Salesforce Oracle Cloud Tencent Cloud Others
Source: Synergy Research Group.

As authorities contemplate the regulatory and supervisory response to fintech,

developments need to be examined in light of the needs in each market. Some of
the issues related to disaggregation and the incorporation of new players into
financial product value chains can be addressed through regulatory guidance on
governance of partnerships and outsourcing, including clear allocation of
responsibilities and supervisory reporting, applied to each individual institution.
Traditionally, regulators have placed responsibility for regulatory compliance with
the licensed institution. However, as mentioned before, this may not work in a
scenario where the licensed institution has much less scope to enforce specific
requirements – for example a small niche player partnering with a big tech or using
a global cloud provider. Thus, many challenges, such as up- or downstream
concentration risks, may need to be addressed at the sector level, eg through
enhanced oversight of systemically important providers. Further, regulatory
authorities might need to increase their ongoing scrutiny of the operational and
process model of licensed institutions and introduce changes to ensure that they
are able to discharge their consumer protection responsibilities.

4.2 Implications for competition policy

Concentration risks in financial service provision may increase under the central
scenario outlined above. Digital platform providers in particular may quickly achieve
dominant positions. On its own, this need not have a detrimental effect on
consumer choice and welfare, even if a small number of dominant players reduces
coordination barriers and may increase the risk of collusion. Provided switching
costs remain low, markets can remain contestable and deliver beneficial outcomes –
even in a barbell market structure. On the other hand, the concentration of big data
may result in data monopolies. Data-driven advantages of larger players can
increase switching costs and effectively lock in certain customers, and ex post
remedies for abuse of this customer lock-in may be difficult. In such cases ex ante
remedies by the sectoral regulators on the ability to switch or interoperability
requirements are better suited.

4
Graph 11 compares ex ante sectoral regulations with ex post competition
interventions. As an example, the largest two mobile payment providers in China
reached a combined market share of 94% in 2019 (FSB 2019). The authorities were
concerned that this could also pose severe systemic risks if one such provider were
to be impaired. In 2018 the People’s Bank of China (PBC) implemented a new
mobile payment regulation requiring all mobile payments to be cleared through the
PBC rather than settled internally within the mobile payment platform provider (Liu
2019). In this way, a sector-specific ex ante approach was able to address concerns
rather than waiting for ex post correction.

Interplay of ex ante sectoral regulations and ex post interventions by competition

authority Graph 11

While overall financial services may remain competitive, in certain specific

product segments monopolies or quasi-monopolies could emerge. Some services,
such as exchanges, have inherent network and scale economies that could result in
only a very few viable providers. While this is not new, the entry of companies from
adjacent spaces or the embedding of financial services into non-financial products
means the market structure of those industries becomes relevant. A monopoly
utility or logistics company could require customers to use its embedded payment
instrument. It could also embed working capital lines that could not be obtained by
most customers from other lenders who do not have the data and leverage of these
providers. This would allow them to effectively dominate a financial services
segment by virtue of the control held in the primary industry. The movement
of highly

4
 concentrated mobile network operators into payments, often setting up their own
proprietary agent networks, is an example of how a player from concentrated
adjacent market that is regulated with respect to competition can carry that market
structure over into a financial product area where the regulation was not attuned to
that issue. Kenya provides a case study of regulatory response to concentrated
market structure and anticompetitive behaviours in digital payments (see Box C).

Box C

Competition in digital payments: the case of Kenya

The Competition Authority of Kenya (CAK) is a state corporation and was established by the Competition Act No. 12
of 2010. It is mandated to promote and safeguard competition in the national economy by effectively enforcing

PROBLEM
ACTION RESULT
 Safaricom was found to
 Safaricom, in a  Agent profitability
have abused its dominant
settlement with CAK, increased by 10%;
market position by
agreed to open up M- consumer savings
entering into exclusive
Pesa agents to all market increased by USD
contracts with M-Pesa
players 33.2bn; bank networks
agents
expanded from 5% in
 Safaricom, in a settlement
 Safaricom was found to 2013 to 15% in 2014
with CAK, agreed to lower
have abused its dominant
service charges and to  Consumers can switch
market position by
inform PSPs of transaction between services as
practicing excessive
price via real-time they can compare
pricing and a lack of
notification after initiation costs; decrease in
transparency on prices
but before completion services charges from
for transactions going
1/2 to 1/10 of original
through its network  In April 2018,
charge
interoperability was
 Safaricom was found to
launched. The three  Transaction costs
have abused its dominant
relevant authorities dropped; business of
market position to
(Communication Authority, competing PSPs
impede account-to-
CAK and the Central Bank increased
account interoperability
of Kenya) cooperate to
facilitate interoperability

competition law and rules economy-wide, to open markets by advocating for the removal of anticompetitive sector
regulation, and to protect consumers and suppliers. CAK has intervened in a number of sector cases. With specific
regard to the market for digital payments, the shapes below report the identified issues, the policy actions taken,
and the results achieved.

Source: Adapted from Bossone and Natarajan (2021).

The proliferation of new products has not always resulted in competition at all
levels of the restructured value chains. The consumer-facing end of payments has
become more diverse, with a range of wallets, affinity and prepaid card programs,
and embedded payments offerings. At the underlying transaction processing level,
though, there is an oligopoly in many markets. In 2016, the UK Payment Systems
Regulator (PSR) reviewed the ownership and competitiveness of infrastructure that
supports the three major UK payment systems – BACS, Faster Payments Service, and
LINK – and found there was no effective competition for the provision of UK
payments infrastructure (UK PSR 2016). Various measures were proposed to
increase competition and remove barriers to entry.

4
 A similar challenge is emerging in the context of open banking and APIs. The
back-end connection to APIs to facilitate data exchange is a high-volume, low-
margin business. But since each bank may have different API specifications, it can be
costly for a specific provider, such as a fintech start-up, to develop connections to
everyone. Aggregation occurs at a middleware level through players like Plaid, Tink,
and Yodlee. These are emerging as dominant networks of API connections, a
potentially oligopolistic layer at the choke point between more competitive
segments. Even in markets where APIs are standardized, API hubs are emerging to
simplify the interface requirements.
While unbundling would tend to reduce anti-competitive product tying, re-
aggregation and embedding potentially put tying at the centre of financial product
economics. In particular, platform models that combine free services that have
network effects with financial services might represent exactly the type of situation
anti-trust authorities were concerned with in establishing anti-tying rules: that a
monopoly provider of one service might require customers to use its other
products. Thus a dominant social network platform that has a quasi-monopoly
position over small local businesses’ connections to their customers might embed
payments in the social networking experience, and make it difficult for a customer
to pay the business through anything other than the network’s payment product.
The more dominant the social network, the more necessary it is for a business to be
present on that platform, effectively reducing competition in payments acceptance.
Dominant technology companies have already run afoul of anti-tying regulations
(eg Apple, Microsoft). While regulators have moved away from considering tying
arrangements to be illegal per se, and focused more on rule-of-reason analysis that
looks at both the benefits and potential costs of tying, these analyses are sure to be
tested as financial services tying moves from the "traditional bank product
exception" space to the sphere of big techs and embedded finance.
Market structures that concentrate data and supercharge network effects could
reduce intermediation costs and broaden inclusion. In many markets, however, the
resulting market power might be seen as more detrimental than these benefits. For
example, in November 2020 the US Department of Justice sued to block the
proposed merger of Visa, a payments company, and Plaid, a fintech data
aggregation platform; the merger was abandoned in early 2021. Chinese regulators
have recently addressed concerns of anti-competitive practices that exclude
competitors in associated digital services such as e-commerce and social media
when providing services to customers (McMorrow 2021). As in other industries,
regulators will have to balance the efficiencies of natural monopolies against
restraining market power. This is not a new consideration in financial services, but is
just one of several trade-offs that become more pronounced as technology re-
shapes financial services.
Different societies will also attach different preferences to market structure
outcomes. Regulatory policies governing entry and scope of activity of new players,
from small fintechs to big tech platforms, will have profound effects on competitive
dynamics both in financial services and in the core industries of cross-over players.
Fintech and big tech will require changes in the way that competition authorities
and financial regulators interact, and calls for more coordination between regulators
(Petralia et al 2019). Approaches to competition policy differ starkly across societies.
Some societies may welcome market structures that concentrate data and
supercharge network effects if they reduce intermediation costs and broaden
inclusion. In other markets, the resulting market power might be seen as more
detrimental than these benefits. Concentration of infrastructure and data in state

4
 hands may be accepted in some societies, while others may be more concerned
about potential extension of state surveillance. As in other industries, regulators will
have to balance the efficiencies of natural monopolies against restraining market
power. This is not a new consideration in financial services, but is just one of several
trade-offs that become more pronounced as technology re-shapes financial
services.

4.3 New dimensions to financial regulation policy trade-offs

Digital innovation may thus give rise to complex trade-offs between three
categories of policy goals. These are: (i) financial stability and market integrity; (ii)
efficiency and competition; and (iii) data privacy and consumer protection. Graph 12
gives a visual representation of these trade-offs as a triangle. This final section
discusses the three edges of the triangle in turn – and which types of policies
authorities are applying to mitigate these trade-offs.

Policy trade-offs from digital transformation in finance Graph 12

Source: authors’ elaboration.

Efficiency and competition vs financial stability and integrity

Regulators have long debated the relationship between competition and financial
stability. There are broadly two schools of thought on the competition-stability
nexus (red arrow in Graph 12). One school of thought argues that greater market
entry, or at least contestability, in the financial sector is desirable, as it fosters
competition and reduces incumbents’ market power. This can lead to greater
efficiency and better outcomes for customers. It can also lead to more diversified
activities (Amidu and Wolfe 2013) and fewer institutions that are individually
systemically important and must be treated as “too-big-to-fail.” Systemic
importance can introduce governance challenges and moral hazard, as well as a
financial sector that is, on aggregate, larger than would be economically optimal
(Feldman 2010). An alternative school of thought emphasizes that greater
competition is not always optimal or conducive to financial stability. Financial
institutions that are more profitable may be able to accumulate

4
more equity capital, have a higher franchise value and therefore act more prudently
(Keeley 1990). This may occur at least up to a point where they incur moral hazard
by becoming too big to fail. An implication may be that some barriers to entry are
justified to help mitigate risks (Hellman et al 2000). This may be particularly true in a
period of rapid market transition. New entrants can target high-margin products, or
skim off the most profitable customers, leaving incumbents with loss-making and
riskier businesses. Regulators may view managing incumbents’ adjustments to
either adoption of new technology, consolidation, or exit with a slower pace of entry
or dampened competition as more welfare-sustaining than rapid and potentially
disruptive exits.
Fintech and big tech entry may turn some existing insights on their head. In
particular, the assumption that entry increases competition may no longer be true in
an environment supercharged by the DNA feedback loop (see Box B). While there is
agreement that contestability matters greatly for the overall degree of competition
(Claessens 2009), and the reservoir of contestants represented by the small niche
services side of the barbell would tend to maintain at least some competitive
pressure on the larger players, big techs with control over key digital platforms in e-
commerce, search or social media may establish and entrench market power very
quickly after entry. Such control may also generate conflicts of interest when both
big techs and their competitors (eg banks) rely on the same core infrastructures
such as networks or cloud computing services.
The traditional focus of competition authorities on a single market, firm size,
pricing and concentration may no longer be well suited to these market outcomes.
Indeed, digital platforms that pursue growth over profits, and data over revenue,
may not be well-captured by competition policy approaches focusing on prices and
consumer welfare (Khan 2017). While indicators of market power like large market
share and high margins provide useful first-order indications, they are insufficient to
determine whether a provider has market power and has the incentives to abuse it.
Moreover, the speed of changes to market structure presents challenges, as firms
can go from “too small to care” to “too large to ignore” to “too big to fail” in only a
few short years (Arner et al 2017). This was apparent particularly with the growth of
the big techs in China, but similar growth is being seen in many other markets,
particularly EMDEs. Large foreign players in particular may become “too powerful to
regulate and supervise.” Ex post remedial action, as with traditional antitrust and
mergers policies, may not be sufficient to address risks.
This may be a reason that some jurisdictions have upgraded their rules and
methodologies for assessing and addressing anticompetitive conduct. In India, for
example, the main e-commerce platforms are prohibited from selling products
supplied by affiliated companies on their websites – including financial products –
to avoid potential conflicts of interest (BIS 2019). In the EU, proposals for a Digital
Markets Act (DMA) and Digital Services Act (DSA) include ex ante requirements on
“gatekeepers” to prevent self-preferencing and bundling or tying, and requirements
on “very large platforms” around reporting, risk assessment and crisis protocols
(European Commission 2020a,b). In the UK, the Digital Competition Expert Panel
(“Furman Report,” HM Treasury 2019) recommended the establishment of a digital
markets unit to develop a competitive code of conduct, enable greater personal
data mobility and advance data openness. In the United States, a US House (2020)
committee recommended rules to prevent discrimination, favouritism, and self-
preferencing, and to strengthen antitrust laws and enforcement. In China,
meanwhile, the State Administration for Market Regulation (SAMR) published draft
guidelines on

4
 antitrust measures for internet companies (including big techs), proposing to block
anticompetitive behaviours such as exclusivity clauses, price discrimination and
market barriers (Restoy 2021).14
There could be stability challenges in a barbell outcome with respect to both
large and small players. As in the traditional competition-stability nexus, greater
franchise values may create incentives for large players to act prudently and
mitigate risk (such as those around cyber security, where the expertise of big techs
is clear). However, systemic importance (too big to fail) may entail pressing risks to
financial stability. Meanwhile, small, specialized players may help to serve specific
market segments, but their lack of diversification may also mean less stability. The
niche providers would not have the diversification to survive a demand shock
affecting their segments, and a specific market segment dependent on only a few
niche providers could be particularly hard hit when a specialized player exits. In
addition, costs to authorities to effectively supervise a large number of small entities
is significant; this would be particularly burdensome for many EMDEs.
Authorities may also need to revisit existing rules on the relationship between
banking and commercial activities. In some jurisdictions (eg the UK and Japan), it is
common for non-financial institutions to own financial institutions subsidiaries,
while in others (eg the US) there is an explicit separation between banking and
commerce.15 As the boundaries between digital financial services and other services
blur, there may be new challenges. For instance, would a big tech that partners with
a small bank and provides financial services have indirect power to influence lending
decisions without having to have an ownership stake? Can cross-subsidization of
financial and non-financial services introduce systemic risks, either from the “real” or
financial side of such links? These and further aspects may require additional
attention.

Efficiency and competition vs privacy and consumer protection

As data become an even more important source of market power, a trade-off
between efficiency / competition and privacy / consumer protection arises. This is
represented by the blue segment in Graph 12, and has implications for financial
inclusion and market structure.
The massive volumes and types of data powering new financial services
offerings have unique characteristics. One special aspect of the use of big data as an
input to production is that data can be used many times and by any number of
firms simultaneously, without being depleted. This is referred to as “non-rivalry.”
Data are often generated at zero marginal cost as a by-product of other digital
services. Re- use and exchange of data can generate large welfare gains, but market
mechanisms may result in re-use and exchange at below optimal levels (World Bank
2021).
The control by providers of their customers’ data maintains privacy but can
reduce competition. The non-rivalry of data can generate increasing returns for data
in both scale and scope (Farboodi et al 2019; Boissay et al 2021). Any single
additional piece of data (eg payments information or a client’s review on a firm’s
product) has

14
Meanwhile, regulatory sandboxes may facilitate new entry into the market. Cornelli et al (2020) show
that in the UK, firms in the regulatory sandbox were able to attract greater funding given lower
information asymmetries and regulatory costs.

4
15
In China, rules stipulate that big techs may only hold a 30% share in financial institutions – hence the
30% stake of Alibaba in Ant Group, and of Tencent in WeBank.

4
additional value when it is combined with an already existing large stock of data. For
this reason, data are more valuable to big tech firms and other firms with a wide
range of business lines, giving rise to so-called digital monopolies. Data may thus
become an “essential facility” in an antitrust sense. Fuelled by big data, algorithmic
price discrimination is able to parse the population of potential customers into finer
and finer subcategories – each matched with a different price. Large multi-product
providers could in principle capture the entire consumer surplus (Bar-Gill 2019).
In the absence of privacy concerns, wider sharing of data could in principle be
more efficient. If markets are competitive, it could be socially desirable to enable
data sharing with providers able to make better use of the data, for example as
manifested by being able to sell a product to the customer and therefore being
willing to pay the data producer for access. This could increase efficiency and
personalization and enhance competition and financial inclusion. This consideration
has been an important driver behind open banking rules, which allow customers to
share their data between providers (eg banks and fintechs), and thus support
competition.
Privacy concerns are valid, however; free sharing of data can harm individuals.
In the case of personal data, open access to data can bring important costs, as
individuals value their privacy (Acquisti et al 2019). Bank transaction data can be an
important input into credit scoring decisions, and it can reveal a wide array of
information about consumers’ preferences, buying habits, location and a host of
other attributes, particularly when combined with other data sources. Open access
to personal data can allow for identity theft, reputational damage and the
manipulation of behavioural biases to sell consumers products that are not in their
interests – to name only a few examples (Ru and Schoar 2016). Open access across
providers could also leave consumers worse off through high, personalized prices. 16
In insurance markets, greater data could allow for more granular pricing, thus
undermining the risk pooling or solidarity function of insurance (IAIS 2017).
Control over personal data may be subject to market failures. These come
about as data are non-rival and only partially excludable (Carrière-Swallow and
Haksar 2019), and due to challenges in obtaining reliably informed consent. This can
invite too much data sharing (Acemoglu et al 2021), hoarding of data, too little
investment in cyber-security, and fragmentation across borders. While it would in
principle be optimal to define “property rights” over data – a decentralized or
“Coasian” solution, in line with Coase’s classic theory of property rights (Coase 1960)
– this may not be easily applied in practice. In particular, large, multi-product
institutions may have access to types of data that others do not have, and given
returns to scale and scope they may be able to bid more for such data than
specialist providers. Thus, data can be the source of further competitive distortions.
In addition, consumers cannot always fully process available information, and the
information available when they grant

16
In a recent paper Brunnermeier et al. (2020) argue that big data and machine learning invert adverse
selection problems. Artificial intelligence allows insurers to infer many statistical insights from the
insurance customer and thereby reverses the information advantage from the customer to the
insurer. This determines an “inverse selection.” A customer knows her multiple attributes, but a
platform and only a platform can connect them. Given this comparative advantage, the insurer can
try to protect its statistical information by offering only a few screening contracts. In this case,
forcing the insurance company to reveal its statistical information to the customer can be welfare-
improving. Their research shows indeed that in a setting with naïve agents that do not perfectly
infer statistical information from the price of offered contracts, price discrimination significantly

4
boosts insurers’ profits.

4
 ‘consent’ may be incomplete. This could result in consumers over-sharing, or under-
valuing their own data property rights.
Given the network effects underlying competition, the competitive playing field
may be levelled more effectively by placing well-designed limits on the use of data.
Introducing some additional rules regarding privacy – while at the same time
allowing selectively for the sharing of some types of data – could increase effective
competition, because the addition of such limitations on the use of data could curb
big techs’ exploitation of network effects. Just as governments have decided that
certain information – such as gender or pre-existing medical conditions – cannot be
used in the pricing of health insurance contracts, there may be a need for
legislatures to define types of personal data that cannot be used in credit decisions
or other types of financial services. Moreover, policies may be needed to give
individuals greater control over their data. Jones and Tonetti (2020) show in a
theoretical model that endowing individuals with data rights may lead to outcomes
that are close to optimal. Design of such policies must also recognize that
mandating data sharing could reduce incentives to produce high quality data, or to
provide valuable services that depend on the provider monetizing data (World Bank
2021).
Data mobility requirements may also help to mitigate trade-offs. Work on
consumer data rights in the EU, Australia and Canada are examples of steps in this
direction. Proposals on data mobility are also included in the EU DMA and DSA
proposals and the UK Digital Competition Expert Panel (2019). Making data
“portable” across providers and increasing data subjects’ control of their data could
help to allow the efficient uses of data while preventing data monopolies from
arising. While some types of data (eg genome data, social contacts) may still be
judged “off-limits” for use in financial services, the ability to port permissible data
between different providers may allow for efficiency and competition in a way that
empowers and benefits consumers, reducing rents by providers.
Beyond data privacy, new entrants are often shifting credit, market, and
operational risks to consumers with limited validation of informed consent or
product appropriateness. Marketplace finance platforms were initially designed as a
regulatory arbitrage, gathering funds through a non-deposit mechanism to avoid
being termed a bank. The platforms introduced competition to banks, reducing
costs to borrowers, and offering higher yields to investors. The increased efficiency
of the platforms, stems in part from the reduced regulatory overhead, minimal
required capital, and absence of deposit insurance premiums (some P2P platforms
do set aside funds to self-insure). Investors in many marketplaces have looked at
P2P as an alternative to bank savings products, but in most cases they bear direct
credit risk without any type of deposit insurance, and often are not aware of that
risk. Robo- advisors and gamified trading platforms expose consumers to market
risks from new instruments with which they may have limited familiarity.
Cryptocurrencies have been subject to volatility risk that might not be anticipated
by those seeking to use them for payments. DLT-based payments embed
operational risks with respect to the exchanges, and even the underlying consensus
mechanisms, that are not always well understood. While Stablecoins seek to address
volatility against a particular basket of currencies, they fundamentally alter the
protection for the holder by removing the guarantee of redemption at par that e-
money issuers and banks provide their customers. Increased usage of digital means
of communication between providers and customers might expose the customers
to clever social engineering attacks. The increased speed of processing of
transactions affords little or no time to rectify any mistakes by customers.

5
 Regulators need to balance the innovation and efficiency introduced by
broader competition with the potential dampening effects of consumer protection
oversight and enforcement. Much depends, as with data privacy, on whether
consent can be adequately ‘informed’ and whether there are reliable objective tests
for product suitability. The challenge of protecting the consumer will become more
acute as the consumer takes on more direct roles in assembling, and even
providing, financial services (eg unintermediated P2P). Depending on how the
regulatory perimeter is set, new approaches to disclosure, determining product
appropriateness, and informed consent need to be developed to ensure the
promise of increased efficiency does not come at the cost of customer losses and
reduced trust.

Privacy and consumer protection vs financial stability and integrity

Data sharing can alleviate problems of asymmetric information and thus could be
beneficial for financial stability and integrity. This potentially introduces a new
trade- off between privacy (and consumer protection more generally) on the one
hand and financial stability and market integrity on the other. This trade-off is
represented by the green segment in Graph 12.
In the credit market, there is ample evidence that more data can improve
stability. Credit registries (CRs), for instance, can counter adverse selection and
reduce information asymmetry between lenders and borrowers. CRs allow loans to
be extended to safe borrowers who had previously been priced out of the market,
resulting in higher aggregate lending (Pagano and Jappelli 1993) and furthering
financial inclusion. They can also counter moral hazard by increasing borrowers’ cost
of default, thus increasing debt repayment (Padilla and Pagano 2000). 17 Conversely,
sharing of credit-related information has the benefit of reducing the information
monopoly a lender has on its borrowers. For example, banks with long-standing
relationships with their borrowers have exclusive information that allows them to
charge higher interest rates and extract other rents from those high-quality
borrowers (Padilla and Pagano 1997). CRs can counter this. Finally, they can reduce
over- indebtedness by revealing borrowers’ debt exposure to all participating
lenders, leading informed borrowers to extent them less credit (Bennardo, Pagano
and Piccolo 2015). Brown and Zehnder (2007) show that an information sharing
institution helps lenders avoid serious losses from short term borrowers.
As the experience with credit registries shows, this is not without its downsides
for consumers. In the United States, for instance, consumer advocates have criticized
systems that screen for past overdraft fees and allow banks to deny account services
to individuals (Baradaran 2015). Transparency in how personal data are used, and
recourse in the event of misuse, provide some mitigation for these potential
exclusionary effects. A large body of regulations has allowed borrowers to gain
insights into their personal credit score and the information collected about them.
In a world of digital innovation, in which a much wider range of data is available for
screening, there may be an even higher premium on giving individuals such access
and rights, commensurate with the growing impact of providers’ use of individuals’
data on availability of services.

17
Doblas-Madrid and Minetti (2013) find that if lenders enter a credit information sharing institution,
their borrowers improve their repayment performance – delinquent payments on leases and loans
decrease.

5
 For example, there are concerns that opaque algorithmic processing of
personal data in banking could result in biases and discrimination. AI/ML-based
credit scoring is supposed to reduce the incidence of non-performing loans, thereby
contributing to stability. However, sophisticated ML algorithms may not be as
neutral as their mathematical nature suggests at first glance. While less human
intervention means that taste-based discrimination may decline, algorithms
developed by humans or trained on historical data can introduce new biases (Morse
and Pence 2020). Using data on US mortgages, Fuster et al (2019) find that Black
and Hispanic borrowers are disproportionately less likely to gain from the
introduction of machine learning in credit scoring models. This suggests that the
algorithm may develop differential effects across groups and increase inequality
(Graph 13, left-hand panel). Borrowers to the left of the solid vertical line represent
“winners,” who are classified as less risky by the more sophisticated algorithm than
by the traditional model. Based on the cumulative share, about 65% of White Non-
Hispanic and Asian borrowers would win, compared with about 50% of Black and
Hispanic borrowers. In another study, fintech does not seem able to close the
gender gap in access to financial services, as women have a lower propensity to
share data than men. Chen et al (2021) find that men are more likely to use fintech
product services than women (27% vs 19%) and that this pattern holds in every
country (except for Peru and India; Graph 13, right-hand panel). Moreover, women
report being more worried about their security when dealing with companies online,
and less willing to share data for better offers on financial services.
Reducing illicit activity is another area where use of personal data can
potentially improve stability and market integrity, but carry costs in terms of privacy
and financial exclusion. For example, anti-money laundering (AML) and combating
the financing of terrorism (CFT) practices require the exchange of detailed
information on all parties involved in a financial transaction. This could result in
exclusion of individuals who cannot present sufficient documentation or have not
had prior access to demonstrate positive patterns that the screening mechanisms
may seek. The data analysis may reveal habits and preferences that are not related
to illicit activity but are nonetheless sensitive or personal. Since sanctions
monitoring, KYC remediation, AML measures, and transaction monitoring are
among the most expensive elements of compliance by banks, the motivation to use
advanced data analytics is strong. Increased automation and use data generated at
low marginal cost could drive down costs and enable more low income customers
to be accredited. However, a general raising of the bar as data coverage of the more
active and/or high income customers improves will exacerbate the data divide
relative to previously excluded segments. Tiered KYC requirements attempt to
address this as a way of balancing data coverage and market integrity requirements.
Discussions around central bank digital currencies put these issues into
particularly sharp focus. Some central banks, such as the People’s Bank of China, are
considering CBDC architectures in which the central bank will have full access to
transactions data so as to support AML/CFT and prudential supervision. Other
central banks are leaning against having access to such data, and looking into so-
called “intermediated” options where private sector intermediaries keep such data
and the central bank only has a ledger of wholesale transactions (Auer et al 2020b;
Carstens 2021). Given the stark divergence in preferences for privacy-stability
tradeoffs across different jurisdictions, it is possible that final designs may also differ
in this regard.

5
Winners and losers of machine learning credit scoring and the fintech gender gap Graph 13

Distributional effect of machine learning in US The fintech gender gap

mortgages1
Cumulative share of customers seeing rate change

0.8

0.6

0.4

0.2

0.0
–1.0 – 0.0 0.5 1. 1.5
0.5 0
Asian White Hispanic
White non-Hispanic 50% Black
1
The change in the log predicted default probability as lenders move from traditional predictive technology (a logit classifier) to machine
learning technology (a random forest classifier) is reported on the horizontal axis. The cumulative share of borrowers from each racial group
who experience a given level of change is reported on the vertical axis.
Sources: Fuster et al (2019); Chen et al (2021).

Addressing these trade-offs, too, will require a societal discussion on data

sharing. Data protection rules like the EU General Data Protection Regulation
(GDPR) and California Consumer Protection Act (CCPA) have represented important
steps in protecting privacy and limiting data use based on consumer consent. Yet
further debate will be needed to determine under which circumstances privacy rules
may be relaxed to improve stability and integrity, and how consumers can be
protected from algorithmic discrimination. This remains a frontier issue in
policymaking in the financial sector and the digital economy more generally, where
different jurisdictions may continue learning from one another going forward.

5. Conclusions

This paper has shown that digital innovation is bringing about economically
meaningful changes in the production of financial services, with implications for the
industrial structure of finance. Improvements in connectivity and computing can help
to enhance efficiency and competition. In many cases, financial services have seen
an un-bundling of different products and services. At the same time, financial
frictions and forces that drove the need for financial intermediators in the first place
have re- asserted themselves. The financial sector may be tending toward a barbell
outcome in market structure, in which large multi-product institutions exist
alongside more specialized niche institutions.
Regulatory and supervisory policy tools will have to adapt. Existing regulatory
perimeters may not adequately cover emerging providers of financial services, and
new players may pose challenges for day-to-day financial supervision. It is
increasingly challenging to balance competition and stability, both among financial
services providers and across myriad players in reconfigured financial product value

5
 chains. Rules for control over data, and which data are allowable for specific
services, will need to be defined – ideally in a way that gives individuals control,
balances competition and efficiency with privacy and consumer protection, and
enhances financial inclusion. New forms of discrimination and bias will need to be
understood and countered. A new balance will have to be struck across consumer
protection, privacy considerations and stability and integrity in a way that respects
societal preferences in different jurisdictions around the world, and within different
societies. Monetary policy, systemic stability, and consumer protection tools, which
have been left largely beyond the scope of this paper, may also need to adapt.
Challenges include access to lender of last resort programs by new providers,
stabilizing credit markets in which non-bank players become major lenders, and
maintaining the viability of deposit insurance programs where new players become
large.
Authorities need to be intentional with respect to market structure. Prior
notions around the trade-offs between competition and stability need to be re-
examined. Longstanding precepts on the separation of banking and commerce are
already being upended. Existing approaches to competition policy and antitrust may
not be well- suited to digital financial services, where measures of pricing and
concentration for one market or firm may have become less informative. The nature
of contestability must be understood anew in markets driven by increasing returns
to scale but also increasingly low entry barriers for niche services. The desirability of
a barbell market configuration will depend in part on consumer behaviours and
skills that determine switching costs in a given market. Market forces are already
being shaped by minimum capital, licensing requirements, activity restrictions and
other policies. Pretending that market forces are just running their course is
inaccurate at best, and at worst could lead to negative outcomes for consumers and
macro stability.
Several authorities are taking promising policy measures to mitigate trade-offs.
Recent entity-based measures around big techs, such as the proposals in the EU,
UK, US and China can take a more ex ante approach to address risks, prevent
anticompetitive behaviours, and in the process promote a level playing field. Data
privacy protections and data mobility requirements can support consumer
protection and may also promote competition. Still, a societal debate will be needed
on what types of data should be permissible for use in financial services and when
privacy rules can be relaxed in the service of stability.
To tread into this new territory effectively, and to balance the necessary policy
goals, authorities will need to collaborate. At the domestic level, central banks and
other financial sector regulators need to cooperate with industry regulators (eg for
telecoms and other utilities that may offer financial services) as well as competition
and data protection authorities. At the international level, authorities are working
together to share expertise on financial regulation of both new entrants and
incumbents, and to coordinate policies. This becomes ever more important given
the cross-border spillovers of anti-trust and data governance decisions, and the
potential to improve service efficiency through harmonization of standards in areas
such as cybercrime prevention, data protection, and interoperability (World Bank
2021). Collaboration can help to ensure regulatory consistency and peer learning
between countries, and ultimately higher welfare for their populations.

5
References

Acemoglu, D, A Makhdoumi, A Malekian and A Ozdaglar 2021 “Too Much Data:

Prices and Inefficiencies in Data Markets” American Economic Journal:
Microeconomics, forthcoming
Acquisti, A, C Taylor and L Wagman 2016 “The economics of privacy” Journal of
Economic Literature, 54(2)
Aghion, P, P Bolton 1997 “A theory of trickle-down growth and development”
Review of Economic Studies, 64(2), 151–72
Alexopoulos, S 2020 “US Mid- and Small-Cap Banks Technology Disruption Report:
With a New Breed of Regional Banks Emerging, Many Are Positioned as Endgame
Winners in the Digital Age” JPMorgan, December 15
Aldasoro, I, J Frost, L Gambacorta and D Whyte 2021 “Covid-19 and cyber risk in the
financial sector” BIS Bulletin 37
Alfonso, V, C Boar, J Frost, L Gambacorta and J Liu 2021 “E-commerce in the
pandemic and beyond” BIS Bulletin 36
Alonso Gispert, T, P Chatain, K Driessen and A Plaitakis 2021 “Regulation and
Supervision of Fintech” Technical Note, World Bank Group Fintech and the Future of
Finance report
Amidu, M and S Wolfe 2013 “Does bank competition and diversification lead to
greater stability? Evidence from emerging markets” Review of Development Finance,
3(3), 152–66
Armantier, O, S Doerr, A Fuster, J Frost and K Shue 2021 “Whom do consumers trust
with their data? US survey evidence” BIS Bulletin 42
Arner, D, J Barberis and R Buckley 2017 “FinTech and RegTech in a Nutshell, and the
Future in a Sandbox” CFA Institute Research Foundation Briefs, 3(4), July
Arner, D, R Auer and J Frost 2020 “Stablecoins: risks, potential and regulation” BIS
Working Paper 905
Arrow, K and G Debreu 1954 “Existence of an equilibrium for a competitive
economy” Econometrica, 22, 265–90
Auer, R 2019 “Embedded supervision: how to build regulation into blockchain
finance” BIS Working Paper 811
Auer, R, G Cornelli and J Frost 2020a “Covid-19, cash and the future of payments”
BIS Bulletin 3
Auer, R, G Cornelli and J Frost 2020b “Rise of the central bank digital currencies:
drivers, approaches and technologies” BIS Working Paper, 880
Autor, D 2019 “World of the Past, Work of the Future,” AEA Papers and Proceedings,
109, 1–32
Baltensperger, E 1980 “Alternative Approaches to the Theory of the Banking Firm”
Journal of Monetary Economics, 6, 1-37
Bank for International Settlements 2019 “Big tech in finance: opportunities and
risks” BIS Annual Economic Report, June

5
 Bank for International Settlements 2020 “Central banks and payments in the digital
era” BIS Annual Economic Report, June
Bank of Thailand 2019 “Permission for Commercial Banks to Undertake Information
Technology Related Services to Support Digital Banking” (unofficial translation),
available online at
https://www.bot.or.th/Thai/FIPCS/Documents/FPG/2562/EngPDF/25620188.pdf,
August
Baradaran, M 2015 How the Other Half Banks: Exclusion, Exploitation, and the Threat
to Democracy, Cambridge, MA, Harvard University Press
Barrett, A 2018 “The Unbundling and Rebundling of Banks”
dxjournalco/2018/09/the- unbundling-rebundling-of-banks/ September 11, 2018
Bar-Gill, O 2019 “Algorithmic price discrimination: when demand is a function of both
preferences and (Mis)perceptions” University of Chicago Law Review, 86(2), 217-254
BBVA 2020 “The five V’s of big data” https://www.bbva.com/en/five-vs-big-data/
Bennardo, A, M Pagano and S Piccolo 2015 “Multiple-Bank Lending, Creditor Rights
and Information Sharing” Review of Finance, 19(2), 519–70
Bodie, Z and RC Merton 2005 “Design of financial systems: towards a synthesis of
function and structure” Journal of Investment Management, 3(1), 1–23
Boissay, F, T Ehlers, L Gambacorta and HS Shin 2021 “Big tech in finance: on the
nexus between data privacy and competition” in R Rau, R Wardrop and L Zingales
(eds) The handbook of technological finance, Palgrave-Macmillan, forthcoming
Bossone, B and H Natarajan, 2021 “Digital Payment Systems Strategy for the MENA
region,” mimeo
Britton, M 2018 “The Importance of The Barbell Economy In 2019”
https://www.mattbritton.com/blog-posts/2018/12/17/the-importance-of-the-
barbell-economy-in-2019
Broeders, D and J Prenio 2018 “Innovative technology in financial supervision
(suptech) – the experience of early users” FSI Insight 9
Brown, M and C Zehnder 2007 “Credit reporting, Relationship Banking and Loan
Repayment” Journal of Money, Credit and Banking 39(8), 1883–18
Brunnermeier, M K, R Lamba, and C Segura-Rodriguez 2020 “Inverse selection”,
mimeo, Princeton
Carletti, E, S Claessens, A Fatas and X Vives 2020 The Bank Business Model in the
post- Covid-19 World The Future of Banking 2, London: CEPR Press
Carrière-Swallow, Y and V Haksar 2019 “The economics and implications of data: an
integrated perspective” IMF Departmental Papers 19(16)
Carstens, A 2021 “Central bank digital currencies: putting a big idea into practice”
Speech, March
Chen, S, S Doerr, J Frost, L Gambacorta and HS Shin 2021 “The fintech gender gap”
BIS Working Papers, 931
Claessens, S 2009 “Competition in the Financial Sector: Overview of Competition
Policies” The World Bank Research Observer 24(1), 83–118

5
Claessens, S and L Laeven 2005 “Financial dependence, banking sector competition,
and economic growth” Journal of the European Economic Association, 3(1), 179–207
Claessens, S, J Frost, G Turner, and F Zhu 2018 “Fintech credit markets around the
world: size, drivers and policy issues” BIS Quarterly Review, September
Coase, R H 1960 “The problem of social cost” Journal of Law and Economics, 3
Coase, RH 1937 “The Nature of the Firm” in PJ Buckley and J Michie (eds) Firms,
Organizations and Contracts, Oxford: Oxford University Press
Cornelli, G, J Frost, L Gambacorta, R Rau, R Wardrop and T Ziegler 2020a “Fintech
and big tech credit: a new database” BIS Working Papers, 887
Cornelli, G, S Doerr, J Frost and L Gambacorta 2020b “Big techs vs G-SIBs:
differences in market funding” mimeo
Cornelli, G, S Doerr, L Gambacorta, and O Merrouche 2020c “Inside the regulatory
sandbox: effects on fintech funding” BIS Working Paper, 901
Cowley, S 2020 “The Tiny Bank That Got Pandemic Aid to 100,000 Small Businesses”
New York Times, June 23, https://www.nytimes.com/2020/06/23/business/paycheck-
protection-program-cross-river-bank.html
Croxson, K, T Valletti, L Gambacorta and J Frost 2021 “Platform-based business
models and financial inclusion” BIS Paper, forthcoming
D’Silva, D, Z Filková, F Packer and S Tiwari 2019 “The design of digital financial
infrastructure: lessons from India” BIS Papers, 106, December
Danielsson, J and R Macrae 2019 “Systemic consequences of outsourcing to the
cloud” VoxEU, 2 December
Deogun, N 1996 “Back to the Fray: Displaced by Mergers, Some Bankers Launch
Their Own Start-Ups --- They Win Over Customers With Personal Service Popular in
Small Towns --- A Barbell-Shaped Industry” The Wall Street Journal, 4 March 1996
Dewatripoint M and J Tirole 1994 The Prudential Regulation of Banks Cambridge,
MA, MIT Press
Didier, T, E Feyen, R Llovet Montanes and O P Ardic Alper 2021 “Global Patterns of
Fintech Activity and Enabling Factors.” World Bank Group Fintech and the Future of
Finance report.
Doblas-Madrid, A and R Minetti 2013 “Sharing Information in the Credit Market:
Contract-Level Evidence from US Firms” Journal of Financial Economics, 109(1), 198-
223
Ehrentraud, J D Garcia Ocampo and C Quevedo Vega 2020 “Regulating fintech
financing: digital banks and fintech platforms” FSI Insight 27
European Banking Authority 2018 “Report on the impact of fintech on incumbent
credit institutions’ business models” July
European Commission 2020a “Proposal for a Regulation of the Council and the
European Parliament on contestable and fair markets in the digital sector (Digital
Markets Act)” December
European Commission 2020b “Proposal for a Regulation of the Council and the
European Parliament on a single market for digital services (Digital Services Act) and
amending Directive 2000/31/EC” December

5
 EY 2020 “A vision for platform-based banking” February
Farboodi M, R Mihet, T Philippon, and L Veldkamp 2019 “Big data and firm
dynamics” NBER Working Papers, 25515, January
Feldman, RJ 2010 “Size and Regulatory Reform in Finance: Important but Difficult
Questions” Scale Economies in Banking Symposium, Federal Reserve Bank of
Minneapolis September
Feyen, E, RR Heffernan, H Natarajan, M Saal and A Sarkar 2021 “World Bank Group
Global Market Survey: Digital Technology and the Future of Finance” World Bank
Group Fintech and the Future of Finance report
Financial Stability Board (FSB) 2017 “Artificial intelligence and machine learning in
financial services: Market developments and financial stability implications”
November
Financial Stability Board (FSB) 2019a “Decentralised financial technologies: Report
on financial stability, regulatory and governance implications” June
Financial Stability Board (FSB) 2019b “Third-party dependencies in cloud services:
Considerations on financial stability implications” December
Financial Stability Board (FSB) 2020 “BigTech in emerging market and developing
economies” October
Frost J 2020 “The Economic Forces Driving Fintech Adoption Across Countries” In M
King and R Nesbitt (eds), The Technological Revolution in Financial Services, Toronto:
University of Toronto Press
Frost, J, L Gambacorta, Y Huang, HS Shin, and P Zbinden 2019 “BigTech and the
changing structure of financial intermediation” Economic Policy, 34(100), 761–99
Frost, J, L Gambacorta and HS Shin 2021 “From Financial Innovation to Inclusion”
IMF Finance & Development, Spring
Fuster, A, P Goldsmith-Pinkham, T Ramadorai, and A Walther 2019 “The effect of
machine learning on credit markets” VoxEU January
Gambacorta, L, Y Huang, H Qiu, and J Wang 2019 “How do machine learning and
non-traditional data affect credit scoring? New evidence from a Chinese fintech
firm” BIS Working Papers, 834
Gambacorta, L, Y Huang, Z Li, H Qiu, and S Chen 2020 “Data vs collateral” BIS
Working Papers, 881, September
Goldfarb, A and C Tucker 2019 “Digital Economics” Journal of Economic Literature,
57(1), 3-43
Hellmann, T, K Murdock, and J Stiglitz 2000 “Liberalization, Moral Hazard in Banking,
and Prudential Regulation: Are Capital Requirements Enough?” American Economic
Review, 90(1), 147–65
HM Treasury 2019 “Unlocking Digital Competition: Report of the Digital
Competition Expert Panel” March
Hileman, G and M Rauchs 2017 “Global Cryptocurrency Benchmarking Study”
Cambridge Centre for Alternative Finance
Holmström, B and J Tirole 1997 “Financial intermediation, loanable funds and the
real sector” Quarterly Journal of Economics 112(3), 663–91

5
International Association of Insurance Supervisors (IAIS) 2017 “FinTech
Developments in the Insurance Industry” February
International Monetary Fund (IMF) and World Bank Group 2019 “Fintech: The
Experience So Far” June
International Monetary Fund (IMF) and World Bank Group 2018 “The Bali Fintech
Agenda” October
Jappelli, T and M Pagano 2000 “Information sharing in credit markets: A survey”
CSEF Working Paper, 36
Jones, C and C Tonetti 2020 “Nonrivalry and the economics of data” American
Economic Review, 110 (9), 2819–58
Kahn C and W Roberds 2009 “Why pay? An introduction to payments economics”
Journal of Financial Intermediation, 18(1), 1–23
Keeley, M 1990 “Deposit insurance, risk and market power in banking” American
Economic Review, 80(5), 1183–1200
Liu, A 2019 “An Analysis of the PBOC’s New Mobile Payment Regulation,” Cato
Journal, Winter
Lukoianova, T and V Rubin 2014 “Veracity Roadmap: Is Big Data Objective, Truthful
and Credible?” Advances in Classification Research Online, 24(1), 4-15
Luohan Academy Report 2019 “Digital Technology and Inclusive Growth” Shenzhen
McKinsey 2020 “Three actions CEOs can take to get value from cloud computing” July
Marous, J 2021 “Will regional banks survive the future” The Financial Brand,
https://thefinancialbrand.com/111327/regional-bank-competition-digital-fintech-
transformation-trends/
Mester, L 2018 “Community Banking and the Community Reinvestment Act” Federal
Reserve Bank of Cleveland speech, 3 October
Mester, L 2010 “Scale Economies in Banking and Financial Regulatory Reforms”
Scale Economies in Banking Symposium, Federal Reserve Bank of Minneapolis,
September
McMorrow, R 2021 “China tech groups given a month to fix antitrust practices,”
Financial Times, 13 April 2021
Morning Consult 2021 “Most Trusted Brands 2021”, available online at
https://morningconsult.com/most-trusted-brands-2021/.
Morse, A and K Pence 2020 “Technological Innovation and discrimination in
household finance” NBER Working Paper, 26739
OECD 2020 “Digital Disruption in Banking and its Impact on Competition”,
http://www.oecd.org/daf/competition/digital-disruption-in-financial-markets.htm
Office of the Comptroller of the Currency (OCC) 1995 “Tying Restrictions: Guidance
on Tying” OCC Bulletin, 1995-20
Oliver Wyman 2019 “The Race for Consumer Trust” March
Padilla, A and M Pagano 1997 “Endogenous communication among lenders and
entrepreneurial incentives” The Review of Financial Studies, 10(1), 205–36
Padilla, A and M Pagano 2000 “Sharing default information as a borrower discipline
device” European Economic Review, 44(10), 1951–80

5
 Pagano, M and T Jappelli 1993 “Information Sharing in Credit Markets” Journal of
Finance, 48(5), 1693–18
Petralia, K, T Philippon, T Rice, and N Veron 2019 “Banking Disrupted? Financial
Intermediation in an Era of Transformational Technology” Geneva Reports on the
World Economy, 22
Pazarbasioglu, C, A Garcia Mora, M Uttamchandani, H Natarajan, E Feyen, and M
Saal 2020 Digital Financial Services World Bank April
Restoy, F 2021 “Fintech regulation: achieving a level playing field” FSI Occasional
Paper 17, February
Ru, H and A Schoar 2016 “Do Credit Card Companies Screen for Behavioral Biases?”
NBER Working Paper, 22360
Rutherford, D 2013 “Barbell (G1)” In Routledge Dictionary of Economics (3rd ed, p
38) Routledge
Schär, F 2021 “Decentralized Finance: On Blockchain- and Smart Contract-Based
Financial Markets” Federal Reserve Bank of St Louis Review, February
Schwartz, B 2005 “The Paradox of Choice: Why More is Less, New York”, NY: Harper
Perennial
Shevlin, R 2019 “The Real Story Behind The PNC-Venmo Clash” Forbes, 18
December Snow 2020 “How the ‘new normal’ is changing cloud usage and strategy”
16 June
Stulz, R 2019 “FinTech, BigTech, and the Future of Banks” Journal of Applied
Corporate Finance 31, 86–97
UK Digital Competition Expert Panel 2019 “Unlocking digital competition” March
UK Payment System Regulator 2016 “Market review into the ownership and
competitiveness of infrastructure provision” July
US House of Representatives (US House) 2020 “Investigation of competition in
digital markets” Majority staff report and recommendations, House Judiciary
Committee's Subcommittee on Antitrust, Commercial and Administrative Law,
October
White, J, Chapsal, A and Yeater, A 2018 “European Union – Two-Sided Markets,
Platforms and Network Effects,” in Global Competition Review - E-Commerce
Competition Enforcement Guide
Williamson, OE 1971 “The Vertical Integration of Production: Market Failure
Considerations” American Economic Review 61(2), 112–23
World Bank 2021 “World Development Report 2021: Data for Better Lives”

6
Previous volumes in this series
No Title Issue date
BIS Papers No 116 CBDCs beyond borders: results from a
June 2021
survey of central banks
BIS Papers No 115 Multi-CBDC arrangements and the
March 2021
future
of cross-border payments
January 2021
BIS Papers No 114 Ready, steady, go? – Results of the third
BIS survey on central bank digital
currency
December 2020
BIS Papers No 113 Financial market development, monetary
policy and financial stability in emerging
market economies
November 2020
BIS Papers No 112 The dawn of fintech in Latin America:
landscape, prospects and challenges
BIS Papers No 111 Inflation dynamics in Asia and the Pacific March 2020
BIS Papers No 110 Measuring the effectiveness of
February 2020
macroprudential policies using
supervisory bank-level data
BIS Papers No 109 The digital economy and
February 2020
financial innovation
BIS Papers No 108 Stress testing in Latin America: A
February 2020
comparison of approaches and
methodologies
BIS Papers No 107 Impending arrival – a sequel to the
January 2020
survey on central bank digital currency
BIS Papers No 106 The design of digital financial
December 2019
infrastructure: lessons from India
BIS Papers No 105 Foreign exchange reserves in Africa:
October 2019
benefits, costs and political economy
considerations
BIS Papers No 104 Reserve management and FX
October 2019
intervention
BIS Papers No 103 Ten years after the Great Financial
June 2019
Crisis:
what has changed?
April 2019
BIS Papers No 102 Asia-Pacific fixed income markets:
evolving structure, participation and
pricing
January 2019
BIS Papers No 101 Proceeding with caution – a survey on
central bank digital currency
BIS Papers No 100 Globalisation and deglobalisation December 2018

All volumes are available on the BIS website (www.bis.org).