Digital currencies have the potential to substantially re-shape the future of banking and financial intermediation.
Whether the provision of a digital currency is by the public sector (central bank digital currency, CBDC) or a by a
private initiative (narrowly referred to in this paper as a stablecoin), the eventual rollout of such new instruments
is likely to provide a significant boost to the retail use of digital assets. At the same time, financial innovations
may create new risks and vulnerabilities whose implications should always be thoroughly assessed.
This paper
analyses the introduction of digital currencies in the network of financial accounts. We identify key channels
through which the effects of these novel instruments propagate in the network, and we reveal significant direct
and indirect consequences for most parts of the financial system.
The international monetary and regulatory community has initiated work on several fronts to prepare for an
orderly transition to digital currencies (see, e.g., G7, 2019; BIS, 2019; Basel Committee, 2019; FSB, 2019). While
the importance of financial innovation per se is commonly recognised, these reports also highlight new threats
to financial stability and call for a regulatory response. Among the potential risks of a disorderly transition is the
possibility that,
depending on the ultimate role of existing financial intermediaries, the commercial banking
system may experience the intractable loss of its fee-generating payment business, erosion of retail deposit
funding and disintermediation of its core lending functions, with adverse consequences for the efficient
allocation of credit to the economy. Additional risks not considered in this paper are associated with money
laundering and digital dollarization. Careful planning and coordination among all the relevant parties seems
essential to prevent damaging disruptions.
A rapidly growing body of academic literature is devoted to the study of the design and implications of digital
currencies. Theoretical models include Andolfatto (2018), Kim and Kwon (2018)
, Agur et al. (2019), Keister and
Sanches (2019), Brunnermeier and Niepelt (2020) and Fernandez-Villaverde et al. (2020). These authors
investigate, with sometimes conflicting results, the effects of different digital currency designs on bank lending
and banks’ deposit market power, cost of funding and aggregate welfare. On the more conceptual side,
Brunnermeier et al. (2019) discuss the effect of these instruments on models of monetary exchange and currency
competition. Moreover, Adrian and Mancini-Griffoli (2019) propose a conceptual framework to categorise digital
monies, and Bullmann et al. (2019) provide a taxonomy of the various models of private digital currencies. In a
quasi-empirical approach using financial balance sheets, Kumhof and Noone (2018), in turn, study the
introduction of CBDC and derive a set of “core principles” that could prevent runs from retail deposits to CBDC.
Finally, Bindseil (2020) analyses the system-wide impact of both a CBDC and private digital currencies and argues
that a two-tiered remuneration system may be sufficient to mitigate the risk of retail deposit runs to the CBDC.
Our starting point is the introduction of a digital currency in financial accounts. An important challenge in using
balance sheets is that, to date, no official consensus exists on the classification of digital currencies in national
accounts statistics. We provide a critical review of the discussion and propose a statistical allocation for the
digital financial assets that are considered in this paper. More specifically,
we consider a CBDC as a deposit
scheme similar to the existing central bank deposit facilities, but with an extended list of counterparties, including
non-financial agents. We classify stablecoins either as a new deposit instrument, termed “non-MFI deposits”, or
a collective investment scheme where the digital instrument is a UCIT-type investment fund share.
Armed with these definitions, we build on the work in Castrén and Kavonius (2013) and Castrén and Rancan
(2014) and incorporate the new financial assets into the “Macro-Network”, a network of bilateral exposures
among the institutional sectors of the economy. We model the introduction of a digital currency as a deposit
shift out of commercial banks to the digital currency. Then, under the different designs,
we model a set of
reactions from the sectors affected by the deposit shift, focusing on the banking sector and the implications that
its adjustment may have on the other sectors. We find that in the process of balance sheet adjustments, the
heterogeneous portfolios of bonds and loans held by the different sectors mean that the set of assets (securities
DIGITAL CURRENCIES IN FINANCIAL NETWORKS
Page 4 EBA STAFF PAPER SERIES
or loans) that one sector may have to sell is not the same as the set of assets that another sector may be willing
to buy.1
Price adjustments are then required to allow the markets to clear.2
Shock simulations give rise to the following main findings. First, we identify the key channels through which the
impact of the introduction of digital currencies propagates to the main sectors of the economy. We show that
even a relatively limited loss of deposits is sufficient to trigger major adjustments in banking-sector balance
sheets, which, in turn, have implications for other sectors, including the “rest of the world” sector and, thereby,
foreign residents and institutions. When the banking sector adjusts to a funding gap by redeeming loans,
households experience the largest impact. When the banking sector reacts, instead, by selling securities, nonfinancial corporations are most affected. Our framework is flexible and capable of accounting for any securities
portfolio structures and rules that govern the adjustment of the accounts of the various sectors.
Second, by invoking network centrality measures, we observe changes in the relative importance of the
individual nodes of the network (the institutional sectors).
The introduction of a CBDC or stablecoin will cause
the sector issuing the digital currency to become a more central player in the network at the expense of the
banking sector, but the process also has important consequences for third parties, such as the “rest of the world”
sector. By affecting the shape of the macro network, the introduction of a digital currency may also affect the
network’s stability properties. Our findings therefore also support the view that the regulation of digital
currencies should take into account wider effects than just the immediate counterparty exposures. Finally, we
show that because the key properties of financial networks are time-varying, it is not only the design of a digital
currency but also the timing of its launch that matter in terms of the impact on the financial system.
The remainder of this paper proceeds as follows.
First, Section 2 presents the data, after which Section 3
proposes an allocation of the different types of digital currencies into the financial accounts. Then, Section 4
introduces the methodology and the macro-network approach to modelling financial interlinkages, with a formal
model relegated to the Annex. Next, Section 5 includes the simulation exercises to assess the dynamic impact
of the introduction of a digital currency. Section 6 then generalises the results by looking at different shock sizes
and assesses the time varying impact on network structures
