1. Introduction

Distributed Ledger Technology (DLT)-or blockchain, the most common DLT implementation underlying cryptocurrencies such as Bitcoin-has been said to be the most disruptive invention since the internet itself (^{Hiesboeck, 2016}). Since its inception in 2009 (^{Nakamoto, 2009}), interest in DLT has grown to the point that it is now being applied in fields as diverse as data management, diamond identification and transaction verification, energy production and consumption, the internet of things, media and content distribution, and, of course, the financial industry (^Mesropyan, 2016). Over three-quarters of all financial institutions worldwide expect to adopt blockchain as part of a production system or process by 2020 (^{PwC, 2017}).

Some businesses are already making use of the technology. Banco Bilbao Vizcaya Argentaria (BBVA), a multinational Spanish banking group, completed the first real-life implementation of an international money transfer using Ripple’s DLT at the end of 2016. The transaction, running on BBVA infrastructure using real money, successfully completed transfers between Spain and Mexico in a matter of seconds (^{BBVA, 2017}).

Almost every major bank around the globe is testing DLT. According to Business Insider UK (^{Meola, 2017}), there are three major factors behind the push for DLT in the finance industry:

Cost savings and efficiency. Banks are dealing with rising costs for maintaining or replacing their aging infrastructure and ensuring compliance with heavy regulatory burdens. Furthermore, banks must deal with increasing economic instability. To that end, DLT-based solutions could generate cost savings of up to $20 billion per year, according to Banco Santander (^{Belinky, Rennick, & Veitch, 2015}).

Competing with startups. Financial technology companies (FinTechs) are using DLT to offer services (such as remittances and international payments) at reduced costs, with greater speed, and with more user-friendly interfaces than major banks. As a result, banks have started to construct their own DLT-based solutions to better compete with these up-and-comers.

New business models. Banks can use DLT-based systems to circumvent central bodies or legacy infrastructure. Banks could potentially develop these systems to create brand new business models that disrupt the current financial ecosystem.

Central banks, though also interested in the technology, have been more cautious in their embrace of DLT. Some central banks, such as the European Central Bank and the Bank of Japan, have declared DLT not mature enough at this stage of development to power the world’s biggest payment systems (^{Koranyi & Evans, 2017}). Likewise, the Bank of Canada stated in a recent report that for critical financial market infrastructures, such as wholesale payment systems, current versions of DLT may not provide an overall net benefit relative to current centralized systems (^{Chapman, Garratt, Hendry, McCormack, & McMahon, 2017}).

In a somewhat more upbeat tone, the United States’ Federal Reserve System has stated that it will continue to work to understand the implications of new payment technologies and models, including DLT and digital currencies, that can facilitate a safe and efficient U.S. payment system (^{Federal Reserve System, 2017}). Other central banks such as the Royal Bank of Scotland (^{Creer et al., 2016}) and the Monetary Authority of Singapore (^MAS, 2017) have built successful DLT-based proof-of-concept prototypes and are now ready to take the next step in their implementation of DLT. All told, two-thirds of central banks worldwide are directly experimenting with DLT protocols (^{Hileman & Rauchs, 2017}).

This paper will do the following:

Describe the research methodology used.

Provide an overview of DLT, with emphasis in those aspects and capabilities of the technology that are relevant to the financial industry in general, and to central banks in particular.

Survey what central banks are currently doing to harness DLT’s potential by reviewing (a) official pronouncements and reports from the central banks themselves, and (b) the academic literature on the matter.

Describe the current limitations of the technology.

2. Methodology

3. A brief introduction to DLT

A distributed ledger is a consensus of replicated, shared, and synchronized digital data geographically spread across multiple entities, sites, countries, or institutions. A blockchain is a type of distributed ledger comprised of unchangeable, digitally recorded data in packages called blocks. These digitally recorded blocks of data are stored in a linear chain. Each block in the chain contains data (e.g., a Bitcoin transaction), which are cryptographically hashed. The blocks of hashed data draw upon the previous block in the chain, ensuring that the data in the overall blockchain have not been tampered with and remain unchanged (^{Blockchain Technologies, 2016}).

This paper considers DLT to be some combination of components, including peer-to-peer networking, distributed data storage, and cryptography that, among other things, can potentially change the way in which the storage, recordkeeping, and transfer of a digital asset is done. The composition of these combinations is dictated by the particular friction or inefficiency a particular implementation of DLT is designed to solve (^{Mills et al., 2016}).

Some of the key components and features in DLT are the following (^{Mills et al., 2016}):

Peer-to-peer node connections. The nodes in a DLT arrangement, which are the devices running the DLT software that individually maintain the shared database records, are connected to each other in order to share and validate information. This structure enables any entity, such as end-users, financial institutions, or financial market intermediaries (FMIs), with a node to share database management responsibilities directly with each other on a peer-to-peer basis.

Distributed histories and ownership. In a DLT arrangement, information regarding records of ownership and transaction histories can be distributed across the nodes in the network. This distribution is the foundation of the technology, with the ledger of transaction histories and ownership positions shared as one common ledger that participants agree is correct.

Use of cryptography. DLT arrangements use cryptography for several purposes, such as identity verification and data encryption. For example, during asset transfers, a form of cryptography known as public key cryptography usually forms the foundation of the transaction validation process. To transfer an asset, a participant may create a digital signature with its non-shared cryptographic credential called a private key. To confirm that the asset in question belongs to the participant initiating the transaction, other participants of the DLT arrangement with the required permissions to act as validators of transactions can verify authenticity of the ledger entry by decrypting it with a mathematical algorithm and the asset owner’s publicly available public key.

Ownership of an asset. Ownership information with respect to an asset can be stored on a ledger within the DLT arrangement, which maintains the ownership positions of all participants in the system.

The DLT protocol defines the asset transfer process. A DLT protocol is a syntax and set of procedures that define how members of the DLT arrangement interact. For a payment transfer, for example, a DLT protocol may lay out validation checks (for example, verify ownership) and conditionality checks (for example, access to sufficient funds or credit). For a securities, commodities, or derivatives transfer, a DLT protocol could provide the conditions around confirmation, clearing, and settlement.

The settlement process. Settlement in a DLT arrangement involves the updating of the common ledger with the new ownership positions of the relevant counterparties. For a distributed ledger, proposed transactions and subsequent positions are broadcast to nodes that maintain a copy of the ledger and ultimately become accepted as the new version of the ledger. The process of having nodes accept a new version of the ledger is commonly referred to as consensus. Two ways to achieve consensus when validating a new transaction are:

Proof of work. Here individuals with suitable computing processing power called miners compete to win the right to validate blocks of transactions by solving a difficult mathematical puzzle. The first miner who completes a new puzzle broadcasts the block and solution to all the other miners and in return “mines” new bitcoins created with that block. Although the problem miners work on is difficult to solve, it is easy to verify. Once the other nodes have seen and verified a new solution, the new block is added to the chain, the transactions in the block are considered settled and miners begin mining a new set of transactions. The way nodes come to agreement about the new block is the consensus mechanism, and the solved puzzle is the proof of work (PoW). This is the approach used by Bitcoin (^{Nakamoto, 2009}).

Notary node. An alternative way to reach agreement is to have a ‘notary’ node that is trusted by everyone and replaces the PoW function. This is the approach used by the Corda environment, a distributed ledger platform developed by the R3 software company (^{Hearn, 2016}).

Application programming interfaces. Application programming interfaces (APIs) are a set of routines, protocols, and tools for building software applications. An API specifies how software components should interact, and within a DLT arrangement APIs can enable the addition of new features or enhancements not native to the distributed ledger protocol itself. For example, they could communicate directly with the underlying protocol of a distributed ledger to effect transfers and gather information. APIs can also provide user-friendly interfaces that make using the technology easier for a broader set of potential users.

Participants may have different roles or functions. Regardless of whether a DLT arrangement is open or closed, participants may be differentiated by the roles they are permitted to play or functions they are permitted to perform. DLT arrangements in which the participants are allowed to perform all activities are often called open or permissionless. Those that restrict participants’ activities are often referred to as closed or permissioned. Cryptocurrency DLT arrangements such as Bitcoin are examples of permissionless systems. The financial industry, however, is focused mainly on developing permissioned systems.

Smart contracts. Smart contracts are coded programs that are used to automate pre-specified transactional events based on agreed upon contractual terms. Like with traditional contracts, a smart contract depends on participants’ consent to its terms. These agreed-upon smart contracts can be used in conjunction with a distributed ledger to self-execute based on information received in the distributed ledger or from other sources. For example, several companies developing DLT products are exploring the use of smart contracts to model corporate debt issuances. In these simulations, a debt-issuing company specifies the parameters of the contract, such as its par value, tenor, and coupon payment structure. Once assigned to an owner, the smart contract would automatically make the required coupon payments until the bond reaches maturity.

4. Literature review

5. Current limitations of DLT

As central Banks continue to study and experiment with DLT, the technology’s limitations have come to light. These will need to be addressed before the central banks can proceed to implement their operational systems using DLT. The limitations fall into the following broad areas:

Speed. Although the speed of end-to-end processing may be adequate, the speed of transaction settlement within the infrastructure itself may be slower than existing centralized systems. For example, DLT arrangements may take longer to achieve settlement when compared with real-time gross settlement systems because the process for validating a transaction and reaching consensus in DLT is potentially more complex than with a central entity (^{BIS, 2017}).

Cost of processing. DLT arrangements can lead to changes in the way costs are allocated among participants. For example, a distributed arrangement in which participants contribute to maintaining and updating a shared ledger allows for the sharing of maintenance across participants. In this sharing of responsibilities, participants operating certain nodes in an arrangement could see increased direct costs for contributing to the operation of the arrangement (^{BIS, 2017}).

Security. Cryptographic tools, such as public key cryptography, play a central role in ensuring the security of existing systems and are of critical importance in DLT arrangements. While current cryptographic tools are considered effective and are widely used today, future technological advancements could render existing cryptographic tools less secure and effective. This issue is of particular concern for an arrangement with a weak governance structure, which may not be able to react quickly enough to emerging security issues and threats. Integration of DLT in existing infrastructures or transition from current systems to DLT-based ones could also generate security breaches that are not inherent in the new technology but could have a strong operational impact (^{BIS, 2017}).

Transparency and privacy. A fundamental requirement for a wholesale payment system is the need for participants to keep their transactions private from parties not involved in the transaction. Proof-of-work systems are ill-suited for these types of large-value systems because they operate under the assumption that all transactions in the system are, at a certain level, publicly observable. In contrast, notary-based DLT systems permit increased privacy because a trusted third party helps validate all transactions. But the lack of transparency in the notary-based system implies that no node in the system, with the possible exception of the notary, has all the information. Therefore, if the information at one or more nodes is corrupted, it may not be possible to reconstruct the entire network since even the notary does not have a full copy of the ledger. This creates the need for backups of individual nodes and a loss of the economies of scale associated with centralized systems. Further, it raises the question of whether the proposed operational-resilience benefits of DLT are possible under the constraint that transactions remain private (^{Chapman et al., 2017}).

Legal settlement finality. Settlement finality is the legally defined moment at which the transfer of an asset or financial instrument, or the discharge of an obligation, is irrevocable and unconditional and not susceptible to being unwound following the bankruptcy or insolvency of a participant. In traditional systems, settlement finality is a clear and well-defined point in time, backed by a strong legal basis. For DLT arrangements, settlement finality may not be as clear. In arrangements that rely on a consensus algorithm to effect settlement finality, there may not necessarily be a single point of settlement finality. Further, the applicable legal framework may not expressly support finality in such cases ^{(Bech & Garratt, 2017}).

Scalability. PCS systems may process hundreds of millions of transactions daily. Consensus algorithms and cryptographic verification introduce latency and limit the number of transfers that some DLT arrangements can process concurrently. Additionally, ledgers that add transactional histories on top of one another, such as blockchains, may challenge storage capacity over time (^{Mills et al., 2016}).

Network effects. If broad adoption of DLT is to take place, the industry will need a critical mass of participants for any application of the technology to be successful. Network effects are derived from the fact that each additional user of a network increases the benefit of the network for existing users. This effect can often lead to a problem for early adoption because the net benefits for early adopters may be negative without sufficient participation, leading to a possible lack of adoption (^{Mills et al., 2016}).

6. Conclusions

Echoing the tremendous interest in DLT by the financial industry in general, central banks are starting to research the possibility of adopting some form of DLT, as they revamp their PCS systems and consider the implications of issuing some form of digital currency.

DLT offers a fundamentally different way to conduct and track financial transactions. It is an innovation that challenges the centralized nature of the existing financial systems in central banks. DLT is still in its infancy, however, and central banks are taking a variety of approaches toward its application. Given the technology’s early stage, a number of challenges to development and adoption remain, including issues around speed, cost of processing, security, transparency and privacy, legal settlement finality, scalability and network effects of the technology.

An area of further research is the full range of applications and use cases by central banks beyond payment clearance and settlement, and the introduction of a complementary or parallel centralized digital currency. Another area deserving further research is the structural changes within the functioning of central banks and the financial industry in general that DLT’s decentralized nature may drive.