SWIFT《區塊鏈對證券交易全流程產生的影響及潛力》報告.pdf

SWIFT INSTITUTE SWIFT INSTITUTE WORKING PAPER NO. 2015-007 THE IMPACT AND POTENTIAL OF BLOCKCHAIN ON THE SECURITIES TRANSACTION LIFECYCLE MICHAEL MAINELLI ALISTAIR MILNE PUBLICATION DATE 09 MAY 2016 The views and opinions expressed in this paper are those of the authors. SWIFT and the SWIFT Institute have not made any editorial review of this paper, therefore the views and opinions do not necessarily reflect those of either SWIFT or the SWIFT Institute. 1 The Impact and Potential of Blockchain on the Securities Transaction Lifecycle *Michael Mainelli ?and Alistair Milne ?Abstract This paper reports the outcome of a series of interviews and focus group meetings with professionals working in post-trade processing and the provision of mutual distributed ledger services. The objective was to elicit and document views on three research hypotheses about the potential impact of mutual distributed ledger technology ‘blockchain’ on post-trade processing global securities markets. These hypotheses are a on the appropriate access to mutual distributed ledger; b on whether change would be piecemeal or ‘big bang’; and c on the extent to which applying mutual distributed ledger in securities settlement would require major changes in business processes. Our research finds that while the use of blockchain to validate operational data in mutual distributed ledgers can yield substantial reductions in both cost and risk, the concept of data sharing itself is far from new. Current interest in mutual distributed ledgers has established significant momentum, but there is a danger of building unrealistic expectations of the extent to which the technology on its own will address the underlying need for co-ordination of business processes both within and between firms. Achieving all the potential benefits from mutual distributed ledgers will require board level buy-in to a substantial commitment of time and resource, and active regulatory support for process re, with relatively little short term payoff. 214 words Keywords Open source software, databases, distributed ledgers, fintech, securities settlement, Bitcoin JEL numbers G29, L17, L89*We are grateful for the support of the SWIFT Institute from a research grant awarded in September 2015 and for participation in interview and focus group meetings and comments on earlier drafts from a large number of individuals see Appendix 3 for a listing of their firms or other affiliations. ?Z/Yen Group, michael_ ?School of Business and Economics, Loughborough University, a.k.l.milnelboro.ac.uk 2 Table of Contents Abstract . 1 1. Introduction . 3 2. Background . 7 3. Hypotheses and Research s . 11 Research hypotheses . 11 Research s 14 4. Findings 15 Two general insights . 15 The continuing need for trusted third parties . 16 Distributed ledgers as a database technology 18 Detailed research findings. 22 Further applications including smart contracts 29 5. The Emperor’s Old Clothes . 31 6. Summary and Conclusions . 35 References . 39 Appendix 1. Public Domain Materials 44 Our own previous work. 44 Terminological disputes what is a distributed ledger 45 The boom in FinTech and blockchain. 48 A digression on the efficiency of the Bitcoin ‘proof of work’. 56 Why is there considered to be so much potential value from the application of distributed ledgers in financial services . 61 More cautious perspectives and the case for co-ordinated and collaborative change 64 Appendix 2. Record of Focus Group Meeting of 25 thNov, 2015 . 69 Initial discussion 69 Discussion of hypotheses . 70 Additional discussion of smart contracts . 78 Appendix 3. List of Inants’ Affiliations . 80 3 1. Introduction Mutual distributed ledgers aka blockchains record transactions and ownership using pervasive, persistent, and permanent data structures replicated across numerous computers. 1The two principal technology components are public-key cryptography and ‘peer-to-peer’ or shared data storage. The end result is a data source that is simultaneously logically ‘central’ while technically ‘distributed’ across the computers on the network. The network of computers using the ledger can consult a single authoritative and immutable ledger of all the data transactions from the origin ‘genesis’ of the data structure. Everyone has the same ‘view’ of the same ‘data’, though they may be retrieving the data from different physical sources. Nowadays the best known mutual distributed ledger is the Bitcoin ‘blockchain’. This provides an indelible record of all of the cryptocurrency’s transactions from when they first began on 3 January 2009, using a consensus process, known as ‘proof of work’ Bitcoin mining to ensuring that the thousands of active nodes on the Bitcoin network do indeed all have the same view of the underlying data to a very high level of cryptographic security. The term blockchain comes from the cryptographic validation of Bitcoin transactions in blocks that combine several transactions together the size of these blocks varies a great deal but can contain a thousand transactions or more. Once a block is validated by proof of work then the transactions it contains are permanently and irreversibly recorded in the blockchain across numerous machines with no single, or ‘master’, central record. Participants in the Bitcoin network can view the ledger transactions in the blockchain from thousands of different machines but always see the same history. Mutual distributed ledgers are not new but, historically, they have suffered from two perceived disadvantages insecurity and complexity. The robustness and relative 1In this paper we employ the following definitions. A ledger is a record of transactions; distributed means divided among several or many, in multiple locations; mutual is shared in common, or owned by a community; a mutual distributed ledger MDL is a record of transactions shared in common and stored in multiple locations; and mutual distributed ledger technology is a technology that provides an immutable record of transactions shared in common and stored in multiple locations. See Appendix 1 and also Mainelli, 2016 for discussion of the terminology. 4 simplicity of the Bitcoin blockchain has drawn favourable comparisons with complex messaging and processes and the multiple requirements for storage used in most financial services, especially in wholesale financial markets’ securities settlement messaging and processes. This, along with the substantial current regulatory and other cost pressures on capital markets firms, have focused many in the industry on the possibility of achieving substantial efficiency gains by applying mutual distributed ledgers to securities settlement. It is though as yet unclear clear to what extent the robustness and simplicity of the Bitcoin blockchain can be replicated when applied to the varied and substantial throughput of global financial market transactions, nor which of the many s of mutual distributed ledger are best suited for post-trade processing and to what extent they should be applied. Despite these uncertainties, strong claims are now being made about the potential of mutual distributed ledgers to reduce costs and risks. A number of initiatives applying mutual distributed ledger to securities settlement are now being pursued, attracting substantial investment from both major banks and venture capital funds. These employ a range of approaches. Some directly employ the Bitcoin blockchain, or a simulacrum thereof. Others employ other open source mutual distributed ledger software such as that offered by Ripple or Ethereum. Others are developing private, proprietary, or semi-public ledgers. The common feature – as highlighted by our preferred terminology ‘mutual distributed ledger’ and ‘mutual distributed ledger technology’ – is providing a common view of ownership and other data across many computers with no centralised record. Understanding of the technology however lags well behind the hype, amongst practitioners, policy makers and industry commentators alike. ‘Blockchain’ technology seems to promise major change for capital markets and other financial services – some say it may ultimately prove to be as important an innovation as the internet itself – but few can say exactly how or why. The objective of this paper is to offer some insight into the potential use of mutual distributed ledger for reducing costs and risks in securities settlement. The findings are based on engagement with practitioners in a series of structured interviews and two focus group meetings supplemented by a review of extensive relevant public domain materials, much of which is described in an appendix. 5 Our principal contribution is investigating three ‘research hypotheses’, not hypotheses that can be confirmed or rejected in a statistical sense but carefully worded statements designed to elicit the fullest possible understanding from our inants of the practical detail involved in the application of blockchain to post- trade processing. Investigating these hypotheses through extensive engagement with a large number of practitioners has highlighted the many challenges that will need to be addressed if mutual distributed ledgers are to be fully adopted in securities settlement. To provide a flavour of our findings, here are a couple of the conclusions emerging from our analysis. A view expressed in several of our interviews is that a root cause of inefficiency in post-trade processes is that the two sides to a trade each maintain their own separate records of the transaction and the resulting counterparty obligations. The consequence is the expenditure of much unnecessary resource reconciling this data with that held by the counterparty at each step of contract cution. This is one reason why mutual distributed ledgers have such appeal to established firms – holding out the possibility of storing all relevant operational ination about trade cutions in agreed and verified at shared by all participants to the trade. It is important to recognise however that agreement on the sharing of trade ination can be bilateral and therefore requires neither that the actual ownership of securities is recorded on a mutual distributed ledger nor any of the various consensus mechanisms employed to ensure consistency of data on a mutual distributed ledger. 2Indeed from a technological perspective the fuss about blockchain is a little surprising since distributed databases have themselves been around for many years and in the securities market context the cryptographic advances incorporated in the Bitcoin blockchain do not add any obvious value over the established approaches to ensuring security widely used in financial services. 2Such a bilateral approach to sharing of data for improving the efficiency of post-trade efficiency seems to be the distinguishing feature of the ‘Corda’ software for financial services, announced in early April 2016 by the industry digital ledger consortium R3, but the possibility of such bilateral data sharing is far from new. It is also used for example in the ‘woven broadcasting’ time stamping and document retri systems of Z/Yen see for the States of Alderney as an open public service example. 6 A related finding is that achieving all of the potential benefits of mutual distributed ledger is likely to be afflicted by the problem of ‘excess inertia’, a problem highlighted in other non-financial contexts by the research literature on technological innovation. This is a market failure that arises when users are ‘locked in’ to existing practice and as a result private profit incentives are not strong enough for individual firms to adopt a more efficient standard or technology, even when adoption would reduce costs for the industry as a whole. In the case of mutual distributed ledgers, this problem of private incentives is reinforced because the full benefits of using mutual distributed ledger appears to involve substantial although as yet unquantified costs in the short to medium term while the anticipated benefits lie largely in the relatively distant future. Mutual distributed ledger may struggle to compete for funding in an environment of limited IT budgets much of which is already committed to regulation and compliance. Our overall judgements are cautious compared to much that has been written about the application of mutual distributed ledgers in financial services. 3While agreeing that mutual distributed ledgers have considerable potential for both cost and risk reduction, a great deal remains to be done in order for this potential to be exploited. The momentum behind a shift to recording data in mutual distributed ledgers does however appear to provide the industry with an opportunity to harmonise business processes and address many long-neglected inefficiencies in post-trade and other financial operations. The paper is organised as follows - Section 2 sketches the background to our work. It considers the challenge of cost reduction in post-trade processing, explains how mutual distributed ledgers such as the Bitcoin blockchain differ from traditional centralised recording of data and why there is now such a high level of interest in their application to securities settlement. 3Two other recent reports also emphasise the practical challenges of employing distributed ledgers in capital market operations are DTCC, 2016; Euroclear and second the perception of current the “FinTech” boom, mplified by the cryptocurrency Bitcoin, as an opportunity for fundamental entrepreneurial-driven change in financial services Firms participating in financial markets today devote enormous amounts of resource to obtaining data, checking data records, and reconciling that data both internally and against the records of other firms. Broadridge, utilising data from Oliver Wyman