Charlie Shin is well-placed to assess the potential value of DLT to CSDs. A computer science graduate who has worked in an AI laboratory, he has participated in a series of J2EE projects since he joined KSD in 1999 and is now working on AI and DLT platforms as a head of the innovative technology team at the Korean CSD. Though he reckons DLT can help in some areas, such as KYC and reconciliations, he told Dominic Hobson he thinks the immaturity of DLT and its lack of standardisation and a widely applicable technical specification argues against its adoption by CSDs in their core systems.
Hobson: What opportunities do you see for central securities depositories (CSDs) in distributed ledger technology (DLT)?
Shin: DLT is potentially useful for optimisation of the collection and distribution of static data, such as Know Your Customer (KYC) information. Another possible application of DLT might be to introduce a higher level of automation in reconciliation. The former utilises the tamper-proof feature of DLT-based information management, while the latter enhances transaction network efficiency.
Hobson: Do you mean that, if all the information about a transaction can be found by every party on a single ledger, there is no need to reconcile multiple ledgers?
Shin: Theoretically, yes. More precisely, reconciliation will be achieved, in much neater and more advanced way, by the nature of DLT, which ensures the integrity of ledgers shared among multiple parties. However, there are still more challenges left to be solved. If the huge amounts of information required for reconciliation reside outside the distributed network, and cannot exist in the same network for technical, legal, or any other reason, will it be still more efficient to use DLT in such cases? We might have to consider these sorts of cases carefully before we conclude anything about using DLT.
Hobson: The financial services industry favours closed or “permissioned” DLT networks, which need governing authorities to decide who can belong, and what they are allowed to do. It is often said that CSDs are a natural choice as governors of permissioned DLT networks. Do you agree?
Shin: Many of us think that there should be a governing body for a permissioned network. But a CSD is not the only institution that could fulfil that role. Other institutions, such as exchanges, could do it. But from the point of view of segregation of duties, it is quite right to think that the CSD is well- placed to play that kind of role.
Hobson: What threats does DLT pose to CSDs?
Shin: The fundamental goal of the technology—decentralisation of processes —can be a threat to the existing business model of a “central” securities depository. However, we must acknowledge that decentralising the process does not necessarily guarantee an increase in the efficiency of the entire financial market. Although some argue that CSDs represent a single point of failure, I would say that there are multiple points of failure in DLT-based networks, which could critically endanger the robustness of the market.
Hobson: If every node in a DLT network has a perfect copy of the data, surely that makes it more robust than a single copy of the data at a central depository?
Shin: When I say “multiple points of failure” I mean that distributed nodes can break down at any time. Just sharing the ledger alone cannot guarantee the robustness of the whole network. Distribution means not only the sharing of the information, but also the sharing of control and power, and that means each meaningful node in a distributed network needs to function properly at all times. As you know, the current technology for DLT is not mature enough for us to just adopt the framework immediately as if it was as tried and tested as the J2EE (Java 2 Platform Enterprise Edition) environment.
Hobson: What lessons does the development of J2EE offer DLT enthusiasts?
Shin: Inter-operability is surely one of the critical aspects of DLT. Achieving inter-operability requires standardisation. That is the most important lesson of the development of J2EE. DLT needs to follow the same standardisation process as J2EE. There are so many competing platforms for DLT that the technology needs a technical specification. Then we can freely use it. I still do not think we should replace our core systems with DLT, but we can use it in some parts of our business once the technical specification is settled.
Hobson: Are you saying, in essence, that DLT is not yet mature enough for systemically important CSDs to adopt?
Shin: I believe DLT will someday evolve into something like J2EE, with standardised functional specifications and stable inter-operability features. I am also sure that leading technology companies will succeed in narrowing the performance gap between legacy systems and DLT-based solutions. Big companies are already deploying DLT to their Cloud solutions. But one fact alone will prevent CSDs like us from adopting the technology for the time being. This is the fact that DLT expands the area that needs to be controlled to the entire network.
Hobson: Does that obstacle to adoption apply to permissioned DLT networks as well?
Shin: Most of the current systems are like a cluster of stars, in which the network revolves around a central intermediary. The boundary of the area which needs to be controlled is limited. The central intermediary has the ability to govern the whole network and all financial transactions that pass through it. If we replace this centralised network with a distributed one, then control is shared among all participants equally. Even if it is permissioned and has a standardised form of governance and trust among the network participants, a distributed network still extends the boundary of control. It becomes more difficult to control risks and co-operate with other parties in a DLT network.
Hobson: You mentioned the technology will one day be mature enough for CSDs to adopt. Assuming it is mature enough, what will govern the readiness of a CSD to adopt DLT?
Shin: I think the main driving force will be the market itself. CSDs are not independent organisations; we exist because we interact with our market participants. The power of DLT essentially comes from networking, which means it will continuously expand its domain into areas where business processes are not systematised or where the systems are not efficient enough. In other words, the readiness of CSDs to adopt DLT will be contingent on external factors. By the time the surrounding marketplace is equipped with DLT, CSDs should be ready to interact with market participants via DLT-based services of their own.
Hobson: Do you not think CSDs, because all market participants have to use their services, are in a powerful position to lead the market to DLT rather than follow it?
Shin: What CSDs must provide is cost-efficient, highly stable and market-friendly services. These are what CSDs from all over the world have been providing for some time already. The financial markets in which CSDs operate are highly competitive so, if we are to replace existing systems with new systems, the new systems must lower cost, increase efficiency, and improve availability. DLT systems, like any new system, must prove they warrant the investment. So when I say CSDs cannot be the driving force behind DLT, it means that we are already quite efficient and competitive, and there is no need for us to switch to DLT until the market demands it. We have a duty to support the market and if the surrounding marketplace changes to DLT services, we should provide supportive services to interact with them.
Hobson: The ASX is replacing its ageing CHESS platform with DLT and offering market participants the option to switch to DLT as well or continue to interact with it using existing technologies. What is your view of the ASX decision?
Shin: I have to say that I was very surprised to hear that news. I hope the project will give us valuable insights into DLT and its adoption.
Hobson: You mentioned inter-operability more than once as important to CSDs, and clearly DLT networks will need to inter-operate if they are to fulfil their potential. Are CSDs natural drivers of inter-operability agreements that could encourage adoption of DLT?
Shin: CSDs are not so much natural drivers of adoption of DLT as born co-ordinators of inter- operability. Again, however, CSDs should follow rather than lead the marketplace. Currently, all clearing and settlement information flows through the CSDs and, if the markets transform themselves into a series of DLT-based networks, CSDs can play a central role in co-ordinating the reorganisation of business processes so that they can inter-operate successfully.
Hobson: To summarise, are you saying that existing technologies can do the job of a CSD at least as well and maybe better and more safely than DLT, and that CSDs should not take the risk of adopting DLT?
Shin: For the existing core services, yes. There are two simple reasons for this. First, the technology market is not mature enough. If we are to adopt a new technology, there should be various kinds of related solutions that we can utilise in monitoring, supporting and securing the core technology, but we do not have them right now. Secondly, the financial market infrastructure (FMIs) of the world have been running their systems with confidentiality, integrity and availability for decades. We do not change our systems without a proper reason. However, for the crypto-assets that are designed to run on DLT, I am afraid we might need a different approach. Among the crypto-asset definitions, I guess it is the asset-backed type of tokens that could have a direct impact on our current financial markets, if they are properly regulated. For such DLT-born assets, we, as FMIs, need to be ready at some point.
Hobson: What advice would you give to CSDs considering whether to adopt DLT, either because their existing platform is ageing, or because they want to offer new services?
Shin: As CSDs, we should focus on ways of maintaining a high level of efficiency and availability and a low level of cost in providing the infrastructure that underpins the financial markets. In that sense, listing our common problems in terms of legacy systems, and identifying ways to mitigate those problems, is a good starting point for a discussion about DLT between CSDs. Furthermore, it would be even more helpful if we could identify the ideal target domains for the adoption of innovative technologies.
Wednesday 10 April 2019
Blockchain: is it an opportunity or threat and what should CSDs do to embrace the future business
Moderator: Vipin Y S Mahabirsingh
Andre Eduardo Demarco, Director of depository operations, B3, Brazil
Lina Hediah, Executive director, Consensys, Dubai
Woochol Shin, Head of innovative technology team, KSD, Korea
Alexander Chekanov, Chief architect, NSD, Russia