Research Summary - Blockchain Based Settlement for Asset Trading

[tjd233]

TLDR:

• Using qualitative and quantitative approaches, the feasibility of a blockchain-based settlement system is analyzed mathematically and by a PoW benchmark settlement model.
• Compared to current legacy systems, the blockchain-based settlement system can reduce settlement time, remove third party intermediaries from the settlement process, decrease counterparty risk, and lower various costs due to increased efficiency.
• Thin, or infrequent, markets may be prime candidates for a blockchain-based settlement system as they currently rely on expensive intermediaries who could be bypassed using peer-to-peer atomic transaction enabled smart contracts.

Core Research Questions

Can securities be settled on a blockchain, and, if so, what are the benefits relative to traditional settlement systems?

Citation

Jonathan Chiu, Thorsten V Koeppl, Blockchain-Based Settlement for Asset Trading, The Review of Financial Studies, Volume 32, Issue 5, May 2019, Pages 1716–1753, https://doi.org/10.1093/rfs/hhy122

Background

• Current security markets rely on central third-party depositories, called Central Securities Depositories, to hold securities which can then be traded using a Delivery Versus Payment system.
• Brokers, custodians, and payment agencies are all involved in the process of security settlement.
• The number of actors required to complete transactions results in typical settlement cycles to be upwards to T+2, T+3, or even longer.
• Blockchain-based settlement systems have been proposed as a way to transform security settlements using a shared database of security ownership.
• This would be updated without the need for multiple intermediaries or a third party, and settlement risk can be contained using smart contracts and atomic transactions. Theoretically, a blockchain-based settlement system may decrease the settlement time and reduce settlement risk.

Glossary:

• Delivery versus payment (DvP): a settlement in the securities market requiring payment be made before the transfer from seller to buyer takes place.
• Central Securities Depository (CSD): a centralized entity holding financial securities which allows for the securities to transfer via electronic book records, enabling the buying/selling of securities.
• Proof-of-Work (PoW): a blockchain consensus protocol requiring miners to use computational power to identify the correct “hash” for the current mining block. The more miners competing to identify the correct hash the more secure the network is.
• T+n: for security trading, settlement of trades typically takes n business days after the initial trade execution date T. For example, most stock trades take two business days to settle after the trade is executed, which would be denoted as T+2.
• Mining: is the search for the correct solution to a constantly evolving hashing problem, which results in a relatively consistent rate of discovery of new blocks on the blockchain.
• Smart contract: a contract which is executed automatically depending on the criteria of the contract being met. They do not need intermediaries to be successfully executed. In terms of asset settlement DvP can be ensured using the autonomous nature of smart contracts.
• Atomic transaction: when a series of transactions either all occur or none occur. This prevents the possibility of partial transaction settlement.
• Fork: a divergence in blockchain code or state information.

Summary

• The authors provide a brief overview of blockchain technology and its potential to improve current settlement arrangements.

  • Blockchain may reduce information costs as it provides a common public ledger able to be accessed by anyone. Current post-trade settlement arrangements rely on third parties to keep records, and are generally slow, inefficient, and costly compared to the potential of a blockchain-based system.

• An explanation of the modelled trading environment is provided.

  • The values of trade price and trade surplus are mapped through lemmas.

• The authors define how payment and asset holding values are denoted on a blockchain.

• A PoW permissionless blockchain for settlement system design is defined, discussing the importance of transaction fees, the optimal block size, and optimal block time.

• The authors use a PoW permissionless blockchain for settlements model to quantitatively assess the optimal block size, block time, and assess the impacts of changes in default risk and changes in block time.

  • They also use the model to compare the system to current legacy systems, analyzing the feasibility of a blockchain-based system.

• Multiple extensions of the modelled blockchain-based system are discussed including; optimal number of miners, heterogeneous investor preferences, heterogeneous transaction sizes, intermediation by brokers, and a potential of utilizing a permissioned blockchain model.

• The authors provide a brief discussion and conclusion section about their key findings.

Method

• The authors use both a qualitative and quantitative approach.
  • They construct lemmas and propositions to theoretically and qualitatively explain how blockchains can be optimized to handle asset settlements.
  • Using a constructed PoW blockchain for securities settlement using data from the Trade Reporting and Compliance Engine (TRACE) reporting system the authors conduct a quantitative analysis of the feasibility of using blockchain to settle asset settlements.
• The authors construct a hypothetical (PoW) blockchain for securities settlement

  • The chain features three distinct features:

    1. The blockchain allows for a DvP mechanism by handling the transfer of payment and securities. This potentially eliminates settlement risk.
    2. The blockchain is permissionless.
    3. The blockchain is designed to control the speed of settlement by utilizing both block time and block size as tools to moderate the settlement speed.
    • Participants can post transaction fees to select how quickly their transaction is settled.

  • The benchmark chain contains the following parameters:

    1. Trading period = 8 hours
    2. Block time = fixed at 5 minutes
    3. Number of trades = 45,000
    4. Individual trade value = $1 million
    5. Maximum default exposure = 3%
    6. Daily valuation shock = 250 trades
    • Essentially, 250 trades completed that day are viewed poorly by the traders after the trades are executed.
    7. Investors attempting a fork incur a cost of $10,000

  • The model is used to predict multiple different trading environments.

Results

Qualitative results — the authors map out a theoretical PoW blockchain used for asset settlements through models and the defining of multiple lemmas. Key results are shown below:

• Trade price and trade surplus are defined as follows when A=personA and B=personB:

And

• The authors show the price of the trade is independent of time but the surplus is decreased when settlement is delayed. They show investors are willing to pay a transaction fee to shorten the settlement period.

• Mined blocks would contain the following information:

  • The authors define how payment and asset holding values are denoted on a blockchain. The following equation defines these values where B=Block and n=subperiod such that n+1 implies the current block(Bn+1) is being mined on top of block(Bn):

• Transaction fees and block size have an inverse relationship. When block size increases, investors’ willingness to pay for a transaction decreases, and vice-versa.
• The optimal block size and block time is set to minimize lag in settlement time.

Quantitative Results — the constructed PoW blockchain model allowed the researchers to estimate optimal trading environments, transaction fees, etc. Key results are discussed below:

• Benchmark PoW chain results:
  • Optimal block size = 774 transactions
    • Block supports 2.6 transactions/second
  • Settlement time per trade = 148 minutes
  • Average fee per trade = $38 or 0.34 basis points per $1 million transaction
• Results of increasing the number of transactions:
  • Increased trade volume results in larger block sizes due to more transaction fees.
  • Shorter average settlement time and lower cost transaction fees.
  • Implies a permissionless blockchain infrastructure is scalable from an economic view.
• Results when trades become time critical, i.e. when investors place a high value on faster trades:
  • Block size increases and average settlement time decreases as investors are willing to pay higher transaction fees to ensure timely settlement.
  • The cost of transaction fees changes non-monotonically. They increase up to the point where the block size increases. Once the block size increases, costs fall, but immediately begin rising until the block size increases, creating a continuous cycle.
    • The general trend line implies the average cost of transaction fees continues to increase over time.
• Results when default risk exposure increases:
  • More mining activity is required to block potential forking behaviors.
    • This decreases the block size, which increases both the average settlement time and the transaction fee.
• How does manipulating the block time impact the system?
  • If block time is decreased from 5 minutes to 4 minutes the following results occur:
    • The optimal block size decreases, transaction fees investors are willing to pay decrease, and incentive to fork remains the same.
• When comparing the current legacy systems, where settlement time is T+2, investors would require a subsidy of $150-$200 to use said system over the proposed benchmark blockchain-based system.
  • This advantage of the blockchain-based system continuously decreases as settlement time decreases. So, when settlement time on the legacy system is equal to 0 (instantaneously) the legacy system is preferred.
• When investor preferences become more heterogeneous in regard to liquidity shocks, block sizes tend to decrease and those investors who prefer early settlement fall into earlier blocks while paying higher transaction fees compared to those who are less time sensitive.
  • “Impatient investors need to bear a larger fraction of the total mining cost than more patient investors.” (pg 1744)

Discussion

• Blockchain-based settlement systems are shown to have three distinct benefits over traditional security settlement systems:
  1. Time-varying settlements that depend on the needs of the market instead of technological constraints.
  2. No intermediaries are required to ensure delivery/payment. Utilizing smart contracts atomic transactions can be conducted.
  3. Improving thin, or infrequent, markets like private equity or cross-border trading. These markets traditionally rely on financial institutions who typically charge high fees. A permissionless blockchain-based settlement system would allow users to bypass the high fees charged by institutions, allowing for indirect cost savings.
• Using a blockchain-based settlement system relies heavily on the volume of transactions, high costs of tampering (attempted forks for example), and limited default exposure.
• Permissionless blockchains vs. permissioned blockchains may have different pros and cons. A key advantage of a permissioned system is that no computational resources are wasted through mining, and instead transaction fees are transferred from investors to the validator(s).
• Results indicate that for blockchain-based settlement to be effective it needs to accommodate broker service so customers can have access to all market types. This may be counterintuitive, as a blockchain-based system may be able to bypass intermediaries, like brokers, all together.
• A key component of legacy settlement systems is privacy regarding who is completing what trade. Blockchain-based settlement systems may be more transparent in nature than the current legacy systems, reducing the level of privacy traders currently have. It is unclear how this may impact the trading environment.

Applicability

Recent events, such as multiple retail trading platforms temporarily limiting/delisting GME trades (read more on how crypto fits in here), have highlighted some potential shortfalls in the current asset/security trading environment. With the current system relying on third-party intermediaries, like closinghouses, the typical stock trade is completed roughly two days after the initial trade is executed. The results of this research indicate a decentralized blockchain-based security settlement platform may be able to make settlements more efficient by bypassing the centralized clearing party. This could allow for peer-to-peer trading utilizing atomic transactions through smart contracts to ensure DvP settlements without intermediaries. Furthermore, the potential of moving asset settlements to a blockchain system could reduce counterparty risk and in turn reduce various costs.

[jmcgirk]

Seems like there are massive potential savings and efficiencies by using blockchain-based settlements. But are there any unique risks that come from blockchain-based asset trading? Would it be as resilient to runs, and how might automation factor in?

[Larry_Bates]

This sounds like a place where the Chainlink Off-chain Reporting Protocol would be a useful solution that would create an off-chain oracle network while reducing transaction fees. It would seem beneficial to have the settlements queued and reviewed off-chain before putting them on-chain for many reasons, including the capacity to correct errors before committing a transaction on-chain.

[Ishgun_Singh]

The key question here is that is a blockchain based transaction clearance approach better than a conventional one? Can the clearing house i.e., a centralized ledger be replaced by a distributed ledger technology?

Allow me to present my opinion on both sides of the coin (ignoring other issues such as scalability, simply focusing on the economic side of this implementation):

1. No – retaining the legacy systems would be a better option.

• Netting is an extremely important component of settlement. It lowers transaction costs, reduces counter party risk, and increases operational efficiency. A blockchain based approach for settlement eliminates the component of netting form the process.
• From the demand side of things, there may not be sufficient demand for a blockchain based approach in the considered scenario. That approach requires pre-funding for settlements. This may be extremely tough to execute. People are accustomed to a T+1/2/3-day settlement time since decades. Changing this would be hurdle and increasing the slope of the adoption curve against time would a challenge at the minimum. The point here is that if the customer does not want simultaneous execution and settlement, why do it at all?

2. Yes – A Blockchain based settlement is a good idea for asset trading.

• If we have an instantaneous settlement, we most likely don’t even need the netting. It’s implicit.
• For the aspect of people being accustomed to delayed settlements, this may be achievable with smart contracts.
• Smart contracts can also be used to reduce back office costs by automating processes.
• Theoretically, a blockchain-based settlement system may decrease the settlement time and reduce settlement risk.