We live in volatile times. Last week a Chinese steel producer Tsingshan Holding Group was reported to face an $8 billion paper loss because of skyrocketing nickel prices.
Periods like this highlight the importance of central counterparties (CCP) that sit between buyers and sellers in financial markets to provide a buffer in case of default. While only three CCPs have ever failed, all three failures occurred in circumstances similar to the current market climate.
Hence it’s an opportune time to explore the fragility of CCPs. And whether or not there’s a better alternative.
In the past decade, the volume of transactions cleared via CCPs has risen substantially. That’s largely because the Great Financial Crisis triggered a G20 agreement to migrate over the counter (OTC) derivatives trading to be centrally cleared. In terms of scale, the notional value of OTC derivatives is more than five times the market capitalization of global stock markets.
As CCPs grow bigger, so do the risks they bear. And some CCPs are most certainly too big to fail. In the meantime, technology has evolved.
A new paper explores whether a CCP could be replaced by a so-called distributed Financial Market Infrastructure (dFMI) using blockchain. Its nine authors include executives from JP Morgan, Fnality and Clearmatics. Fnality is the payments infrastructure, formerly known as the Utility Settlement Coin, backed by 15 major financial institutions, which plans to launch a UK synthetic wholesale CBDC, to be followed by other jurisdictions. Clearmatics came up with the concept behind Fnality.
What is a dFMI?
Unlike the CCP model with a central party sitting in the middle of every trade, with a dFMI, all transactions are executed peer to peer using a blockchain network governed by its participants. The market participants operate the network nodes, and the regulator can have permission to see all data, providing a view over an entire market.
CCPs exist to deal with risks, but many of those risks disappear with automation.
For example, delays in calculating and settling variation margins in volatile markets increase risk. dFMIs can calculate margin requirements far more regularly and settle them instantly using tokenized cash or other tokenized assets. Fnality is an example of a settlement token, and HQLAX tokenizes collateral.
In the case of one of the three CCP failures by France’s CLAM in 1974, the lack of margin adjustment was a major factor in its demise.
Short settlement times and atomic transactions are a feature of blockchain in capital markets, such as delivery versus payment (DvP) for assets, or payment versus payment (PvP) for foreign exchange. While shorter settlement times most definitely reduce risk, so does the netting done by current CCPs. And dFMIs can potentially implement multilateral netting as well.
Numerous advantages and possible limitations of a dFMI are explored in the paper, with just a couple analyzed here.
There are examples where CCPs act in their own interest, which can result in their clearing members bearing additional losses. CCPs have multiple buffers to deal with defaults by a counterparty, starting with the collateral or margin posted by a clearing member. The CCP puts up its own capital as a buffer and clearing members contribute to a default fund. However, the amount that clearing members chip in is often far higher than the CCP and there are multiple examples where incentives may not be fully aligned.
In late 2018, Nasdaq Clearing AB in Sweden had a close call. A single trader got on the wrong side of a trade, resulting in losses of €114 million in excess of its collateral. Nasdaq commodity capital contributed €7 million and the members’ default fund had to draw down €107 million. And insufficient variation margin was also a major factor here.
“Aas (the trader) was not required to pay any additional margin, even though the position made up a large proportion of the Nordic power market,” noted the BIS. “The reasons for this are unclear, but some observers have suggested that margin-setting may sometimes reflect competitive pressures.”
And others might be more direct and call that incentive misalignment.
With a dFMI, the loss is borne by the parties directly involved in a transaction. A buffer fund is still needed, including for issues like cyberisks, but only the members would contribute according to codified rules. If the contributions are in proportion to the risks they contribute, this aligns incentives.
There are a reasonable number of issues that dFMIs have to overcome to be considered viable. Scalability issues with blockchain networks are well documented with evolving solutions.
Perhaps the biggest issue is the challenge to the status quo. One example mentioned in the paper is regulators might prefer oversight of a single CCP entity rather than a consortium operated network. The document briefly suggests that perhaps the public Ethereum blockchain network could become a dFMI.
We’d additionally argue that perhaps it’s not just a matter of oversight, but also the speed of crisis decision-making by a single entity might be faster.
Another status quo issue is that dFMIs require interoperability, something that clearing houses have tended to resist.
Does blockchain at the edges reduce dFMI advantages?
In the meantime, we’ve seen blockchain deployed at financial institutions, reducing their own risks and hence CCP risks.
For example, numerous initiatives automate margin variation calculations and settlement, such as JP Morgan derivative margin payments and Vanguard and Symbiont for over the counter (OTC) forex derivatives.
There are also moves by incumbent CCPs to adopt blockchain while maintaining conventional CCP governance. An example is the DTCC’s Project Ion, designed for netted T+0 settlement of bilateral transactions for U.S. stocks. However, the trades pass from the DLT platform to the centralized system.
Hence, a point not covered in the paper, is that some of these blockchain deployments have a knock-on effect on the CCP’s risks. So even if it were objectively provable that a dFMI is superior to a CCP, by already reaping some of blockchain’s risk-reducing benefits, the CCPs are defending their turf. So far.
An alternative view is that blockchain-based infrastructures are inevitable and just a matter of time and scale. The building blocks of a blockchain-powered world are starting to fall into place. These include tokenized money (Fnality), tokenized collateral (HQLAx, Deutsche Börse), and regulated security token exchanges (SIX Digital Exchange, SDX). However, there are different degrees of centralization in each case, including governance.
The real question is, what will the network governance of dFMIs look like? Could it be distributed, such as Ethereum? Will it involve institution-owned collaborations such as Fnality? Or might control remain with centralized entities like SIX, the Deutsche Börse, the DTCC, or even a regulated startup?