Potential netting benefits from expanded central clearing in Canada’s fixed-income market

Introduction

A weekend with friends can end with a lot of small IOUs. Maybe someone buys lunch for the friend who forgot their wallet, while another purchases tickets to a show so everyone sits together. Sorting through all these IOUs can be tedious, especially when many cancel each other out. What if one person collected all the IOUs to simplify everyone’s payment into a single, small amount? A central counterparty (CCP) does something similar for large financial institutions with billions of dollars of payment obligations and associated risk exposures.

Since March 2020, CCPs for fixed-income securities have received renewed attention after investors sold large volumes of fixed-income securities to meet cash needs at the onset of the COVID‑19 pandemic. Normally, dealers might absorb these sales, but in this instance they did not have the risk-taking capacity to absorb sales of this size (Fontaine et al. 2021). In many countries, including Canada, central banks stepped in to restore market conditions. The Bank of Canada did this through asset-purchase programs to meet demand for selling and extended emergency funding through sale and repurchase agreements, also known as repos.1 Academics and policy-makers have suggested that more use of CCPs may make fixed-income markets more resilient and reduce the need for central banks to intervene (Group of Thirty Working Group on Treasury Market Liquidity 2021; Liang and Parkinson 2020).

We consider whether more central clearing would enhance the resilience of Canadian fixed-income markets. Central clearing can help dealers reduce their risk exposures, freeing up their capacity to absorb large asset sales when shocks occur. A CCP reduces risk exposure by becoming the buyer to every seller and the seller to every buyer for the transactions it clears. This lets market participants replace risk exposures from multiple counterparties with a single, smaller risk exposure to the CCP through a process called multilateral netting.2 Using data on Canadian fixed-income trades, we estimate the potential for multilateral netting under hypothetical scenarios where central clearing is expanded beyond status quo levels.

Central clearing could be expanded if more market participants join the CCP. All major dealers and four public pension plans are members of Canada’s fixed-income CCP, the Canadian Derivatives Clearing Corporation (CDCC). CDCC’s eligibility criteria would allow entities such as provinces, federal and provincial Crown corporations, public pension funds and the Bank of Canada to join as Limited Clearing Members (LCMs).3 Institutions who become LCMs pay higher per-trade fees because they do not contribute to the CCP’s emergency capital that is used if a member defaults. It is less clear which membership models would let other buy-side market participants join, such as private pension plans, large asset managers, insurance companies and hedge funds. These entities may have higher credit risk or lack the operational capability to participate in CDCC.

Other ways to expand central clearing include expanding eligible securities and the share of trades that current members centrally clear. Currently, eligible Canadian fixed-income securities are Canada Mortgage Bonds and securities issued by the Government of Canada and several Canadian provinces. In our sample, we estimate that CDCC-eligible securities made up around 46% of all repo collateral used by Canadian market participants during the 2020 and 2021 calendar years. Much of the remaining 54% were US Treasury securities used in repos with US-based counterparties who are not eligible to become CDCC members. We limit our study to repos that use CDCC-eligible securities. Almost half of these repos were conducted between CDCC members, of which 74% were centrally cleared.

Overall, our analysis suggests that expanding central clearing to more buy-side institutions could significantly reduce dealers’ risk exposures from repos, though not eliminate them. The most substantial netting benefits for dealers would come from including all buy-side institutions, but this would require new types of CCP membership. Nevertheless, we still find netting benefits in an intermediate scenario where CCP membership is expanded to institutions eligible to become LCMs. A large share of these benefits results from the Bank of Canada being an eligible LCM. One limitation to expanding CCP membership is that netting benefits for eligible buy-side institutions would likely be substantially smaller than for dealers. This suggests the buy side may need additional incentives to join the CCP and generate potential systemic benefits.

Balance sheet netting of repos is a primary benefit of central clearing

Anecdotes from Canadian market participants suggest a primary benefit of central clearing fixed income is balance sheet netting. Balance sheet netting allows banks to calculate repo risk exposures on a net, rather than gross, basis for regulatory and accounting purposes. Specifically, repos can be offset by reverse repos if they have the same maturity date, use the same settlement mechanism and, importantly, are conducted with the same counterparty.4 Balance sheet netting is possible without central clearing, but it is limited to transactions between two counterparties. Central clearing reduces this constraint since transactions with multiple counterparties are replaced with a single counterparty, the CCP.5

Many dealers in Canadian fixed income are owned by banks and therefore subject to regulatory constraints. An important example of a regulatory constraint is the leverage ratio. The Office of the Superintendent of Financial Institutions (OSFI) requires Canadian banks to maintain a minimum ratio of capital relative to risk exposure.6 Balance sheet netting allows banks to reduce their risk measure and improve their leverage ratios. As well, balance sheet netting may affect how banks internally allocate capital across business lines and report exposures on year-end financial statements.

Greater balance sheet netting may also make central banks’ repo operations more effective in a crisis. In periods when market-wide funding conditions are stressed, central bank repos provide emergency funding for banks to use directly or facilitate lending to the buy side.7 If the central bank, banks and buy-side institutions are CCP members, banks can conduct reverse repos with the buy side that offset central bank repos. From the banks’ point of view, the CCP is the sole counterparty to these transactions. Banks can net these transactions without increasing their risk exposures. This can lower the prices banks charge when lending to the buy side, increasing the impact and effectiveness of central bank repos. However, this would have to be weighed against the central bank’s operational costs of centrally clearing and the need for central bank and CCP policies on pricing and collateral haircuts to be compatible.

Methodology

To analyze the effect of expanded central clearing, we estimate current levels of balance sheet netting using data on Canadian fixed-income trades that include information on which repos are centrally cleared. Our measure of balance sheet netting is the total percentage decrease in gross repos outstanding for dealers. We then calculate this measure as if certain groups of bilaterally cleared repos were centrally cleared, assuming no other changes to the counterparties involved or the terms of trade. This methodology follows Fleming and Keane (2021); details can be found in the Appendix.

Specifically, we compare three scenarios of expanded central clearing relative to the status quo:

• current members clearing all their trades with each other
• repos by all counterparties who are currently CDCC members or eligible to become LCMs under CDCC’s criteria
• full central clearing of all repos by all counterparties

The last scenario poses questions about which clearing models would permit all counterparties, but indicates the potential netting benefits under expanded clearing models.

A limitation of our analysis is we assume no other changes in response to changes in central clearing. For example, if additional repos were centrally cleared, market participants might be motivated to adjust the repo maturity dates or use a greater share of eligible securities to take full advantage of netting. However, new participants and current members in the CCP may not use it for all trades. It is therefore difficult to know how our results might be biased. Our results indicate potential netting benefits only under current business practices.

Expanded central clearing may increase dealers’ balance sheet capacity

We estimate dealers may experience significantly higher balance sheet netting if central clearing was expanded. On average, status quo netting of repos collateralized by eligible securities is estimated to be around 34%, compared with 37% if all inter-member trades were centrally cleared (Chart 1). The modest increase for inter-member trades suggests existing members already have a high degree of netting given current CCP membership. In the scenario of full central clearing, netting is around 58% on average, a substantial 24 percentage point increase.

We estimate an increase of 8 percentage points on average, from 34% to 42%, if all eligible LCMs join CDCC and centrally clear all their trades. In the last six months of the sample, the Bank of Canada accounts for over half of this potential increase. This is due to the increasing volumes of securities repo operations (SROs) that make the Bank of Canada’s bond holdings available to market participants as collateral.8

Chart 1: Dealers may experience material balance sheet netting benefits

Source: Bank of Canada, MTRSLast observation: December 2021

Potential balance sheet netting benefits for the buy side are smaller

To analyze how netting may provide pension funds with the incentive to become clearing members, we estimate corresponding netting benefits for those who are eligible to become LCMs but are not currently clearing members. We limit our analysis to these entities since their prospect of becoming clearing members is better defined than for non-eligible entities. We use the same methodology as in the previous section but calculate netting from the perspective of eligible pension funds rather than dealers.

We estimate that if pension funds eligible to become LCMs were to join the CCP, their netting would increase moderately, by an average of 3 percentage points, from 7% to around 10%, for repos collateralized by eligible securities (Chart 2). This is smaller than the benefits to dealers of all eligible LCMs joining (3 compared to 8 percentage points). One reason is that pension funds tend to trade with a smaller number of counterparties: around four per day, compared with around 46 per day for dealers. Pension funds have smaller benefits from central clearing because they already have a high degree of bilateral balance sheet netting given this smaller number of counterparties. As well, the Canadian buy side tends to use longer-term repos, which are more difficult to net due to their varying maturities.

Chart 2: The benefits of balance sheet netting are substantially less for public pension funds

Source: Bank of Canada, MTRSLast observation: December 2021

Expanded central clearing has associated risks and limitations to overcome

We focus on the benefits of expanded central clearing, but the accompanying risks should be acknowledged. One trade-off with central clearing is that while counterparty credit risk may decrease, a CCP depends on its members to provide cash if one member defaults (King et al. 2020). This may place additional strain on surviving CCP members when markets are under stress. As well, a CCP concentrates operational risk, and an operational outage at a CCP could affect the entire financial system.

Our analysis also highlights potential limitations to expanding central clearing to include more buy-side institutions. Although dealers may benefit significantly from multilateral netting with more buy-side membership, it is not clear whether the buy side has enough incentive to join the CCP. Our findings indicate that prospective buy-side members may experience only modest netting benefits. Moreover, they may not gain as much from the netting benefits they receive since they are not subject to the same regulatory and risk constraints. Other benefits are harder to quantify and include improved counterparty risk management and the potential for improved pricing in normal times. The buy side could receive better prices as a result of central clearing if dealers’ costs were lowered and this change was reflected in prices for their clients. It is unclear whether these benefits outweigh the initial and ongoing costs of central clearing, such as the need to provide collateral. Finally, it is unclear which clearing models may accommodate the Canadian buy-side institutions that are not currently eligible to become LCMs. Other jurisdictions are using or developing sponsored clearing models that may allow these entities to become clearing members (Hraska and Pozmanter 2021). Our analysis suggests this is a productive avenue for future work.

Appendix

Data

We use data that span a sample period of January 2020 to December 2021. Our primary source of data is the Market Trade Reporting System (MTRS), which contains transaction-level information on repos and cash trades, including reporting dealer, client, volume, price, security, transaction date, settlement date, maturity date and whether trades were centrally cleared. Since MTRS exempts reporting of trades with the Bank of Canada, we augment our sample with data on the Bank of Canada’s largest repo operations in the sample: SROs, which provide specific bonds as collateral for market participants, and term repos, which were a stable, emergency source of funding during the market turmoil in 2020. These data sources cover the majority of repos conducted among Canadian counterparties. We restrict our analysis to CDCC-eligible securities. Finally, we remove outliers in the data by dropping trades with extremely large par quantities, extremely long repo terms and extremely low prices; we winsorize prices at the 1% and 99% levels.

Methodology

We compute balance sheet netting as the aggregate percentage decrease in gross repos outstanding for dealers:

\(\displaystyle\\[10pt]{Balance\ Sheet\ Netting}_t=\) \(\displaystyle\,\frac{\sum_{i,j}{({Gross\ Exposure}_{i,j,t}-\ {Net\ Exposure}_{i,j,t})}}{\sum_{i,j}{(Gross\ Exposure}_{i,j,t})}\)

\(\displaystyle{Gross\ Exposure}_{i,j,t}\) \(\displaystyle=\,{Repo}_{i,j,t}\)

\(\displaystyle{Net\ Exposure}_{i,j,t}\) \(\displaystyle=\,Max\ ({Repo}_{i,j,t}-{Reverse\ Repo\ }_{i,j,t}\ ,\ 0)\)

for each dealer \(i\), date \(t\), and netting group \(j\), Repo is the dollar value of the dealer’s total repos outstanding, and Reverse Repo is the dollar value of a dealer’s total reverse repos outstanding. Netting groups indexed by \(j\) correspond to the balance sheet netting criteria defined by unique combinations of repo maturity date and counterparty, which is CDCC in the case of centrally cleared repos and a dealer’s client in the case of bilaterally cleared transactions.9 We make the conservative assumption that dealers take full advantage of bilateral netting with all non-CDCC counterparties. All trade quantities used in the paper are market value (price times par quantity).

Endnotes

References

1. Fleming, M. J. and F. M. Keane. 2021. “The Netting Efficiencies of Marketwide Central Clearing.” Federal Reserve Bank of New York Staff Report No. 964.
2. Fontaine, J.-S., C. Garriott, J. Johal, J. Lee and A. Uthemann. 2021. “COVID‑19 Crisis: Lessons Learned for Future Policy Research.” Bank of Canada Staff Discussion Paper No. 2021-2.
3. Group of Thirty Working Group on Treasury Market Liquidity. 2021. “U.S. Treasury Markets: Steps Toward Increased Resilience.” Group of Thirty.
4. Hraska, J. and M. Pozmanter. 2021. “More Clearing, Less Risk: Increasing Centrally Cleared Activity in the U.S. Treasury Cash Market.” DTCC Whitepaper
5. King, T., T. Nesmith, A. Paulson and T. Prono. 2020. “Central Clearing and Systemic Liquidity Risk.” Finance and Economics Discussion Series (FEDS) Working Paper No. 2020-9.
6. Liang, N. and P. Parkinson. 2020. “Enhancing Liquidity of the U.S. Treasury Market Under Stress.” Hutchins Center Working Paper No. 72.