Introduction

The Canadian Overnight Repo Rate Average (CORRA) is a measure of the average cost of overnight Government of Canada (GoC) general collateral in Canadian dollar repurchase agreements (repos).1 Before September 2023, CORRA was consistently at or slightly below the Bank of Canada’s target for the overnight rate. However, in the autumn of 2023, CORRA started being consistently a few basis points (bps) above the Bank’s target. This was somewhat unexpected given that the Bank has a floor system and excess settlement balances. As a result, in early January 2024, the Bank began conducting overnight repo operations to provide temporary overnight liquidity to primary dealers (PDs) in GoC securities in exchange for non specific GoC bonds or treasury bills (i.e., general collateral).

Overnight repos are routine operations that are part of the Bank’s operational framework to implement monetary policy and reinforce the policy rate. Immediate commentary by market participants questioned whether the Bank’s quantitative tightening (QT) program, which reduces the level of settlement balances, was causing the upward pressure on CORRA and whether the Bank’s reinforcement of its target rate with overnight repos was a sign that QT was coming to an end.

Understanding where pressure in the overnight funding market is coming from is particularly important for the implementation of monetary policy. This paper identifies and discusses drivers of the recent upward pressure on CORRA. Although QT has reduced the level of settlement balances, we find that market positioning is likely the main driver of the recent upward pressure on CORRA, although other structural factors also play a role. To better understand these drivers, we first provide a brief overview of the main repo market participants in Canada and the role of repos in the bond market.

Who uses repos?

In Canada, major users of repos in the GoC fixed-income market are similar to users of other major bond markets.2 We describe each type of major user below.

Hedge funds

Hedge funds are increasingly active in the GoC bond market, and they employ a host of strategies to try to earn incremental returns. Popular strategies include attempting to predict outright yield or curve movements (e.g., 2s10s, 10s30s) based on their views on macroeconomic factors such as economic data, monetary policy, fiscal policy and geopolitical risk.3 Hedge funds can also seek to implement relative value strategies, which include, for example, making bets on relative value switches between off-the-run bonds and current on-the-run bonds. In addition, hedge funds have implemented cash-futures basis trades, a popular and documented hedge fund strategy in the US Treasury market. These trades try to profit from price discrepancies between the bond futures market and the cash market.4 Hedge funds typically use leverage through the repo market to finance any long bond positions (i.e., a leveraged long5) and to source specific collateral to cover any short positions while executing their trading strategies (Figure 1).

Primary dealers in Government of Canada securities

Canada’s 11 PDs are key intermediaries in the repo market. They make markets in GoC securities. As a result, at the end of each day they will hold long positions in a subset of GoC International Securities Identification Numbers (ISINs) while also holding short positions in another subset of ISINs. These positions would generally result from only the PDs’ market making with clients (i.e., the PDs would not incorporate their own market views). However, in practice, PD cash bond desks (cash desks for short) have room to adjust their inventory to reflect their anticipated moves in outright yield levels or shape of the curve (e.g., flatteners, steepeners, butterflies) and expected changes in relative value (e.g., movements in bonds that trade rich/cheap to the GoC curve). Like hedge funds, PD cash desks need financing for any long positions and will source specific bonds to cover any short positions (Figure 2). The funding and sourcing of such positions are mainly done in the repo market and are the responsibility of each PD’s repo desk.

Pension funds

Canada’s large pension funds are sophisticated users of the repo market—they have made great use of leverage, funding their large beta fixed-income portfolios using repos with PDs. In addition, specific fixed-income teams within pension funds often have alpha budgets, where they attempt to earn returns above and beyond any passive beta fixed-income portfolio by implementing strategies like those of hedge funds and using the repo market for funding. Finally, pension funds use repos as a quick way to raise funds to make up any shortage in daily cash balances or as a means of bridge financing to help fund major investment acquisitions. Like hedge funds and PD cash desks, pension funds typically finance such positions through PD repo desks.

Others

Other users of the repo market include bank treasuries, asset managers and, to a much lesser extent, foreign official institutions. Bank treasuries can be a source of cash for their affiliated repo desk to fund long bond positions of their cash desk or sometimes provide funds directly into the repo market. Several bank treasuries have indicated that if they fund their respective repo desks, that funding would typically be provided at CORRA. Official institutions have shown a small propensity to lend out some of their bonds, and when they do, they tend to make collateral switch trades, where they lend specific bonds that trade special6 and borrow general collateral, earning a positive spread.

Repos are the plumbing of the bond market

The cash bond, interest rate swap, bond volatility and swaps volatility markets often get the most attention within any major developed fixed-income market. Often overlooked is the importance the repo market plays in providing the inner workings—or the plumbing—for both the cash bond and swaps markets to work efficiently. Figure 1 and Figure 2 illustrate the flow of cash and bonds when market participants (e.g., hedge funds, pension funds and PDs) rely on the repo market to fund purchases. Repos provide the ability to go long or short bonds7 very quickly, which leads to increased liquidity of the bond market. In the absence of a well-functioning repo market, PDs would be significantly less able to facilitate trading.8 If increased demand on repo desks to fund these long positions is large enough, repo desks will ultimately increase their repo rates, leading to upward pressure on CORRA, all else being equal.

Drivers of the recent upward pressure on CORRA

An increase in long bond positions

A significant factor contributing to the recent upward pressure on CORRA has been an increase in long positioning in GoC bonds, which is being funded through increased borrowing in the repo market. As in other developed bond markets, GoC bond yields declined in late 2023 as market expectations shifted to the view that global monetary policy tightening has run its course and 2024 may see early and pronounced rate cuts. With this shift in expectations, a number of trading strategies become attractive, including increasing outright long positions and yield curve steepeners. Meanwhile, there have been signs of a significant volume of cash-futures basis trading (i.e., long bonds, short futures). All these strategies being executed led to buying across the GoC yield curve from various types of investors (e.g., hedge funds and pension funds) who use or rely on the repo market to fund these long positions. In addition, the prospects of a start of a rate cutting cycle most likely gave PD cash desks an adequate level of comfort to increase their bond inventory (i.e., increase long positions), also through the repo market.

The blue line in Chart 1 represents the total amount of reverse repo less repo outstanding between dealers and all clients (including all maturities, overnight and term from the perspective of PDs). When the net amount is positive, it means that PDs are doing more reverse repo than repo with clients—that is, they are buying more general collateral from clients and lending clients more cash. Therefore, as the blue line rises, PD repo desks are increasing the amount of funding they are providing to clients through the repo market.

Chart 1: Increased demand for repo funding was putting upward pressure on CORRA

As the 2023 Canadian bank financial year-end (October 31, 2023) and calendar year-end approached, the typical funding pressures developed. After the Canadian bank year-end (and a relatively “dovish” November Federal Open Market Committee meeting), PD repo desks saw an upswing in reverse repo usage from clients. This upswing was driven by hedge funds whose positioning turned from net short to net long (Chart 2). This increase in demand for repo funding contributed to higher repo rates and pushed CORRA higher.

Chart 2: The increased demand in repo funding came from hedge funds

Overall, the roughly $20 billion increase in total client repo funding since the autumn has come predominantly from hedge funds.9 Further analysis also reveals a significant buildup in positioning across a basket of bonds that are the cheapest to deliver among the 2-year, 5-year and 10-year futures bond contracts. Such activity provides strong evidence of how increasingly prominent the cash-futures basis trade has become in Canada.10 We have confirmed this anecdotally in discussions with PDs.

Apart from hedge funds, the increase in repo financing for total clients (from $35 billion in July 2023 to $55 billion in October 2023) comes mainly from Canada’s pension funds (Chart 2). They were likely using the opportunity in 2023 to shift into more GoC securities as yields neared or surpassed 4%—the highest level reached in the previous decade—in the 10-year and 30-year sectors (Chart 3).

Chart 3: Government of Canada yields reached levels last seen over 10 years ago

GoC long positioning at Canada’s six largest banks, including their respective PD cash desks, continued to grow throughout 2023 (Chart 4). All else equal, this increases the demand on PD repo desks to sell general collateral to raise cash to fund these positions, putting upward pressure on repo rates and CORRA.

Chart 4: Upward pressure on CORRA increased as Canada’s six largest banks increased their net long positions

Most recently, pressures on CORRA have waned. This is likely due to the reintroduction of the Receiver General daily auctions, a decrease in market expectations for global central bank rate cuts, and significantly less repo financing needed by hedge funds because of the March bond futures expiry.

Tight cross-currency basis

Results of our discussions with repo desks point to a tightening (i.e., higher) basis between CORRA and the Secured Overnight Finance Rate (SOFR)11—the CORRA-SOFR cross-currency (XCCY) basis—as an additional reason for pressure on CORRA. When foreign money market rates and XCCY swaps are attractive, some market participants that hold a long position in the Canadian dollar can lend Canadian dollars, borrow a foreign currency and invest those funds in a foreign money market instrument. This drains the supply of cash available in the Canadian repo market because it moves cash into a foreign currency.12 During this period of upward pressure on CORRA, the level of front-end CORRA-SOFR XCCY swaps made this trade attractive to those that hold a long position in the Canadian dollar (Chart 5). As an example of this type of trade, Canadian pensions and bank treasuries have been known to lend Canadian dollars, borrow Japanese yen via a XCCY swap and buy a short-dated Japanese treasury bill. Using market rates from mid-February 2024, we find that making this trade using a one-month Japanese treasury bill would yield a Canadian-dollar return of about 5.18%—22 bps higher than a one-month GoC treasury bill and 23 bps higher than a one-month Canadian-dollar general collateral repo. The degree to which these types of trades are being implemented requires further investigation, but it has been pointed to by several market participants.

Chart 5: Three-month CORRA-SOFR cross-currency basis swap has tightened since autumn 2023

Other factors

In addition to the long bond positioning and the tightening of the XCCY basis, other factors were likely at play, but their importance is less clear and likely more of a secondary effect.

Concentrated distribution of liquidity providers

Despite there being excess settlement balances in the system as a whole, pressures from liquidity scarcity can emerge from distributional issues. In the overnight general collateral repo market, the distribution of cash lenders has been concentrated in a small subset of primary dealers and bank treasuries (Chart 6). The overall share of Lynx balances among participants has also remained concentrated and relatively stable (Chart 7).

In light of recent pressure, some market participants have questioned whether QT can continue for as long as planned or whether it needs to end sooner. Some market commentators have suggested that the steady-state level of settlement balances could be in the $70 billion to $100 billion range, according to a few of their own estimates. However, there is no consensus among market participants, and other commentators have generally agreed with the Bank’s estimated range.13 In a recent survey of Lynx participants conducted at the beginning of 2024, respondents indicated that the $20 billion to $60 billion range is still appropriate and achievable, although many expressed the opinion that the Bank may need to maintain balances toward the upper end of this range.14 Nevertheless, as QT continues, the Bank’s holdings of GoC securities will decline. All else being equal, this will reduce the amount of settlement balances in the system, which will increase pressure to redistribute settlement balances across participants.

Chart 6: Two primary dealers provide the majority of cash in the overnight general collateral market

Chart 7: Settlement balances of the top three Lynx participants have remained stable

Persistence in CORRA

Dealers typically benchmark the funding rate for the current day (t) to the latest available CORRA rate (which is based on t-1 transactions), leading to a persistence of the rate. This is because both treasuries and clients—two of three main funding sources for repo desks—have been using the CORRA published in the morning as an indication of prevailing funding costs. On the treasury side, although the relationship between the treasury and the repo desk varies from dealer to dealer, some treasuries use the CORRA published in the morning for internal transfer pricing to the repo desk. For clients, CORRA serves as a transparent benchmark level at which to transact with dealers. Some clients have informal agreements with some dealers to lend funds to them at the previous day’s CORRA. Thus, higher published CORRA on any one day—due to the previous day’s large imbalance between supply and demand—may mean dealers have fewer incentives to lower funding rates at the start of the day. This in turn leads to higher rates over the remainder of the day, all things being equal. This effect of ratcheting up interest rates is hard to undo without a more marked and likely relatively persistent decline in funding needs across the market.

Balance sheet efficiency

The interbank repo market among primary dealers is centrally cleared through the Canadian Derivatives Clearing Corporation (CDCC). A main benefit of central clearing for individual dealers is the balance sheet relief it provides from netting offsetting risk exposures.15 In other words, offsetting positions (reverse repo vs. repo) cleared on CDCC are netted against each other. If they net to zero, they do not increase the size of the PD’s balance sheet. In this example, a PD would not incur charges from its treasury desk that would result from an increase in the size of its balance sheet. Such balance sheet efficiency is beneficial day to day and can be particularly important around financial reporting periods (i.e., quarter- or year-ends). Hence, the netting benefits provide an incentive for some participants to use the interbank market, even though it could be more expensive than other sources of funding (e.g., sourcing from buy-side participants seeking to lend funds). This then puts upward pressure on CORRA. In addition, the limited adoption of central clearing in Canada from the buy side compounds this issue.

Impact of banking regulation

Some regulations placed on global banks, including banks in Canada, after the 2008–09 global financial crisis support financial stability overall but may result in relatively lower repo liquidity and more volatile pricing. These regulations can increase the cost or, to a certain point, limit the ability of banks to engage in liquidity or maturity transformation.

Conclusion

We find that the increase in long positions requiring repo financing is the main driver of the recent upward pressure on CORRA. However, we identify several other contributing factors that, to varying degrees, are present in the background. Some of these are market-based while others are more structural in nature. At the same time, the Bank has maintained its policy of QT, and assets have continued to mature from its balance sheet, putting downward pressure on settlement balances.

  1. 1. The repo plays a vital role in the functioning of any major developed bond market. A repo is the sale of eligible security to a cash lender with the requirement that the borrower of cash repurchase the eligible security at a future date. The price quoted for a repo transaction is a repo rate, which is effectively the lending/borrowing rate for the cash leg, or portion, in the repo transaction. It is the main funding tool enabling market participants to easily take long or short positions in Government of Canada securities, for example, which supports market liquidity and market-making activities.[←]
  2. 2. To learn more about Canadian repo markets, see C. Garriott and K. Gray, “Canadian Repo Market Ecology,” Bank of Canada Staff Discussion Paper No. 2016-8 (March 2016).[←]
  3. 3. A 2s10s and 10s30s are examples of yield curve trades that attempt to profit on changes in the difference in yields, or spread, between two maturity points. Hence, a 2s10s trade is based on changes in the spread between the 2-year yield and the 10-year yield.[←]
  4. 4. To learn more, see F. Avalos and V. Sushko, “Box A: Margin leverage and vulnerabilities in US Treasury futures,” Bank for International Settlements BIS Quarterly Review (September 2023): 4–5; and D. Barth, R. J. Kahn and R. Mann, “Recent Developments in Hedge Funds’ Treasury Futures and Repo Positions: is the Basis Trade ‘Back’?” Board of Governors of the Federal Reserve System FEDS Notes (August 30, 2023).[←]
  5. 5. A leveraged long bond position involves purchasing a bond and subsequently lending the whole position in the repo market for cash, which is then used to pay for the initial purchase of the bond.[←]
  6. 6. A bond is said to trade special when its repo rate is below the general collateral rate.[←]
  7. 7. A trader opens a long position in a security when they expect that the security will increase in value in the future. The opposite of a long position is a short position. A short position is generally the sale of a security the seller does not own. A trader who sells short expects that the price of the security will decrease.[←]
  8. 8. To learn more, see J.-S. Fontaine, J. Selody and C. Wilkins, “Improving the Resilience of Core Funding Markets,” Bank of Canada Financial System Review (December 2009): 41–46.[←]
  9. 9. To learn more, see J. Sandhu and R. Vala, “Do hedge funds support liquidity in the Government of Canada bond market?” Bank of Canada Staff Analytical Note No. 2023-11 (August 2023).[←]
  10. 10. To learn more, see J. Sandhu and R. Vala (2023).[←]
  11. 11. SOFR is the US equivalent of CORRA.[←]
  12. 12. This makes it less likely that the Canadian cash can find its way into the Canadian repo market as the XCCY desk that is now long in Canadian dollars would look to sell Canadian dollars in the XCCY or foreign exchange market instead.[←]
  13. 13. To learn more, see N. Bulusu, M. McNeely, K. McRae and J. Witmer, “Estimating the Appropriate Quantity of Settlement Balances in a Floor System,” Bank of Canada Staff Discussion Paper No. 2023-26 (October 2023); and T. Gravelle, “The Bank of Canada’s market liquidity programs: Lessons from a pandemic” (speech delivered to the National Bank Financial Services Conference, Montréal, Quebec, March 29, 2023).[←]
  14. 14. Lynx participants continued to cite the need for a sufficient buffer of settlement balances to avoid being short at the end of each payment cycle. They also said settlement balances play an important role in their portfolios of high-quality liquid assets.[←]
  15. 15. To learn more, see J. Z. Chen, J. Chen, S. Ghosh, M. Pandey and A. Walton, “Potential netting benefits from expanded central clearing in Canada’s fixed-income market,” Bank of Canada Staff Analytical Note No. 2022-8 (June 2022).[←]

Acknowledgements

We would like to acknowledge indispensable guidance and feedback from Mark de Guzman. Special thanks to Deputy Governor Toni Gravelle for valuable comments. The authors also acknowledge Danny Auger, Sharaf Ahsan, Michael Mueller, Adrian Walton, Rishi Vala, Jabir Sandhu and Malcolm Fisher for their helpful views and input. The views expressed in this work belong to the authors and do not represent the official views of the Bank of Canada.

Disclaimer

Bank of Canada staff analytical notes are short articles that focus on topical issues relevant to the current economic and financial context, produced independently from the Bank’s Governing Council. This work may support or challenge prevailing policy orthodoxy. Therefore, the views expressed in this note are solely those of the authors and may differ from official Bank of Canada views. No responsibility for them should be attributed to the Bank.

Content Type(s): Staff research, Staff analytical notes
Topic(s): Financial markets, Interest rates, Lender of last resort, Monetary policy implementation
JEL Code(s): D, D4, D5, D53, E, E4, E43, E44, E5, E52, G, G1, G12

DOI: https://doi.org/10.34989/san-2024-4

On this page
Table of contents