Et tu, CDOR?

With the demise of the London Interbank Offered Rate (LIBOR) and its succession by the Secured Overnight Financing Rate (SOFR) entering its final stages, the spotlight is now on Canada as it moves to phase out its primary benchmark interest rate of the Canadian Dollar Offered Rate (CDOR), to be replaced with the Canadian Overnight Repo Rate Average (CORRA), a CAD-denominated, risk-free rate alternative derived from the repo markets. This week’s publication by the Canadian Alternative Reference Rate working group (CARR) of recommended CDOR fallback language for loan products has superseded the current market approach of including generic benchmark replacement language in loan contracts, with immediate effect.

What you need to know

• The CARR publication sets out recommended language for a hard-wired approach for CDOR loans to flip to Term CORRA on June 28, 2024 (and not before), with a fallback to CORRA compounded‐in‐arrears, including suggested credit spread adjustments of 29.547 bps and 32.138 bps for one-month and three-month loan tenors respectively. The spreads are not meant to be prescriptive, but rather are intended to be used as a negotiation starting point for loan counterparties.
• An officially recognized Term CORRA rate has not yet been adopted, though CARR has indicated it believes one will be available by Q3 2023 ahead of the June 28, 2024 CDOR sunset date.
• Borrowing mechanics for CORRA Loans are not provided for in this iteration of the CARR recommended language, but CARR expects that the CORRA mechanics will look substantially similar to those used for LIBOR or SOFR loans. CARR has already published interest rate conventions for loans based on overnight CORRA compounded‐in‐arrears, and will publish recommended conventions for loans based on Term CORRA if CARR proceeds with establishing a Term CORRA benchmark rate.
• There are some significant differences between the approach taken by CARR and its recommended fallback language for CDOR transition versus the approach taken by the Alternative Reference Rates Committee (ARRC) with respect to LIBOR’s replacement by SOFR.
• Once the move from CDOR to CORRA becomes effective, CARR has suggested abandoning bankers’ acceptances (BA) funding and removing related provisions from loan contracts.

Background

LIBOR’s checkered past

Market observers may recall the LIBOR scandals that made headlines in 2012, when the implicated banks either settled with regulators or were subject to significant fines (the largest being $2.5 billion imposed on a single institution). In 2015, three traders were jailed for their roles in LIBOR manipulation.

LIBOR’s shortcomings are due to the method by which its rate is derived. Specifically, LIBOR is the average rate at which banks can obtain unsecured funding in the London interbank market. LIBOR is a "polled" rate, meaning it relies on the subjective judgment of individuals acting on behalf of submitting banks responding to the question of what rate their institution could borrow funds, by asking for and then accepting interbank offers in a reasonable market size.

At best, this makes LIBOR susceptible to human error and fallibility; at worst, it makes it subject to manipulation. Submitting banks stand to profit from rate movements on their existing loan and interest rate derivatives portfolios by inflating, deflating or otherwise obscuring their true borrowing costs. The issue is magnified in times of market duress, because a higher borrowing rate implies more credit risk for the polled bank, which exacerbates counterparty risk and propagates a negative feedback loop, which in extreme circumstances like the GFC can lead to contagion and a run-on-the-bank scenario.

CDOR, guilty by association

With the LIBOR scandals tarring all interbank offered rates with the same red brush—including CDOR—the stage was set for global loan, bond and derivatives markets to move away from interbank benchmark rates towards more robust and resilient alternative rates. The lineup of candidate replacement rates would not rely on human forecasting, but would instead be based on historical transaction data where credit risk has been substantially removed from the rate-derivation calculus. The most durable of these alternative rates are the so-called risk-free rates such as SOFR, Sterling Overnight Index Average (SONIA) and CORRA. The risk-free aspects of these benchmark rates are discussed below.

CDOR was originally developed in the 1980s as the basis for pricing BA related credit facilities and is currently the primary interest rate benchmark in Canada. It is referenced in more than $20 trillion of gross notional exposure across the Canadian wholesale financial system, including in derivatives, bonds, and loans. CARR was established in March 2018 to review the efficiency and robustness of CDOR and guide benchmark reform efforts in line with those undertaken by the ARRC in the United States, a task force convened by the Federal Reserve Board to shepherd the transition of LIBOR to SOFR in USD and Eurodollar loan markets.

It is worth noting that Canada’s federal regulator of financial institutions, The Office of the Superintendent of Financial Institutions (OSFI), has endorsed CARR’s efforts and recommendations to transition the Canadian financial system from CDOR to primarily use CORRA. As supporting reasons, OSFI has cited the move as elevating Canadian benchmark standards and aligning Canada with other jurisdictions which are adopting similar standards.

Anatomy of an interest rate

LIBOR and CDOR are fundamentally different from SOFR and CORRA

To make sense of CARR’s recommended fallback language, it is critical to understand the differences between the “old rates” (LIBOR and CDOR) and the “new rates” (SOFR and CORRA).

LIBOR and CDOR are both unsecured term rates, with LIBOR being an interbank borrowing rate (what is your bank’s cost to borrow?) and CDOR being a lending rate (at what rate is your bank prepared to lend to other Canadian banks?). Because both LIBOR and CDOR involve assuming unsecured credit and repayment risk of another bank, they are referred to as “credit sensitive rates” and necessarily include a credit risk premium reflecting the implied probability that the borrowing bank fails to repay the loan. Moreover, both LIBOR and CDOR are term rates insofar as they reflect the cost of funding for a future period of time, typically one month or three months. As such, LIBOR and CDOR funding can be thought of as a series of mini fixed rate loans of recurring one- or three-month periods. Accordingly, both rates also include a term premium reflecting the time-value-of-money (in a positive yield curve environment).

By contrast, SOFR and CORRA are secured overnight loans that are derived from the overnight repo market. This is significant because the reference loans from which those rates are derived are secured by collateral, typically high-credit quality government bonds or treasury bills, which separates the loan exposure from the credit risk of the borrower counterparty and replaces it with the liquidity risk of the security that is backstopping it. Since SOFR and CORRA are based on overnight secured borrowing costs from the previous business day, there is no time-value-of-money component’ to them. Specifically, CORRA is a measure of the cost of overnight general collateral funding in Canadian dollars using Government of Canada treasury bills and bonds as collateral for repos.

Credit spread adjustments

While lenders and borrowers across the land may be inclined to dismiss the CDOR transition to CORRA as merely more red tape and another field day for lawyers, the rubber hits the proverbial road and the financial statements when it comes to credit spread adjustments.

Recall that at bottom, the purpose of a benchmark interest rate is to be a proxy for the economic cost of borrowing in a particular currency, which in turn reflects the monetary policy of the related sovereign. In the case of CDOR, the underlying cost of money is layered upon and clouded by the additional variables of counterparty credit risk and the time value of money. Through that lens, CORRA can be viewed as a purer reflection of macroeconomic financing costs because it excludes those variables, and as such, the CORRA rate is nominally lower than CDOR. As a result, an additional spread must be factored in in order to account for those two variables and “convert” CORRA to the economic equivalent of CDOR. The amplitude of the credit spreads is a commercial point that will directly impact the borrower’s cost of funds (or the lender’s rate of return on lending).

In the case of the LIBOR transition to SOFR, the ARRC somewhat arbitrarily determined the suggested SOFR credit spreads based on five years of historical spread data as of March 25, 2021. Notably, this time period included the market and credit shocks of early 2020 from the COVID pandemic, though inflation was then well below the current inflation prints for the United States and Canada, registering in the 7-9% range on a year-over-year basis (which bears on federal monetary policy and, in turn, benchmark interest rates). As reflected in this week’s publication, CARR has adopted the ARRC approach and taken the step of publishing hard-coded credit spread adjustments for CORRA-CDOR, reflecting the historic economic difference between CORRA and CDOR calculated as of May 16, 2022. This is the date that the administrator of CDOR (Refinitiv Benchmark Services (UK) Limited) announced that the publication of CDOR rates for all tenors will cease on June 28, 2024.

Key takeaways from the CARR recommended language

• The CDOR to CORRA transition differs from the LIBOR to SOFR playbook in several important respects.
  • Unlike the LIBOR transition, CARR has not published an “amendment approach” alternative whereby the parties agree to agree to replace the current benchmark rate at a later date.
  • In contrast to the ARRC approach, CARR has not provided for an “early opt-in election” permitting the agent bank or lender to initiate the change to CORRA before the June 28, 2024 CDOR cessation date. Fortunately, CARR’s approach should simplify matters for banks from an operational perspective, since it will avoid a staggered transition of outstanding loans to CORRA over time. The entirety of the CDOR loan portfolio will remain on CDOR until June 28, 2024, after which all loans will toggle to CORRA concurrently.
  • After CDOR loans flip to CORRA, the agent or lender will have the option to discontinue BA funding. Unlike LIBOR funding mechanics which generally work for SOFR, BA mechanics do not map over to CORRA-based funding.
• CARR’s recommended language provides for a two-step benchmark replacement framework:
  1. if a recognized Term CORRA rate is available by June 28, 2024, then CDOR loans will covert to Term CORRA after that date; and
  2. if Term CORRA is not available at such time, then CDOR loans will toggle to CORRA compounded-in-arrears where the overnight CORRA rate is observed and compounded over the term of the interest period (and therefore not known in advance on the borrowing date). The CARR recommended language does provide for the ability to “go back up the waterfall” if a recognized Term SOFR rate becomes available after June 2024 when the loans transitioned to CORRA compounded-in-arrears.
• The CDOR transition mechanics and related language for loan-based products differ from CARR’s recommendations for floating rate notes, in part because the latter is typically not easily amended due to widely held noteholders and the need for unanimous consent.
• The borrowing mechanics for CORRA loans are not provided for in yesterday’s iteration of the CARR recommended language, but CARR expects that the CORRA mechanics will look substantially similar to those used for LIBOR or SOFR loans. CARR has already published loan conventions for loans based on overnight CORRA compounded‐in‐arrears, and will publish recommended conventions for loans based on Term CORRA if CARR goes ahead with establishing a Term CORRA benchmark.
• The CARR recommended language includes hardcoded credit spread adjustment which crystallized as of the date of the RBSL announcement (May 16, 2022) and are consistent with the credit spread adjustments published by Bloomberg to be used in ISDA documentation. The spreads are not meant to be prescriptive, but rather are intended to be used as a negotiation starting point. Loan participants should proceed with caution before committing to lock into the fixed recommended spreads for longer term durations, especially given the current uncertainty in the macro and geopolitical environment.