[co-author: Ashley Longman]
It may not be Taylor Swift’s next song, but an early payment changes the lender-borrower relationship. In an exchange, we all know there are consequences. Rather than a “break cost,” the swap market simply calls it an early termination fee. In lending, traditionally, less time/energy was spent negotiating terms requiring the borrower to compensate a lender for breach costs. Today, however, it is a hot topic. Specifically, in the context of SOFR loan prepayments where the concept of “break costs” is a nebulous/unclear concept, leaving some market participants to wonder, in a post-LIBOR world, “Are there still “break costs” for loans simply due to prepayment?”
Hot plug: The term SOFR has break costs for the same reason as LIBOR. The same goes for Daily Simple SOFR, although the costs may be lower than the Term SOFR. As explained in more detail below, “break costs” describe the costs to the bank of managing expected cash flows between SOFR assets and liabilities (largely funded by SOFR assets). Breakage costs are not opportunity costs. The following outlines some aspects of bank risk management in broad strokes, but the basic concept is there, and that basic concept justifies a cost break.
LIBOR-World: Breaking costs = LIBOR deposits
In the days of LIBOR, it was not uncommon to see a loan include a concept of compensating the bank for “breaking costs” (even if there was no prepayment penalty). While the same loan provided any guidance as to the calculation of any breakage fees, it was often based on the concept of what the bank might otherwise have earned on a “LIBOR deposit” or otherwise put its funds to use. This provision may be the source of this confusion. It clearly links break costs to a concept of opportunity costs, i.e. costs reflecting what the bank would otherwise have gained by purchasing a LIBOR asset at the start of the interest period.
SOFR-World: Breakage costs = ???
Now parties are seeing prepayment of a SOFR loan and struggling to justify the breakage fee because, to paraphrase, “If the bank needs a SOFR return for the rest of the interest period, it can simply participate in the overnight repo market”. First, let’s leave out whether it’s even possible for a bank to be prepaid on day T and have the same funds available for the repo market that night. The thought is not wrong, but if you prepay forward SOFR, the chances of forward SOFR being perfectly equal to real SOFR are close to zero. The question is not what the bank might have gained otherwise. The question is whether the bank will have met expectations for its cash outflows (eg, SOFR liabilities) based on the receipt of forward SOFR from the borrower. If the bank does not receive forward SOFR to cover its SOFR liabilities, the bank may need to find another SOFR asset and purchase it (or write off a SOFR liability) as part of its broader management of the bank’s entire risk portfolio. The bank won’t just bet (or expect) that the overnight repo market will be close enough to the forward SOFR.
Context: 2020 Banking and risk management
Some parties may view the business of the bank as making loans and then using the income from that loan to fund other loans or pay normal expenses. It’s the old “Bailey Building and Loan” model, and George Baily is everybody’s banker. It’s your bank of the 1920s/30s (and the best Christmas movie, sorry elf).
In today’s banking industry: The goal is to minimize the time a bank has static money on its books. If cash leaves the bank, then the bank generates revenue from (i) fees and/or (ii) spreads. Here, “spreads” really means any concept where the bank asks a customer to pay $XYZ for a product, and that value is higher than cost to the bank to provide this product.
What is also important to realize here: Cash outflows are not always a one-time event (eg, the granting of a loan) – i.e. a bank may hold a liability that requires many recurring payments. Example: an exchange. The diagram below shows how banks will think about swaps and fund the bank’s obligations on that swap. It’s not “unique” or “Ed, of course… in some cases, but not all”. A big nuance missing here is that sometimes the ‘hedging asset’ can be unique (as described below using a swap) or the bank can engage in portfolio hedging (so not 1 to 1 ). Also, rather than a cost of “1.0% per month”, the cost could be a one-time payment for the purchase of SOFR paid tickets/titles.
In any case, the bank does not commit to a SOFR liability (here, a commitment to pay SOFR on a swap) without an asset that is expected to produce the return on SOFR. An important function of treasury/risk management teams is to manage this risk. The bank does not want to store the risk ofWell I don’t know what SOFR will do tomorrow, but I bet we can afford it.” To this end, the bank wants a Term SOFR asset to pay for any Term SOFR liability.
Outage Costs = Risk Management Costs
Banks are trying to maintain cash outside from the bank in sync with incoming money in from (i) the Federal Reserve, (ii) assets purchased or (iii) assets produced by the bank. The problem with (i): you have to repay that loan. The problem with (ii): these usually cost more than (iii). This makes (iii) a fairly efficient means of financing banking activities.
This is why your SOFR loans always require a provision for cancellation fees. If the borrower repays early, the bank will incur expenses because it needs to adjust its expectations of the SOFR revenue from that loan. If a term SOFR loan is prepaid on Monday, then on Tuesday the bank needs an asset to produce that term SOFR expectation. Good risk management here is not “Hey, just do overnight restsbecause the forward SOFR rate on January 1 is not impacted by a sudden drop in SOFR rates during this interest period. So: For the bank to pay forward SOFR, it needs a forward SOFR asset. If he can’t find a perfect match, then not only are the costs of buying an imperfect asset “breaking costs”, but any delta that costs the bank money must also be included in the “breaking costs”.
To be fair, the costs may be low on a loan-by-loan basis, but a bank will want to consider this on a portfolio basis. For a SOFR Daily Simple loan, the costs may be even lower, but there is always a good chance that a bank will not have funds at the door on T-day, and the same funds are used on the repo market that night (so there will always be a cost to the bank to plug the hole during the period when the prepaid funds cannot otherwise be tapped).
These are not prepayment fees/penalties. It is simply the bank demanding that the borrower return the entire bank for incurring new costs to manage a new risk created by the prepayment.
There may be other costs as well, but the above is one cost that makes sense to me and provides at least one good reason why (i) lenders should retain “break-fee” provisions and ( ii) a borrower trying to negotiate it, should at least appreciate the bank is not unreasonable. Ultimately, it’s up for negotiation (like anything else), but perhaps that context can steer the negotiation away from focusing solely on whether or not there are breakout costs.
If breakage fees exist, how much $$$$ are the breakage fees?
Good question. Some LIBOR loans never went into this detail, others did and based it on a LIBOR deposit. A possible answer would be to use the USD SOFR ICE swap rate (HERE), whenever ICE starts posting it. This rate, as of the prepayment date, applied to the prepaid amount for the remainder of the interest period or fiscal quarter, on an ACT/360 basis, should be a fair estimate of the cost to the bank. That said, it can be argued that a premium could be added as there are transaction costs beyond what can be captured by the USD SOFR ICE swap rate. To be clear, there are also other possible measures (for example, simply opt for Term SOFR or Term SOFR + Adjustment for the remainder of the period of interest). Perhaps that is where the parties should focus their energy today. Not on the existence of breakage costs, but rather on “They exist, but breaking up with our new partner SOFR is expensive.”