LIBOR replacement in client products and prospective pricing

Question

I am asking whether the industry, or single banks, have made up their minds already on how to replace the 'missing' interbank risk compensation component in their variable rate credit products when transitioning from LIBOR to SOFR, or whether (and how) they are about to change the financing (or hedging) mix to more closely reflect the RFR-based products.

If you feel that an answer to this question is overly opinion-based, I am of course happy to reformulate it in order to make it more helpful for the community.

My understanding of the status quo: LIBOR

Putting the LIBOR scandal aside, I understand that LIBOR is an inherently credit risky interbank rate reflecting the average (perceived) funding cost for a given tenor, say 3M. Let's assume our bank's funding cost $y_b$ to be linked to the interbanking market, and hence to LIBOR, as a somewhat stable spread $s_b$ over LIBOR. If my bank hands out credit to a customer at LIBOR plus a fixed spread, I am able to pass on the industry's 'average' refinancing cost to them, and hence also the corresponding risk from increasing LIBOR. The client rate $y_c$ consists of LIBOR plus a client spread $s_c$ $$ \begin{align} y_c&=LIBOR + s_c \\ &=LIBOR + m + k + \bar{s}_b+\alpha\sigma(s_b) \end{align} $$ The client spread $s_c$ is commonly fixed at closing of the deal. (Of course, there are provisions in place that adjust the spread when the client's creditworthiness deteriorates.) Here, $m$ are other undefined margin requirements, $k$ is a client credit risk component, and $\bar{s}_b$ and $\sigma(s_b)$ are my bank's 'average' spread and spread variability (think:vol), and $\alpha$ gives us some margin of error.

In this world, I think I could quite well price my client's contract at some well understood rate $s_c$ over LIBOR and refinance / hedge most of the interest rate risk conveniently at LIBOR for a steady income of (very roughly) $m$.

Question

In the future, with LIBOR out of the equation: Are banks, and if so how are banks, able to pass on parts of their (average) refinancing rates (and thus risk) to their customers? Will there be some form of regime shift where

• banks begin to assume the interbanking risk and charge more to their customers?
• and/or: refinancing will be based on some fudged RFR-curve with individualized banks spreads?
• and/or: we might see liquidity in banking-industry-CDS-index-swaps (yes, CDX-swaps) used to hedge financing risk, with hedging cost passed on to the customer?

Answer 1

This question is the subject of much current debate amongst regulators and banks. You are absolutely correct , many banks are alarmed that the demise of Libor will make their asset-liability management more difficult. Historically , banks have originated loans linked to Libor , reasoning that they will be able to pass along fluctuations in their funding costs to their loan customers. Note that it is indeed the cost of unsecured financing that is of interest here. Banks raise unsecured funding via at least two methods (a) deposits, which are relatively cheap and (b) wholesale unsecured financing such as interbank loans and bank commercial paper.

In the US there is currently a vocal community of regional banks that are in favor of creating a replacement for Libor that retains the credit spread above the RFR. Several indices have been created which are candidates for this role, the most prominent being Ameribor and BSBY. My understanding is that there is some loan origination linked to these new indices already. There is also some loan origination linked to the RFR and some that is still linked to Libor. The Fed has made clear that it expects no more loan origination linked to Libor after Jan 2022, so this is becoming an urgent situation. My reading is that the Fed tolerateS Ameribor , BSBY and other alternatives for loan origination but they reject the use of these for widespread interest rate derivatives trading, lest they reproduce the exact same situation that occurred with Libor. They prefer the derivatives market to be based on the RFRs which are based on a much higher volume of trades than Ameribor , BSBY or any other unsecured lending index.

For more information , try googling the minutes of the Alternative Reference Rate Committee (Arrc) which is a committee convened by the Fed and attended by major banks and investors. There is also information online about Ameribor and BSBY. See below for a letter to Fed from regional banks.

https://www.politico.com/f/?id=0000016d-d15d-d0d8-af6d-f77d6c5f0001

Answer 2

Here's my take on this. Firstly, in my experience, banks do not raise financing as a function of Libor, but rather as a function of the local Central Bank policy rate.

As an example, the Czech Central Bank provides a 14-day reverse-repo facility ("the main financing rate"), whereby it lends govies for two weeks in exchange for cash, on which it pays interest $r_{Central}$.

As a result, if a local bank A wants to borrow from a local bank B, local bank A will need to pay at least $r_{central}+s$, where $s$ is some small spread: otherwise bank B has no incentive to lend money to A and run credit risk on A, because it can just deposit cash at the Central bank at rate $r_{central}$ hassle-free.

Therefore it is the Central Bank deposit rate which provides floor to interbank borrowing, and the local Czech Libor trades at a (pretty much) fixed spread to the Central Bank deposit facility rate.

The same principal holds for USA-based or Eurozone-based banks: the local central bank always has some policy deposit rate (overnight, one-week, two-week...), and this policy rate drives the interbank borrowing. Libor then tends to be quoted at a spread to these central bank policy rates.

Libor has always been a fictional rate: no one will lend you money unsecured for 3 months!

Let's say that USA-based banks can borrow from each other at $r_{Fed-Fuds}$ + spread (or $r_{SOFR}$ + spread). If a bank keeps borrowing at this $r_{Fed-Fuds}$ and needs to keep rolling it over, it can hedge this via a Fed-Funds OIS Swap: i.e. they agree to receive compounded $r_{Fed-Fuds}$ on a quarterly basis, and agree to pay semi-annual (or quarterly) fixed $r_{fixed}$.

That way, they still need to keep rolling over their borrowing in the interbank market at $r_{Fed-Fuds}$ on an overnight basis, but the OIS swap will allow them to be reimbursed "in hindsight" every quarter and turn this overnight compounded Fed-Funds rate into a fixed.

In the end, they can then choose to lend money to clients at the rate $r_{fixed}$ + spread, or they can choose to index the lending to client at $r_{Fed-Funds}$, averaged over the past quarter, updated on a quarterly basis...

Edit: interestingly, just a few days after the question was asked, the FT reports that Bank of America has issued its first ever syndicated loan linked to SOFR (https://www.ft.com/content/a94fa64c-2723-4d90-93fe-a218683148e5):