Should you sell at a loss?
Credit unions often face the challenge of managing loans or fixed-income assets that were originated or purchased in a lower interest rate environment and are now “underwater” from a pricing perspective, resulting in a loss if sold. However, credit unions willing to accept these losses, the proceeds can be then used to originate new loans or purchase higher-yielding loan participations. This strategy can generate significant economic upside in certain cases.
As an initial example, consider a credit union that originated a 72-month auto loan a year ago during a time when market rates for a loan of that term and other risk attributes (e.g., credit score, DTI, etc.) were 6%. A year later, interest rates have increased and that same loan would have a market rate of 8%. Should the credit union sell that loan, ignoring servicing fees and assuming an 18% prepayment rate, that loan would sell for approximately 96.50% of par value. (Note that all figures are illustrative.)
The credit union now has a choice — it can sell the loan at a 3.5 point loss (assuming the loan was originate or purchased at par) and immediately invest the proceeds in a new market-rate loan, either by originating one or purchasing a participation in one, or it can take the monthly principal and interest generated by the existing loan and use that to originate or purchase smaller amounts of new, market-rate loans over time. In the former case, the credit union would immediately have a lesser principal amount invested that is earning interest at a higher rate; in the latter case, the credit union would have a greater principal amount invested that is earning interest at a lower rate until the original loan fully pays off (to keep things simple, we’ll assume that rates stay constant).
It turns out that, in the long run in a perfectly efficient market, both scenarios end up in the same economic place. Selling the loan at a loss and purchasing market-rate loans or participations generates immediate higher income through higher interest rates, even though it is earned on a smaller balance which, over time, offsets the loss. On the other hand, retaining the loan results in lower income due to the lower interest rates, even though it is earned on a larger balance. Over time, paydowns are used to invest in (or originate) market-rate loans, eventually converting the loan into 100% market-rate loans.
The following graph summarizes the difference between these two alternative approaches over time. The graph shows both the difference in balance as well as the cumulative difference in interest received over time. As the additional interest that is received in the new loan is applied to buy more loans, the difference shrinks, with the total cumulative interest differential eventually exactly offsetting the initial loss take.
So, if things will eventually end up in the same economic place, why should a credit union even bother?
There is one key reason for a credit union to sell at a loss. Should the credit union sell “riskless” assets (e.g., Treasurys, Agency MBS, etc), and invest in or originate spread products such as loans, the increased yield can more than compensate for the initial loss in part because those assets do not have the same credit and liquidity profile.
For instance, if a credit union had purchased the 5-year Treasury three years ago in November 2020 at a yield of roughly 0.4%, and sold it 3 years later as a 2-year, which is trading at a roughly 4.6% yield, the price would be around 92, an 8-point loss. From our analysis above, we know that should the credit union make the sale, take the loss, and buy a lesser amount of 4.6%-yielding Treasurys, the interest income would increase, but in the long term, the alternatives of holding or selling would end up breaking even. However, if the credit union sells the Treasurys and purchases a loan participation in, for instance, direct auto loans, which are currently trading at a loss-adjusted spread of roughly 200 bps over Treasurys, they would instead own a 6.6% net yielding asset. (It’s important to note that, unlike the first analysis where losses were ignored as they would equally impact both alternatives, here, since Treasurys are considered “riskless” securities, we are reducing the yield in the loans by expected losses and only using the loss-adjusted yield.)
In this case, the extra yield earned on the loan participation makes up for the loss realized on the sale of the Treasury and then some, with the breakeven occurring in about 16 months. The following graphs show how the balances and interest evolve over time, assuming full reinvestment of all cash flows in new loans:
As a note, the wider the loss-adjusted spread, the shorter the breakeven period; if, for instance, the loss-adjusted spread in our example were 300 bps instead of 200 bps, the break even reduces to 14 months.
Unlike a loan, a Treasury security pays only interest on a semi-annual basis, with the principal paid at maturity, with no possibility of early payment (with the exception of callable securities). Combined with the low interest rate on the Treasury and the materially higher loss-adjusted rate on the loan, this allows the participation to make up for the loss, and in fact, come out ahead over a relatively short timeline.
In the past, many credit unions had excess cash from pandemic stimulus that they couldn’t utilize in their own originations or loan participations. As a result, they invested in Treasurys or related securities. Today, these credit unions are in a favorable position to take advantage of this trade. With interest rates currently higher and spreads wider, it is an opportune time for credit unions to increase their realized net interest margin and potentially recover any realized losses they may have incurred. This strategy can be implemented in a relatively short period of time, particularly through loan participations.
