The Fed’s losses have passed $100 Billion: What’s next?

There has just been a landmark event at the Federal Reserve: Its accumulated operating losses have passed $100 billion. This startling number, which would previously been thought impossible, was reported on Sept. 14, 2023, in the Fed’s H.4.1 Release.

It is essential to understand that these are not mark-to-market paper losses; they are real cash losses resulting from the Fed’s expenses being to this remarkable extent greater than all its revenue. These are equally losses to the U.S. Treasury and thus costs to the taxpayers; nonetheless the Fed keeps officially insisting that its losses don’t matter. (Meanwhile the Fed’s mark-to-market losses are more than $ 1 trillion in the most recent public report.)

The Fed started posting operating losses in September 2022, so it has taken only 12 months to produce this ocean of red ink. Looking forward, annualizing the year-to-date results suggests the Fed’s net loss for calendar year 2023 will be about $117 billion. 

We must add the $17 billion it lost in the last four months of 2022, so at the end of 2023, the accumulated losses of the Fed may be about $134 billion. That means the losses will have run through the Fed’s entire capital of $43 billion, plus another $91 billion, for losses of more than three times its capital. What comes next?

The losses will continue into 2024. As long as short-term interest rates stay at their current historically normal levels of around 5 percent (that is, if we have “normal for longer”), what the Fed has to pay on its deposits and borrowings will create a punishing negative spread against the Fed’s trillions of very long-term, low yielding investments. 

Since the Fed bought heavily at the top of the market and the bottom in yields, these investments yield on average only about 2 percent. It doesn’t take a Ph.D. in finance to see that lending at 2 percent while borrowing at 5 percent will not be a winner.

The Fed owns $5 trillion of Treasury securities, of which $4 trillion have more than one year left to run, $2.3 trillion more than five years and $1.5 trillion more than 10 years.  (One of its investments is in the Treasury 1.25 percent of 2050, for example.)  

The average yield on these Treasury securities is 1.96 percent, according to the most recent Fed Quarterly Financial Report. The Fed also owns $2.5 trillion of 30-year mortgage-backed securities (MBS), of which $2.4 trillion have remaining maturities of more than 10 years. 

The MBS have an average yield of 2.20 percent. Combined, that suggests an overall yield of 2.04 percent, which compares to a current funding cost of deposits and repurchase agreements of about 5.37 percent. Voila the Fed’s problem: A negative spread of more than 3 percent on investments with very low yields locked in for years to come. In short, the Fed made itself into a gigantic version of a 1980s savings & loan.

An estimate of the Fed’s running rate of losses in very rounded numbers is as follows.  It has $7.5 trillion of investments yielding 2 percent. Of this, $2.3 trillion are financed at zero interest cost by circulating dollar bills, so the Fed is making about $46 billion a year from its government-granted monopoly of currency issuance. There are about $5.2 trillion of remaining investments financed at a -3 percent spread for an annual loss of $156 billion. There are $9 billion in overhead expenses. The approximate annualized running rate is thus:

                           Profit from currency monopoly                        $46 billion

                           Loss on leveraged investments                        ($156 billion)

                           Operating expenses                                           (9 billion)

                            Net (Loss)                                                            ($119 billion)

By mid-2022, the Fed knew it had serious operating losses looming and published a projection of them. Its base case projection was for a peak cumulative loss of $60 billion — the reported $100 billion loss is already a lot more than that, let alone the $134 billion loss likely by the end of 2023. And the ultimate peak losses will be far greater still — if 2024 is anywhere near as bad as 2023, that alone takes it far over $200 billion. (The Fed also considered a remote, “tail risk event” of its mark to market losses reaching $1.1 trillion. With its current mark to market loss, the “tail risk event” has already happened.)

The Fed should regularly provide Congress updated calculations of the possible path of its future losses. From what has been reported so far, we know that to struggle back to break even, the Fed will have to wait what looks like years for its underwater investments to roll off, or short-term rates would have to fall a lot, or some combination of these. 

We might guess that, for example, it will take a runoff of $3 trillion of its long-term investments with a drop in short-term interest rates to 4 percent to get close to break even.

Failing that, and in the meantime, the Fed itself, the Treasury and the taxpayers will be suffering continuing huge losses. We will simultaneously be running an interesting test of the Fed’s claim that these losses don’t matter.

Alex J. Pollock is a senior fellow at the Mises Institute, the author of “Finance and Philosophy—Why We’re Always Surprised,” and co-author of “Surprised Again!  The Covid Crisis and the New Market Bubble.”