Bill Nelson on Fed losses
David Beckworth has a great podcast with Bill Nelson. At one point Nelson discussed recent Fed losses due to rising interest rates (which depress the market value of bonds held by the Fed.)
Nelson: Sure. Yeah. So as you know, I send out these periodic emails on monetary policy to anybody who wants to receive them. And if they’ve your listeners who want to receive those emails, should just email me, they’re free. And I’m happy to add you. But I sent one out a couple weeks ago, pointing out that the Fed probably lost about $500 billion in the first quarter. And that we’ll know in a month or so when the Fed releases its quarterly update of its balance sheet, or maybe even sooner when the New York Fed releases its annual projection of Fed income and balance sheet. But in any case, the logic was fairly simple. The Fed’s interest rates rose very sharply in the first quarter. On average the yield curve rose about one and a quarter percentage points.
And the duration of the Fed’s securities portfolio, it indicates is about five years. And so if you just do the math, that’s a 6% loss. And a 6% loss on eight and a half trillion dollars in securities is about $500 billion. So, that’s a loss on their books. Some might argue that that doesn’t matter because they’ll never realize that loss. But I, really as any good economist, would completely disagree with that. And one way to think about it is really the same view that you were just describing that Seth articulated, which is that the present discounted value of remittances are lower by $500 billion. And you can think about that either as the loss on the securities or the rise in the IORB rate. But basically they’re going to be remitting to Treasury and therefore you and I, and everyone else are going to be paying higher taxes with a present value today of $500 billion. That’s just what the math tells you.
And the losses of course could be considerably higher as interest rates keep going up. And I agree that primarily the problem here is political. But it is a real loss. If you think about the consolidated balance sheet of the Federal Government, the Fed’s QE actions shortened up that balance sheet. They replaced long-term debt with short-term debt. They’re borrowing through the Overnight RRP Facility and through reserve balances, that’s the new debt. And they retired as it were long-term debt that they bought. So the consolidated balance sheet of the U.S. Government now has a shorter duration. Which means that the government is now more exposed to the rising short-term interest rates then they would’ve been if the Fed had not taken these actions. And that’s a real effect.
Now a hasten to add that a comprehensive view of this needs to take into account the benefits that accrued from the Fed’s actions. So, insofar as the Fed’s QE stimulated the economy, and that boosts tax revenues, and that’s something that needs to be considered as well. Now, I think you could probably make a pretty good case that the Fed’s three trillion or so of purchases around in the spring of 2020 to sort of save the financial system were extraordinarily beneficial. In terms of the long flow-based QE program that they then launched into, I am less convinced that that actually added a lot of value. But many would disagree. And so it’s that whole picture that needs to be considered.
For simplicity, I’m going to focus on the Fed’s holdings of T-bonds. They also hold some MBSs, but nothing important hinges on that distinction, at least for the purposes of this post.
How should we think about these losses being absorbed by the Fed? Let’s start with the fact that in a fiscal sense the Fed is part of the federal government. So when the value of T-bonds declines, that’s a gain for the Treasury and an equal loss for the Fed. For the consolidated federal government balance sheet it’s a wash.
Nelson suggests that the loss from rising interest rates is real. Does that contradict what I’ve been saying? Not really, it depends how one frames the question.
Image a world with a Treasury that borrows long-term, but there is no Fed. In that case, rising interest rates would benefit the Treasury in the sense that the market value of their debt would be smaller than if they had faced rising rates with nothing but short-term debt. Their previous decision to borrow long would have been wise, in retrospect.
Now imagine the Fed is created and starts buying up some of that long-term Treasury debt, and replaces it with interest bearing reserves (a short-term liability.) The Fed has effectively shifted the consolidated federal balance sheet toward shorter maturities. So the federal government gains less than otherwise from a period of rising interest rates. In that sense the Fed has imposed a loss (as Nelson suggests). The Fed’s $500 billion loss reduces the Treasury’s even larger gain that results from having the market value of its previously issued long-term debt fall by much more than $500 billion.
Given that the Treasury is a huge borrower, it might seem odd for me to claim that the Treasury gains from higher interest rates. Consider this statement by Nelson, which seems to suggest the opposite:
Which means that the government is now more exposed to the rising short-term interest rates then they would’ve been if the Fed had not taken these actions.
Once again, there is no contradiction once you understand what’s going on. Rising interest rates affect the Treasury in two distinct ways. First, the Treasury benefits from a decline in the market value of its existing liabilities. But only longer-term liabilities; T-bill prices are barely affected at all. Second, the Treasury is hurt because it has to pay higher rates on future borrowing.
Now let’s consider a scenario where the higher interest rates reflect higher inflation (the Fisher effect). In that case, the Treasury is an unambiguous winner. The real value of its existing debt declines, and the real interest rate paid on new debt does not increase. This is one way the government benefits from the inflation tax. (Others include seignorage and increased capital gains tax revenue.)
In the past year, real interest rates have risen somewhat, but remain very low. (Near zero on 5 and 10-year bonds, and negative on shorter maturities.) So I suspect that the high inflation of 2021-22 has been a net plus for the federal government, despite somewhat higher real interest rates for new borrowing.
PS. It’s possible the Fed might need a bailout (although I doubt it.) But if it does, it’s mostly a symbolic issue, sort of like the symbolic Social Security “trust fund”. Of course symbolic issues can become political, and I’d expect a Fed bankruptcy to be highly controversial.
PPS. My argument might not seem to apply to Fed holdings of MBSs, but keep in mind that those are close substitutes for T-bonds. From the perspective of the consolidated federal balance sheet, the distinction between the Fed buying T-bonds and MBSs is not that important. It has some importance for capital allocation, but even there I believe the big problem lies elsewhere (the Treasury guarantee of MBSs, for instance.)