The U.S. Fiscal Outlook: Serious, Doable, and Waiting for a Deadline

Transcript

Lightly edited for readability.

Presentation

Jason Furman: Thanks for having me and for this discussion, and I look forward to what Aviva and Oya have to say. What I have to say is that the problem is serious, but it’s doable, and we’re waiting for a deadline. That would be the summary of where we’re going.

And I should start out by saying I’ve ranged over my life from extremely concerned about deficits to extremely blasé about them. And at the current moment, I feel slightly in the middle of those views.

I’m going to talk about the U.S. fiscal situation, just to give you a little bit of context in terms of where we are with the numbers right now; ask the question of how worried we should be; and then finally, what should be done. And the short version is going to be this. On the U.S. fiscal situation: under most scenarios, debt is rising as a share of the economy. It’s not sustainable. Something needs to be done. How worried we should be is harder to tell, and how immediate the problems it poses are. And then finally, in terms of what can be done, the adjustments, I think, are a little bit smaller than many people think. It doesn’t mean adjustments aren’t needed.

The U.S. Fiscal Situation

So let’s start with the U.S. fiscal situation. This is debt held by the public. The first thing to see is we’ve been here before, at 100 percent of GDP. The difference is the last time we were here, at the moment World War II ended, debt started to plummet as a share of GDP and fell for the next 30 years. This time around, we went through the financial crisis, where debt went up too little in my view. We went through COVID, where debt went up too much in my view. Either way, it should have gone up. And the difference is that coming out of them, under most projections, the debt is expected to continue to rise. I’m showing you the CBO baseline, and I’m also showing you my preferred baseline, which I will explain in a moment.

If you look at the federal deficit, the flow, in some ways that’s more extraordinary. It’s currently stabilized at around 6 percent of GDP. The only times we have ever seen 6 percent of GDP or more in our history were World War I, World War II, the financial crisis, and COVID. So this is completely uncharted territory for normal times. In fact, if you look at normal years that exclude events like those, the average has been 2.2 percent of GDP. So we’re way, way beyond that.

If you look at the deficit to GDP for advanced economies, this is one of the metrics where the United States can be proud and can truthfully say it is number one. No one else comes close at the moment. And part of why this is a concern is that U.S. debt to GDP isn’t the highest, but it’s towards the upper end of the range for these countries, and then the flow is going in the wrong direction. So countries like Italy—and Oya will say a lot more, and maybe she’ll think I’m being too nice to Italy by comparing them favorably to the United States—have a higher debt to GDP, but they’re running a primary surplus, so they’re reducing it now. We have reasonably high debt and are adding to it.

Now, the question one should ask is: what are the assumptions underlying all of these forecasts? A good place to start is the Congressional Budget Office. If you do, the debt is rising to 120 percent of GDP after a decade, and we would need to cut spending or raise taxes by 2.1 percent immediately to stabilize it over the next decade.

They leave out, in my view, a whole lot of plausible policy changes. There are expiring tax cuts. There are spending cuts being phased in that are very unpopular with at least one party. Tariffs are unpopular sometimes with some people and all the time with everyone in this building. And defense spending is basically set to decline continuously as a share of GDP in a way that might not be consistent. For what I call plausible policy changes, I take all of those and just multiply by 0.5. So I take half of all of those, and that’s about seven-tenths of a percent of GDP in new spending or tax cuts that’s being added to get to something that’s sort of like a current-policy baseline.

The second thing is the Congressional Budget Office is predicting interest rates that are about 75 basis points lower than where markets are today. Moreover, if you look at forward pricing, markets think rates are going to stay where they are now. If anything, they think they’re going to rise, not fall, whereas the Congressional Budget Office thinks they’re going to fall. So I use current market interest rates. It’s actually a little bit more optimistic—a little bit lower interest rates—than even using forward rates would be.

But then the last thing is that the Congressional Budget Office incorporated an estimate of the effects of productivity growth from AI into their baseline, but their estimate was really, really minuscule. I think it’s unlikely that they’re correct. So in my preferred baseline, I’m adding half a point for faster productivity growth, and that matters quite a lot. And all three of these adjustments end up getting you back to roughly the same place as the CBO baseline, with the productivity making up for the extra spending and the extra interest rates.

Now, there’s a much wider range of possibilities, and this table illustrates that wider range of possibilities. In all of these, I’m holding so-called plausible government policy fixed, so all of these are a current-policy-ish baseline. If you look at the upper left, the zero-zero, that corresponds to CBO. The 0.75 on interest rates and 0.5 on productivity, that 119, corresponds to my preferred scenario. You need productivity boosts of about 1.25 percentage points to feel comfortable. And by the way, as you go out on productivity, you should probably go higher on interest rates too. I didn’t build in CBO’s interest rate rule for productivity; I neutralized that in this chart, so they’re two separate variables.

So 1.25—that would mean productivity growth at about a 3 percent annual rate. I don’t think that’s outlandish to expect, but to me, it’s like the 85th percentile of where we could be over the next decade, maybe the 90th percentile. And you don’t want to make policy at the 90th percentile. You probably don’t even want to make it at the 50th percentile. You probably want to make it more at the 25th percentile, in terms of optimal control and planning for bad things that might possibly happen.

So the last thing I’d say is that we spend lots of time arguing over what the spending assumptions are, how fast health care is going to grow, what’s going on with interest rates. If you look at most of the debt we have today, it’s from large discrete events. And these are not 100-year floods, because there have been three of these in the last 75 years. And in some sense, all of our debt today is attributable to these events, with debt then being reduced in between. And that’s another thing we need to be very mindful of: we can do all this forecasting, and we don’t have these sorts of things built in. And these sorts of things are very asymmetric. There are events that can cost you 50 percent of GDP in two years. There aren’t a lot of events that save you 50 percent of GDP in two years, except maybe a hyperinflation.

How Worried Should We Be?

How worried should we be about all of it? Here’s where I’m just much, much less sure. And I think if you’re really worried, you do have to do the thought experiment: you’re in the year 2000, and somebody comes and tells you the United States is going to have a debt to GDP of 100 percent and a deficit to GDP of 6 percent. What does the world look like? And we’ve had this just massive change in our fiscal situation from 2000, and the unemployment rate is basically the same, and Treasury rates are a bit lower. And if you look at the real rate, it’s even lower compared to 2000 than what is shown here.

Now, this is not “deficits don’t matter.” This isn’t any sort of causal analysis. We don’t know what the world would be like, but it puts some bound on the magnitude at which they are currently mattering. The way I think about it is that throughout the fiscal debates of the last 25 years, we always had more fiscal space than the fiscal hawks would have us believe. It’s also the case that we’ve used a huge amount of fiscal space. So in terms of how much we have left, that I’m much more nervous about. But you go back in a time machine, and people were more worried about this than they should have been. That doesn’t mean that they’re wrong.

Now, when you look at the magnitudes of the consequences, there are sort of the known knowns, which are like termites in the woodwork, where each year it’s a little bit more than the last. The blue line was CBO’s baseline last year for the debt. The black line is CBO’s hypothetical scenario of stabilizing the debt at 100 percent of GDP. And the orange bars are on a different scale; they show the percent difference between those. And there are a few things to see. If you look on a five-year time horizon, you’re talking about less than a 1 percent change in disposable personal income. No wonder no one wants to stake their presidency on the fourth or third little orange bar there. When you look out further, it grows quite large. It grows to something like 4 percent in terms of disposable personal income.

Of course, the unknown unknowns are the bigger deal. There’s the linear effect: two or three basis points per percentage point of debt added to our interest rates. So interest rates are 75 basis points higher now because of what we’ve done recently, all that type of stuff. The bigger concern is nonlinear dynamics. And the problem there is that there’s just no good relationship between them and debt to GDP. We’ve seen countries with much higher debt to GDP than we have without crises. We’ve seen countries with much lower debt to GDP than we have with crises. In some ways, the Truss event is a reminder that institutions matter and perception matters quite a lot, just as much as the numbers.

What Can Be Done?

Last, what can be done? I’m going to scope the size of the problem. I think Aviva is going to talk more about the substance of what you would fit into it. And the first thing to say is, let’s say we want to do a 2 percent of GDP adjustment, maybe 2.5 percent of GDP. That’s not totally unprecedented. In 1990, we did 1.8. In 1993, we did 1.5. So you take the two of those together, you’re at 3.3. That’s almost certainly enough for where we are now. Now, they had defense spending cuts—the post–Cold War peace dividend—that we don’t have to work with now, but it’s doable. Other deficit reductions we’ve done since then have tended to be smaller, but some have been half a point, including, by the way, the tariffs, which are not the best way to reduce the deficit, in my view. And there have been a lot of other things that have increased the deficit at the same time. That’s not shown here; I’m just showing the negative numbers, not all the positive ones. So I think it’s doable.

What might force action? A fiscal crisis: a rapid run-up in interest rates. A second is public opinion shifting. And this is where public opinion is now: the share naming the deficit or debt as the nation’s most important problem. In 2010 and 2011, when we had a 9 percent unemployment rate, this was our nation’s top problem for 10 or 15 percent of Americans. Now we have a 4.1 percent unemployment rate, and it is the top problem for almost no one. But you can see also that public opinion can shift quite quickly, and maybe someone will catalyze on this in a presidential campaign.

The last way, though, I don’t think is super likely, but I think it’s much more likely than any other way of approaching this, which is that Social Security’s retirement trust fund is exhausted in 2032, and Medicare in 2033. The fiscal gap in the two of those programs is about 1.75 percent of GDP. So if you cared about the debt on a present-value basis and you addressed both of these—and by the way, I’m just doing numbers here, so you could do it by raising taxes, you could do it by cutting spending, you could do it by whatever combination—just numerically solving the two of these, plus doing PAYGO, plus having a limited set of emergencies, gets you pretty close. In some sense, they’re like accounting forcing events, and I think we should try to hold on to them, use them, and shape them well.

So, to summarize: the debt and deficit are high and rising. Higher productivity might get us out of the problem, but it feels like, not a fringe tail, but a bit of a tail. The debt is undoubtedly costing our economy. We’re seeing it in rates today. I don’t think it’s huge. The most attractive forcing event is Social Security and Medicare. And then, finally, I do want to say: if you ask me my worry list, this is on there, but it’s not number one, two, or three. And what we can do to have a more people- and future-oriented budget is sort of beyond the scope of my discussion. But I’m hoping Aviva speaks to it more eloquently than I would have had I tried. Thank you.

Discussion

[Aviva Aron-Dine and Oya Celasun then gave their presentations, after which the moderator, Maurice Obstfeld, put one question to the panel.]

Maurice Obstfeld: I’m going to ask the panelists one quick question—it’s probably not an easy question—and then we’ll move on to the Q&A. One of the striking facts that Jason’s presentation made clear is that the public just doesn’t care. And it seems like policymakers also don’t care, in many cases, certainly not in the U.S. There’s very little evidence that it’s high on the agenda. But there was a time when people cared and when Congress cared. The very nice Brookings paper by Alan Auerbach and Danny Yagan showed this: there was a time when Congress would react to forecasts of higher deficits. This hasn’t been the case for about a quarter century. So my question is, how did we get here? And is there a way, short of crises, that we as economists can try to educate the public and educate policymakers about the risks that we can clearly see but they seem to be ignoring?

Jason Furman: Let me first comment on that nice paper by Auerbach and Yagan that I’m sure everyone here has committed to memory. What they found was that it used to be that when deficits were higher, you were more likely to see tax-increase or spending-cut legislation. That was true before 2000. It hasn’t been true since 2000. I think it would be fair to say the tone of the paper wasn’t just documenting that change but was a little bit lamenting it: what happened to people caring?

I think for much of the period they document, the people were right, and the old reaction function would have been wrong. We had interest rates near zero. We had heavily constrained demand. As Aviva was saying, it’s not just whether spending is worthwhile for the resources; it might create jobs, so even less worthwhile spending could be justified. So in some sense, the fiscal scolds were scolding for that whole period, and they were wrong in real time, in ways that were knowable. And the political system, if anything, was at times overly concerned about the deficits, as it was in 2010 and 2011, which showed up as too little stimulus. That wouldn’t show up in their data set. Now, I think it’s correct to lament it today. I think their point has become true.

So part of it may be an optimistic reading—which is probably way overoptimistic—that people roughly cared about deficits more, most of the time, when deficits mattered more. In the ‘80s and early ‘90s, the deficit really was an issue. Interest rates were higher. It receded as an issue when it wasn’t as important. And maybe, with some lag, it’ll come back as one. I think that’s one possibility.

Another possibility is admitting that if you warn every year for decades, at some point people stop listening to you. So having the credibility to say, “I wasn’t worried a decade ago, I am worried now, and here’s why,” might persuade two people. And for the other hundred million people, I think it will take an advertising campaign by the Peterson Foundation to reach them.

[Aviva Aron-Dine noted that the poll asked people to name the single most important problem facing the country, and that she would not name the deficit either.]

Jason Furman: Let me say one small thing about that. There are other data series, like how much of a concern it is. They don’t go back as far, but they’re pretty correlated with this one. So if you look at how concerned people are, it tells a similar story.

[Aviva Aron-Dine and Oya Celasun then answered the question.]

Questions from the Audience

Audience member: I had a question about the recent run-up in bond yields around the world. Is this like a canary in the coal mine? Is this something that could accelerate a crisis? Or is the financial media making a mountain out of a molehill?

Jason Furman: I’m happy to go first. First of all, you were correct in saying around the world. It’s not just that over the last three months yields have risen at similar paces in different countries; it’s week to week. They’re very correlated. So it looks like there’s a common factor basically everywhere except Japan. So the first thing to say is that this is global. That isn’t to say it isn’t caused by U.S. fiscal policy—it’s a global capital market.

Second, I’m slightly skeptical of the fiscal explanations, in that since June, the most recent run-up—or if you look at a longer time horizon—we haven’t gotten a whole lot of fiscal news. In fact, if anything, in the United States I thought the fiscal situation was going to be really bad, and instead it’s just pretty bad, or quite bad. That’s because the tariff revenue exceeded what I expected, and the spending cuts they did exceeded what I expected. So I don’t see the fiscal news or the fiscal surprise. Are interest rates 50 or 75 basis points higher because of the run-up in debt over the last 20 years? I think that’s quite plausible. Are they 50 basis points higher than they were three months ago because of the run-up in debt over the last 20 years? I find that implausible.

So what are some of the other things? Certainly, when you look at spreads, there’s some repricing. If you’re asking why France is different from Germany, you want to look to events in France, and that has a lot to do with it. So you get some of the spreads there. But in general, the recent run-up, at least to me, looks like central banks are extremely credible: inflation expectations haven’t risen, but they haven’t risen because people think central banks are going to raise rates a bunch to stop inflation. So we’re expecting a higher rate path because we’re expecting more inflationary pressure, and we think they’re going to defeat it. This is much more about inflation, the macroeconomy and the central bank’s response function. Underneath it is this bedrock of the debt, but that’s not really what the last year and a half, or the last half year, is about.

Audience member: My question is about what might happen if the AI bubble continues to grow until next year and then pops, and that leads to an increase in unemployment, slows growth down, and particularly slows TFP growth down for, let’s say, several years. How would that affect the deficit, revenue elasticity, and all of the other problems we’ve been talking about?

Jason Furman: I have views, but I just answered the last one, so if somebody wants to go first, you can. No one seems that eager to.

First of all, you didn’t ask whether there’d be a bubble, and I can give you all the reasons there’d be a bubble—except nothing about my stock portfolio, which is all buy-and-hold index funds, reflects any of those views, so I question whether I really have those views. For the valuations to be sustained, you need the productivity growth to materialize, you need the capabilities to keep improving, and you need to be able to profit from them. For me, the better the capabilities get, the more you can use the fifth-best model, and it’ll be more than good enough for you. And that’s a big threat to the business model at the top. So that’s why I’d be concerned about a bubble.

Conditional on one, to me it feels more like 2000 than the global financial crisis. Financial crises are largely when you thought something was money—it was worth a dollar for a dollar, whether that’s your bank account or your money market fund or the AAA bond you were using as collateral—and you discover that thing isn’t actually money. It isn’t worth a dollar. And when that happens, there are just huge ramifications throughout the system. Here, all of this is already properly understood to be risky. So it’s more like a stock that goes down: at the end of that you have less money, but it isn’t collateral running all through the financial system. So to me, it would feel more like a mild recession. The Fed has a lot of room to cut rates. We have really low demand in every other sector—in construction, at least—so you could make up for some of that.

So in some sense, I think this has some impact on what we’ve been talking about, but it’s more like bumping around over a two-year horizon. The bigger question is whether we get productivity growth from AI, and that’s to some degree orthogonal to the question of whether it’s a bubble. I think our fiscal situation and our growth all rise and fall on that over a 10- or 20-year horizon: how the technology works out on the supply side, not what happens on the demand and financial side.