Should We Reduce Federal Budget Deficits Now?
Dean Douglas Elmendorf
April 10, 2019

Good morning and welcome. I’m delighted to be back at Brookings and honored to have been asked by Bill to speak on this important occasion.

Bill Gale has been one of the most thoughtful and interesting voices on U.S. tax and budget policy for 30 years. If you have not spent enough time with Bill or Bill’s writings to know that, then you should definitely read the book and find out why that has been true. If you have spent enough time with Bill and his writings to know how thoughtful and interesting he is, then you do not need me to tell you to read the book, you know you should read it. If anyone here does not have a copy yet, I expect you to buy one on your way out today. I have learned a lot from Bill over many years, and I learned a great deal more from this terrific book.

I also want to say, here at the start, how grateful I am to Bill for giving me the opportunity to join Brookings in 2006. At that point I had been in Washington for more than a dozen years, but I had led a mostly sheltered professional life: I had been lucky enough to have a number of fascinating jobs that gave me the chance to engage in important policy discussions behind the scenes, but I had had very few opportunities to participate in the public policy debate. Bill decided to take a chance on bringing me out of the shadows and into the public domain. That position at Brookings changed my life. I will always be grateful to Bill for taking that chance and giving me that opportunity.

Bill’s book Fiscal Therapy is about making choices. As he explains, federal debt is larger relative to the size of the economy today than at almost any point in the nation’s history, and under current policies, will continue to rise relative to the size of the economy indefinitely. That path is not sustainable, and therefore we will be forced, at some point, to make choices that increase federal taxes, decrease federal spending, or both.

But Bill views the necessity of making choices as an opportunity as well. I think that is exactly right. As Bill explains, the federal government is not taxing or spending in ways that advance our national interest very effectively. Productivity growth is slow by historical standards, yet we have reduced federal investment in infrastructure and research and development to nearly the smallest percentage of total output in my lifetime. Inequality has risen significantly and economic mobility has declined, yet we have not responded with policies to expand access to high-quality education and training for either young people or adults. The share of the American population over age 65 is rising to unprecedented levels, which increases federal spending on our largest and most popular benefit programs, yet we have reduced taxes to a smaller percentage of national income than for any sustained period in 80 years. And the tax system we have is neither efficient nor equitable.

We can do better. We need to do better. And the way to do better, as Bill says, is to “focus on realistic solutions” based on “facts and evidence.” There is much that economists and other policy analysts do not know about how the world works, and experts should prescribe fiscal therapy with some humility, just as other sorts of experts should have humility when they prescribe physical therapy. Nonetheless, there is much that economists and other policy analysts do know about the effects of alternative policies, because of careful thinking, historical experience, and serious research.

Bill’s outstanding book presents the crucial facts and evidence about federal budget policy that elected officials should use in making budget choices. I agree with many of Bill’s specific recommendations for changes in tax and spending policies, and I am really delighted to see so much good sense on the federal budget offered in one place.

Rather than run through Bill’s menu, however, I have decided to focus my comments on one issue--in particular, an issue where my views may differ somewhat from Bill’s views and the views of the other panelists. That issue is whether we should be trying to reduce budget deficits in the near term. I think we should not be trying to do that. Specifically, I think the evidence shows that reducing projected deficits will be necessary ultimately but would be a mistake to do in the near term. Let me explain.

My starting point is the low level of interest rates. The U.S. economy is now roughly at full employment, which is wonderful, and the federal funds rate is only 2½ percent. Most members of the Federal Open Market Committee expect a slight further increase at some point, while financial-market participants are split about whether the next move in the funds rate will be up or down. Thus, 2½ percent seems pretty close to the equilibrium funds rate in today’s economy. The yield on 10-year Treasury notes is also about 2½ percent. That implies that financial-market participants expect short-term interest rates to stay close to 2½ percent, on average, over the next decade. Moreover, longer-term Treasury rates are consistent with short-term rates staying close to the current level, on average, for a long time.

Of course, predicting interest rates is very difficult. If I was good at it, I’d have made my fortune and be speaking to you via video link from an island in the South Pacific. So, one should take interest-rate predictions with a grain of salt. Nonetheless, both logic and empirical evidence imply that a future of low interest rates is likely.

As the financial crisis was beginning a dozen years ago, the picture of interest rates that many of us had in our heads was of low rates in the 1960s, a peak of rates in the 1970s and 1980s because of high inflation and the subsequent effort to drive inflation down, and then again low rates in the 90s and 00s. Interest rates fell in the financial crisis and Great Recession, which was unsurprising, but then failed to rise significantly as the economy recovered, which was surprising. If you draw a picture of interest rates today, and drop off the years before 1990 so you can rescale, you see a pronounced, consistent decline in interest rates over the past 30 years.

A pattern of thirty years’ standing is not a flash in the pan. In addition, a similar decline in rates can be seen in many countries around the world. Moreover, as Lukasz Rachel and Larry Summers argued at the Brookings Papers on Economic Activity conference last month, the global decline in interest rates has occurred even though public debt has surged in many countries.

As this pattern of globally low interest rates has become more apparent, a number of analysts have examined a large number of possible explanations. Some of the possibilities involve a declining demand for capital, stemming from diminishing capital intensity of production, slower trend growth of the labor force caused by demographic changes, and slower growth in total factor productivity. Other possibilities involve specific features of Treasury securities that might push down their yield relative to the return on capital. And there are other possibilities as well. I think a fair summary of this growing literature is that explaining interest rates, even in retrospect, is a complex and uncertain business, but that there are good reasons for rates to stay low for the foreseeable future.

If interest rates really will stay low for a long time, what does that mean for our subject today? Persistently low interest rates represent “a sea change for [federal] budget policy.” That conclusion is a quotation from a paper that Louise Sheiner and I wrote in 2015 and published in the Journal of Economic Perspectives in 2017. It is well understood that low interest rates change the dynamics of federal debt: For any given primary deficit—that is, for any given deficit excluding interest payments—lower rates mean that debt will rise more slowly, all else equal. It is less well understood that low interest rates change the desirable levels of deficits and debt.

Analyzing the implications of low interest rates for federal debt is complicated in part because the implications vary to some extent depending on the reasons that interest rates are low. Louise and I presented a summary version of our analysis in our published JEP paper and a longer, more technical, version in an unpublished paper on the Brookings web site. Our bottom line was that “many—though not all—of the factors that may be contributing to the historically low level of interest rates imply that … federal debt … should be substantially larger than [it] would be otherwise.” Of course, you should not just take our word for it. But other economists, including some more-distinguished economists, have reached similar conclusions. And I am not aware of any research that has examined the implications of low interest rates for optimal budget policy and reached opposing conclusions.

Therefore, it is now appropriate to conclude that the United States should have substantially more federal debt than most economists would have argued several years ago. That conclusion further implies that there is much less urgency to putting federal debt on a sustainable path than most economists would have argued several years ago.

This line of thought does not say that federal debt can rise indefinitely relative to total output, which is what would happen under current laws or policies regarding federal revenues and spending. I will emphasize: Debt cannot rise indefinitely relative to total output, and none of the recent research to which I am referring implies otherwise. Therefore, reducing projected federal budget deficits will be necessary at some point. The crucial question becomes, when should deficits be reduced? I think the right answer to that question is: not now.

There are legitimate reasons to believe that we should reduce federal budget deficits in the near term, but I think those arguments are less consequential than they first appear and are significantly outweighed by the arguments for substantially delaying action. Let me explain.

One legitimate argument for acting soon is that interest rates might rise unexpectedly. As I noted earlier, all interest-rate predictions should be taken with a grain of salt, and current predictions might be wrong. Therefore, I think we should hedge our bets and make policy decisions that would not produce terrible outcomes even if interest rates turn out to be higher than current projections. We should also consider what policies would not produce terrible outcomes if rates turn out to be lower than current projections. Indeed, Larry Ball, Greg Mankiw, and I wrote a paper 25 years ago in which we analyzed the implications for budget policy of the risk that interest rates could rise unexpectedly and economic growth fall unexpectedly, making a previously sustainable debt suddenly unsustainable. We noted that Treasury interest rates have generally been low historically because they provide insurance against damaging, low-probability events, and we concluded that sensible policymakers should take out insurance against the risk of a debt suddenly becoming unsustainable by accumulating less debt than otherwise.

However, as I noted earlier, current interest rates are quite consistent with the evolution of rates over the past 30 years around the world. Moreover, interest rates were generally quite low before the run-up in the 1970s, so the anomalous period looks increasingly to be the 1970s and 1980s, not the period since then. As a result, I do think we should hedge our bets against a run-up in interest rates, but only to a limited extent.

A second legitimate argument for acting soon to reduce projected budget deficits is that significant changes in budget policy should be enacted well before those changes need to take effect, so that the people whose benefits or taxes are being changed can plan ahead. For example, we are still phasing in an increase in the full retirement age for Social Security that was legislated almost 40 years ago—and the gradual pace of that increase has helped some people prepare.

However, most changes in tax rules and spending programs do not need to be announced far in advance. Moderate increases in income or payroll tax rates, for example, do not require significant advance notice, and many people are not very far-sighted about government policies even when they are already in law. After all, if everyone was saving an optimal amount based on their expected retirement income, we would have much less need for Social Security than we do. As a result, I worry only a little about enacting changes well in advance of their implementation.

A third legitimate argument for acting soon to reduce projected budget deficits is that elected officials tend to avoid hard budget choices and therefore analysts should keep the pressure on. We see this sort of avoidance in a collection of federal budget decisions over recent decades, in decisions by state and local government leaders to promise pensions to public employees but not to fund those pensions, and in other ways.

However, this inclination toward budget laxity is hardly absolute: Our policymakers did not cut taxes or increase spending sufficiently in the Great Recession, with significant negative consequences for Americans. German policymakers performed even worse, with terrible consequences for people across the Eurozone. I did think in 2011 and 2012 that the Congress would have been more likely to enact short-term fiscal stimulus if our long-term fiscal outlook had been brighter. But I think now that our long-term fiscal outlook cannot conceivably become bright enough soon enough to significantly alter the backdrop for future stimulus decisions and that those decisions will need to be debated on their own merits.

So, if those arguments for near-term action for reducing budget deficits are not as strong as they first appear, what are the arguments against near-term action? I think there are two:

First, increases in federal taxes or reductions in federal spending would slow the economy in the short run, and with the federal funds rate as low as it is, the contractionary effect of fiscal tightening would be dangerous. Fiscal tightening would reduce aggregate demand and, all else equal, reduce output and employment in the short run—even though it would otherwise boost output in the long run. Under the circumstances that we have viewed as “normal” for most of our professional lives, that contractionary impulse would be offset by reductions in the funds rate arranged by the Federal Reserve. But current circumstances are not normal in the traditional sense, because the Fed does not have much room to cut the funds rate. Therefore, reducing budget deficits today would present a significant risk of reducing output and employment, with all of the attendant economic and social costs.

The second reason that we should not take action to reduce federal budget deficits in the near term is that most of the reduction should stem from an increase in revenue, and Republican leaders are not ready to support revenue increases. I think we should focus on increasing revenue because the spending reductions that would be most likely to occur would be worse for our society at this point than letting federal debt continue to rise. Here is why.

There is no appetite in either major party for significant cuts in the large federal benefit programs. The current Republican president has promised to maintain Social Security, Medicare, and Medicaid—although his actual proposals contradict that promise—and a Republican-controlled House, over 8 years, voted on no legislation that would substantially reduce federal benefits except for repeal of the Affordable Care Act. If the Republicans will not cut those large benefit programs, then surely the Democrats will not either. There is also little appetite in either party for significant cuts in annually appropriated nondefense spending. Such spending is already nearly the smallest percentage of national income in more than half a century, and no matter how much some elected officials like to complain about federal spending in the abstract, they know that their constituents want their food inspected, their borders patrolled, and their highways repaired. In addition, there is little appetite among a significant number of policymakers in substantially cutting defense spending.

That leaves only the benefits focused on lower- and middle-income Americans. If federal spending was cut in the near term, this is the category of spending that would bear the brunt of the cuts. That makes no sense to me. Contrary to some assertions, the current amount of benefits does not provide a comfortable “hammock” for people who do not or cannot work. A growing body of evidence suggests that at least some of those benefits enhance rather than diminish the life prospects of children who receive the benefits. And we have seen rising income inequality and declining upward mobility in this country. So, why should we cut benefits for the people who have gained the least from overall economic growth over the past several decades? In my view, the social costs of such cuts would outweigh the long-term benefits of putting the federal budget on a more sustainable path.

Therefore, I conclude that, for both economic and social reasons, we should not aim to reduce federal budget deficits in the near term. However, we will need to do so eventually. And in the meantime, we should work hard to improve “the way we tax and spend,” as Bill Gale argues in his excellent book. Bill offers us all a terrific set of ideas for how to do that, and I hope that analysts and policymakers follow his lead. Thank you.