Achieving a Sustainable Federal Budget

Thursday, October 20, 2011

Douglas W. Elmendorf, director of the Congressional Budget Office, discusses the effects of revenue increases and spending cuts on the projected budget deficit.

This meeting is part of the C. Peter McColough series on International Economics presented by the Corporate Program and the Maurice R. Greenberg Center for Geoeconomic Studies.

CHRYSTIA FREELAND: I'm Chrystia Freeland, your presider over our breakfast this morning. I was delighted when the council invited me to do this, because we're going to have the chance to hear from, I think, one of the most important guys in Washington, and therefore in the world economy right now, Doug Elmendorf, who is the director of the Congressional Budget Office.

I'm a journalist, so I did some reporting beforehand, and I heard lots of fabulous comments about him. The best one was someone said to me: Doug is one of the only guys in Washington who cares, above all, about getting it right. So I don't know what that says about everybody else in Washington -- (laughter) -- but it was a real tribute to him and to how he looks at the issues. And he also said to me beforehand that he doesn't have any particular point he wants to make because he just looks for the truth and what's really going on.

So how good is that?

I want to ask everyone please to turn off all of your devices. I'm sure you have many of them. Also to remind everyone that this meeting is on the record, so that's good news for my journalist colleagues who are at the back table.

There are also CFR members in Washington who are participating in -- via teleconference. And there is another meeting tonight, for real gluttons for punishment, Professor Jonathan Steinberg as part of the National History Center Series.

How we're going to do it this morning is Doug will make about 10 minutes of remarks from the podium; then, as you can see from these seats here, he and I -- I will ask him questions for about 20 minutes; then we'll throw it open to everybody else.

So please, Doug.

DOUGLAS ELMENDORF: Thank you, Chrystia, for that terribly extravagant introduction. It's a little hard to live up to that. I am delighted to be here.

As you know, a special congressional committee is currently trying to agree on policies to reduce future federal budget deficits. The challenges and choices that committee faces are the same challenges and choices faced by all of our elected leaders and by all of us as citizens. The aging of the U.S. population and the rising costs for health care mean that the combination of budget policies that worked in the past cannot be maintained in the future.

Therefore, achieving a sustainable federal budget will require the United States to deviate from the policies of the past in at least one of the following ways. We will need to significantly increase revenues relative to GDP, or we will need to make major changes in the sorts of benefits provided to Americans when they become older, or we will need to substantially reduce the role of the rest of the federal government in the economy and the society.

Let me explain why the choices are so stark. Begin with Social Security and the major health care programs, by which I mean Medicare; Medicaid; the Children's Health Insurance Program, or CHIP; and in the future, the subsidies to be provided through new insurance exchanges. Under current law, we expect that spending on those programs will total about 12 percent of GDP in 2021, compared with about 7 percent, on average, during the past 40 years -- an increase of 5 percent of GDP.

All other federal spending, apart from interest payments on the debt, has averaged about 11 1/2 percent of GDP during the past 40 years. That includes outlays for defense, the largest single piece; unemployment compensation; food stamps; veterans' benefits; federal, military and civilian retirement benefits; transportation; education and training; health research and so on.

Suppose then that Social Security and the major health care programs continued as they are in current law, and that all other programs were operated in line with their historical relationship to the size of the economy over the past 40 years. In that case, total noninterest spending by 2021 would be the sum of 12 percent for Social Security and the major health care programs, and 11 1/2 percent for all these other noninterest programs -- or a total of almost 24 percent of GDP. Well, revenues have averaged about 18 percent of GDP during the past 40 years. Thus, by 2021, our historical policies on the spending side would push noninterest spending above historical revenues by about 6 percent of GDP.

And that gap explains why the past combination of policies cannot be repeated when it comes to the federal budget. The pieces that used to add up don't anymore with a much larger population of older people and with much higher costs for health spending.

So what choices might we make for the future? Well, one choice is to follow current law. Where would that lead us? Well, as I said, spending on Social Security and the major health care programs would be about 12 percent of GDP by 2021. In contrast, all other spending -- again, apart from interest payments -- would be much smaller as a share of the economy in 2021 than it has been on average during the past several decades.

And that reduction stems primarily from the path of annual appropriations, discretionary spending, that was set in place in the caps enacted in the summer's debt ceiling agreement. And we projected this category of all other non-interest spending will be about 8 percent of GDP in 2021, well below the 11 1/2 percent that it has averaged over the past 40 years.

So the other side of the budget, revenues are currently about 15 percent of GDP. With a weakness in the economy, and the tax cuts enacted partly in response to that weakness, revenues during the past three years have been the smallest shares of GDP since 1950.

But we project that revenues will rise sharply relative to GDP in the coming decade owing to a combination of expected economic growth, some underlying features of a tax system, and very importantly, the expiration of a number of important provisions of a tax code scheduled to occur under current law.

And altogether we project that revenues will reach about 21 percent of GDP in 2021 under current law -- about 3 percentage points above the 40-year average and the highest share of GDP recorded.

So current law will produce a combination of revenues at a historically high level and spending, apart from Social Security and the major health care programs, at a historically low share of GDP. And those differences from the past mean that the great increase in spending on Social Security and the major health care programs can be accommodated with a fairly small budget deficit. Under current law, we think the budget deficit will be about a percent or percent and a half of GDP in 2021 and that debt held by the public will actually be declining relative to GDP in that year.

Of course, though, many analysts and policymakers and citizens have expressed concern about those differences in revenues and spending relative to the historical pattern. Let me summarize the spending concerns first and then touch on the tax concerns.

Spending, apart from Social Security and the major health care programs, falls into three broad categories, each of which will reach historically low levels relative to GDP by the end of the decade under current law. The first of these categories is so-called "mandatory spending," most of which constitutes what people normally call "entitlements." That includes food stamps, unemployment compensation, some veterans' benefits, federal retirement benefits and other things. Such spending has averaged nearly 3 percent of GDP over the past 40 years. It will be about half as large relative to GDP, 1 1/2 percent, under current law by 2021.

The second category here is defense spending, which has averaged about 5 percent of GDP during the past 40 years. The caps on discretionary spending enacted this summer don't distinguish, except for the first few years, between defense and nondefense spending. It's a cap on overall discretionary spending.

But if one supposes that that cap is applied proportionately to defense and nondefense spending, then defense spending would be 2 1/2 percent of GDP in 2021, not counting the cost of any overseas conflicts, which would represent an exception to the caps, essentially. That 2 1/2 percent of GDP would be the lowest share of GDP that defense spending has represented during the past 40 and more years.

And the third category of the spending is nondefense discretionary spending. This goes to transportation, to health research, to veterans' health care, education and training, housing, international affairs, administration of justice, and other things. This category of nondefense discretionary spending has averaged about 4 percent of GDP during the past 40 years, with a good deal of variation but no evident trend. In other words, our policy implicitly has been to have this category of spending grow with the size and income of the population.

But if the caps on discretionary spending were applied proportionately to defense and nondefense, then nondefense discretionary spending would be less than 3 percent of GDP by 2021, its lowest mark in more than 40 years.

So each of the three broad pieces of spending, apart from Social Security and the major health programs, will be at their lowest shares of GDP by the end of the decade, and that's why many observers argue that the resources headed for those purposes will be inadequate to meet the country's needs.

Let me now turn to revenues. As I noted earlier, a key reason that we project revenues to rise so much over the coming decade relative to the size of the economy is the expiration of some tax provisions scheduled under current law, and those changes would produce a tax system that is noticeably different from the one that people have become accustomed to.

Scheduled to expire at the end of this year are the provisions that limit the reach of the alternative minimum tax, and other things. Scheduled to expire at the end of next year are the reduction in tax rates on expansion of credits and deductions enacted in 2001 and 2003 and 2009.

If these expiring tax provisions were extended rather than being allowed to expire as under current law, revenues in 2021 would not be the 21 percent of GDP that we project under current law, but instead about 18 percent of GDP -- close to the historical average. And as you know, many policymakers and many citizens have expressed support for extending many parts of the expiring provisions.

So where does that leave us? Well, the policymakers and citizens who want to attain sustainable deficits but who object to the current law trajectories that will produce revenues above the historical average and spending below the historical average -- spending apart from Social Security and the major health programs -- if you object to those paths but want to achieve a sustainable federal budget, then one needs to propose and support changes in the remaining large piece of the budget, which is spending on Social Security and the major health care programs. Moreover, even those who are satisfied with these paths in the coming decade generally focus on Social Security and the health care programs in considering the budget pressures that will arise beyond these next 10 years as the population ages further and as health care costs rise further.

As I said earlier, we project that spending on Social Security and the major health care programs will total about 12 percent of GDP in 2021, compared with the 40-year average of about 7 percent of GDP. That increase of 5 percentage points is obviously very large relative to the other pieces of the federal budget. To put spending for those programs back to their historical average, if you wanted to let all of the other pieces follow their historical averages as well, would require about a 40-percent cut in spending on Social Security and the major health care programs by 2021. That would require really dramatic changes in who's eligible for those programs or how much is spent on each beneficiary, or both. And of course, many policymakers and citizens have expressed concern about changes in that direction as well.

But we are truly on the horns of a dilemma. We cannot achieve a sustainable federal budget without changing at least one major aspect of budget policy in a permanent and significant way. We simply cannot sustain the spending programs and policies of the past with the revenues that we've been accustomed to collecting. We and other citizens will have to either pay more for government services and programs, accept less in those services or programs, or both.

Thank you. I'll stop there. (Applause.)

FREELAND: OK. Well, thank you very much, Doug. All those people who said nice things about you to me beforehand didn't warn me also how depressing at some level hearing you lay things out with such clarity is.

An area that I wanted to start with is health care. And I shudder to ask you a question about this with Peter Orszag sitting here looking at us.

ELMENDORF: And I shudder to answer with Peter here, but (we'll attempt it ?).

FREELAND: Well, good for you -- right. Yes. So you -- in looking at health care, how much of it is -- how much of the solution can lie not just in thinking about the numbers, but in thinking of effectiveness? As you know certainly extremely well, the U.S. compared to other developed Western countries doesn't get much bang out of its health care buck.

ELMENDORF: Yes, there's a great deal of evidence that the U.S. health care system provides less-good care at higher costs than it could. And that comes partly from comparisons with other countries, also very importantly from comparisons across regions of the United States. Medicare spending, which has been studied most in this regard, varies a great deal across parts of the country even after one adjusts for the sorts of things that one would want to adjust for, like the underlying health of the population and the cost of doing business in certain areas. Even after that sort of adjustment, there's a very large variation in spending that does not have a correlated difference in the health of the -- of the people.

And Peter and others have talked at some length about the opportunities that that presents -- the opportunity to reduce spending while making people -- keeping them at least as healthy or making them healthier. The challenge is to go from that fact to the specific policy changes. Health care is delivered in different ways in different parts of the country. The practice patterns of physicians vary a good deal across the country. And one cannot -- Miami is a high-cost area, Minneapolis is a low-cost area. One cannot turn Miami into Minneapolis simply by taking 30 percent off the checks that are sent. One needs to change the way health care is practiced.

And I think this point is well-understood by people who are practicing health care. There's a, I think, terrific ferment among health care practitioners in trying to find ways to achieve savings without harming and hopefully with helping people's health. And that's true, by the way, in Minneapolis, as well as Miami. It's not that the places that are low-cost relatively are sitting -- resting on their laurels.

But it's not straightforward. Making these sorts of changes is not easy. One small example, perhaps -- or not so small -- might be health IT. As many people have observed, when you go to the hospital, you can often end up giving a lot of information that somebody writes down by hand, and you can go to the next department and give the same information that somebody else who also writes it down by hand. And there's a lot of effort under way to improve use of IT by doctors and in hospitals.

But -- and in fact, the federal government is trying to support that effort in some powerful ways, but it's not a thing that can happen overnight. It's actually complicated to do. And there are up- front costs in some cases -- monetary costs and also up-front costs in how the work is organized.

So I think there's a -- there's a general consensus about some of the directions in which the health care delivery system needs to change, but much less agreement and just knowledge about the specific sorts of changes that should be made. And what that means from a federal policy point of view is that it's more difficult to know what federal policy should be doing to accelerate this process of change. And the major legislation enacted last year included a lot of provisions designed to experiment with different ways of delivering care and paying providers for care and with ways to make those experiments be -- then be pushed into the -- out -- throughout the health care system.

But just which of those policies and which of those changes in the way health care is delivered will work is not clear. And so I think this -- the problem we have is that this is a process that can take some time. But we don't really have the time because of the nature of the budget challenge. And that I think -- so I think it is -- it is a race, essentially, whether the health care system can evolve in ways that deliver more -- care more efficiently at a fast enough pace to alleviate the strain on the federal budget, state and local government budgets and private budgets quickly enough to reduce the -- reduce the problems that all these different entities are facing in paying these bills.

FREELAND: If you were a politically omnipotent health care czar able to resolve all of these issues you've discussed, what kind of savings as a percentage of GDP could you get out of making health care more efficient?

ELMENDORF: That's a -- that's a fair question but not an answerable one. I -- (laughter). The essence of CBO is that we don't make policy recommendations.


ELMENDORF: Let's be clear about that.

FREELAND: That's why I'm offering you a chance to be the czar. (Laughter.)

ELMENDORF: And even with that special CFR "czar exception," I can't make policy recommendations.

How much one can save depends on what one is willing to do and what chances one is willing to take. So if one -- you know, there's a lot of tools, but there's no analytic answer about how far one should push them.

I'll just give you one quick example. If you charged -- if you made people pay more for the health care they receive out of pocket -- and you can do this in Medicare, for example, by restructuring the way Medicare's copayments and deductibles work and by restructuring the way Medigap policies wrap around Medicare. If you change that, then people will spend less on health care. The evidence, I think, is quite clear about that. But it's -- the evidence also says that people will spend less on some things that -- where they probably should spend less, and they'll spend less on some things where they probably shouldn't spend less. So if you make them pay a lot more, then you'll save more money, but there will be some consequences of that.

So there isn't an analytic --

FREELAND: Like they'll get sicker?

ELMENDORF: So some of them will -- some of them will get sicker. And particularly, I think, low-income people who pay more, the evidence suggests, will scrimp on some things that are penny-wise -- turn out to be penny-wise but pound-foolish. There isn't an analytic answer to how big copayments should be. So there isn't an analytic answer to how much one can save.

FREELAND: Earlier this week I interviewed Jeff Immelt. And one of his sort of big points he wanted to make was that the overriding concern should be with growth and that, in some ways, the political conversation, in his view, is in the wrong place and that growth really fixes everything. To what extent -- you know, you've painted a very clear picture for us of the challenges America faces. To what extent do varying levels of growth change that picture?

ELMENDORF: Growth matters a good deal, no doubt. And I think there is a different sort of policy debate about what would be good for our growth of the economy in the short run, in the medium run, in the long run. But it's not possible for us to grow our way entirely out of the challenges that I -- that I posed in my (talk ?).

The challenges that I talked about, the numbers I gave are based on our current economic projection. There are people who think that our current economic projection is too optimistic. Actually, Peter Orszag is one of those people, and he's written to that effect. If we were to lower the lower the forecast, that would make those problems worse. If there were better news on the economy and we hoisted up our GDP path, that would make those problems somewhat smaller. But the problem can't basically be solved in that way.

Now, every piece helps. I don't want to suggest that it's not important for the budget picture to have a strong economy. All I'm saying is that, well, we'll still need to make difficult choices about whether to cut back on the sorts of programs and services we're used to from the federal government or increase the money we're going to put in. And that choice will still be there in a quantitatively significant way for any plausible path of GDP growth.

FREELAND: And what if a person is even more pessimistic than Peter is and takes a really gloomy view of sort of, you know, three to five years of economic stagnation? How grim does the budget picture get? What --

ELMENDORF: I don't have -- I don't have -- I don't have precise numbers at hand, but it can get quite a bit worse. I mean, there's -- you know, and I should say our forecast is not what I would think of as crazy optimistic. We actually expect rather slow economic growth. We look for the unemployment rate to get down below 6 percent, I think, by 2017. So we're not looking for some miraculous recovery in the near term.

But you know, a very important part of the widening of budget deficits over the past few years has been the weakness in the economy. Another important part has been the policies that were -- that were adopted to respond to that weakness. But you know, over the last four years, I guess, we have basically doubled the amount of outstanding debt held by the public from the federal government. And I think maybe half of that is just due to the -- to the economic path, with maybe another half, very, very roughly, due to the policies that have been -- that have been enacted.

So the -- if one continues with a -- with a shortfall of GDP relative our potential level of output, one continues with the sort of shortfall we've been seeing for many years to come, then we will -- we will rack up very significant increases in -- very large deficits and very significant increases in debt, and it will make these challenges harder.

FREELAND: Your job in some ways is to be, you know, the guy holding the ring, the guy that people can trust to tell the truth about this stuff. And your issue has now, you know, become one of the hottest political issues in the country. To what extent has that heat added light, from your perspective? You know, the focus on deficits, how truthful and, you know, dealing with the facts is the debate?

ELMENDORF: Well, political debates involve a great deal of heat as well as light, but I think there has been much more attention from citizens and from my bosses in the Congress on budget issues over the past few years than there had been in the period before that, and I think that has really helped many people understand better the challenges that we confront.

Of course, that attention needs -- should be and is competing with the attention being offered to many other policy issues. Even though I run a budget office, that does not mean that we at the CBO believe the only issues that should be discussed are budget issues. (Chuckles.)

FREELAND: Surely not.

ELMENDORF: Obviously, a great deal of attention is being paid and should be paid to the terrible slump that the U.S. economy is still in. And of course, there are very important issues in foreign policy and so on. And so when one thinks about what Congress is doing, one needs to recognize that they're trying to grapple with a broad range of topics. And in fact we do work, I should say, on a lot of topics, not just the overall budget picture.

So I don't think it's a bad thing in any way that these issues I talked about here get only part of the attention of the policymakers, but I do think that the discussions that have been going on, and even some of the failed negotiations that have gone on, have been important in illuminating the choices for members of Congress and, I think to some extent, for the American public as well.

FREELAND: How about tax reform? To what extent can that be a fix?

ELMENDORF: Well, it depends a little what one means by tax reform. So if one defines tax reform as making revenue-neutral reductions in tax rates and increase -- broadening of the tax base, then, well, revenue-neutral means that it's not changing the amount of revenue being collected. If one means by tax reform lowering rates and broadening the tax base in a way that ultimately brings in more revenue, then that can make an important difference, depending, of course, on how much extra revenue one brings in.

Certainly if one wants to bring in more revenue, it is increasingly important to have a tax system that does that in an efficient way. I mean, if one is collecting a very small amount of revenue, it's not so important to the overall economy whether you collect it in an efficient or inefficient way. The more you're collecting, the more it matters for the economy to collect it in an efficient way. So I think in that sense, tax reform, if viewed as a way of making the tax code more efficient, can fit naturally with trying to raise additional revenue.

But there are many people who don't view tax reform as having that piece of raising revenue. They simply want to -- want to make a more efficient tax code, which is very important and useful in its owner right, but then one needs to turn to other aspects of the budget to address the deficits.

FREELAND: What will you be looking for from the supercommittee? And how much hope do you hold out for those discussions?

ELMENDORF: Well, I won't -- I guess I can't talk about my hope. We don't make political forecasts at CBO. We find economic and budget forecasting quite difficult enough.

What we are doing right now very, very intensively is giving supercommittee members and staffs estimates of the budgetary effects of different proposals that they are considering for reducing budget deficits. And we've done a tremendous amount of that. We do that all the time, of course, for the various committees of the Congress, but we're doing that right now for the deficit reduction committee.

So I guess all I can say is that they are clearly engaged very seriously in discussions about what policy changes might be made; but where that process will lead, whether agreement will be reached on large or small amounts of deficit reduction, I just don't know.

FREELAND: When it comes to scoring various proposals, is there a cheat that annoys you the most, or that we should sort of be most kind of aware that people try to sneak in?

ELMENDORF: Well, CBO has worked for many years at trying to do our analysis of proposals in ways that illuminates the true effects of proposals on the budget and on the economy and isn't -- tries not to be distorted by efforts to game the procedures we follow.

I think one of the most -- but distinguishing between a game and a legitimate policy choice is a difficult matter and not one that is really given to us to offer authoritative opinion on. (Chuckles.) So I'll give you an important example, which is policies that are enacted for short periods of time. Well, there are many reasons why Congress might want to enact policies for short periods of time, from a policy perspective. It's also true that if you enact certain policies for short periods of time they can look less expensive than if you enacted them for a longer period of time.

And so what we try to do is -- in our work, is to talk about the longer-term effects of policies if they were continued, even if they are initially enacted for a short period of time. In the budget projections that we provide, our baseline projections follow current law, because they need to, because it's not up to us to say what we think the Congress really means or will ultimately do.

However, over the past several years, as more and more policies have been enacted for short periods of time, and that many policymakers talk about continuing indefinitely, and thus as the current law baseline becomes less and less useful as a guide to the underlying stance of fiscal policy, we have given greater prominence in our projections to an alternative projection that doesn't follow current law but instead tries to capture, as best we can, current budget policies. That is intrinsically more difficult, because your view of current policy and mine might be different and we -- so that can never replace our projections under current law, but I think it is a very important complement. And if you look at our work, you will find very prominently right up front not just a current law projection, but also a projection based on extending expiring tax provisions and some other changes that many people on the Hill talk about continuing.

And so I think in that area that's the way in which we're trying to avoid being gamed in some way. There are other aspects as well, but probably the issue of temporary versus permanent is the most -- quantitatively most significant of those.

FREELAND: OK. Thank you very much. I think I will throw it open to questions -- many, as we anticipated.


Q: Thank you. Herb London, Hudson Institute. It strikes me that when you have a $1 1/2 trillion deficit the proposal that I'm about to bring to your attention may not be so significant; and yet there's a conspicuous absence in any discussion of fraud in Social Security, Medicare and Medicaid -- now accounting, at least by my estimate, to close to $200 billion a year, and perhaps could even have a chilling effect on further expenses.

The actual expenditures in looking at these matters, investigation as well as litigation, account for two milles -- two milles for every dollar spent. I didn't say two cents; I said two milles. So there's virtually no investigation that takes place. If that were to occur, the recovery is quite significant and might have an effect on the total expenditures that take place in these three categories.

ELMENDORF: So I think there actually isn't a lack of discussion of that among the people I work with on the Hill. There's a(n) ongoing interest among members of Congress in trying to find ways to improve the integrity of payments made through Social Security and Medicare and other programs and to -- on the other side of the budget, to improve compliance with the tax code. In fact, the deficit agreement from this summer included provisions designed to increase the sort of surveillance that you're describing, the efforts to reduce that sort of abuse.

I think your estimate of the amount of fraud is higher than others that I've seen. (And while it seems to recognize that ?), it's not that there aren't some efforts being made already, but there's no doubt that if one made more efforts to crack down on abuses, that one would bring in -- well, reduce spending that goes out, or on the tax side increase the revenue that comes in. I think that actually is a -- is a subject of ongoing discussion.

But it's not -- it's not trivial to do. On the tax side, for example, a number of the provisions that were put in place in the past few years to try to increase compliance are being taken out -- (chuckles) -- or being discussed to be taken out, because of the burden that they impose on law-abiding taxpayers.

And I think as people think on the spending side -- for example, if one made payments more slowly to providers in Medicare, that would give more opportunity to check whether the payment was appropriate before it was sent. It would also make it harder for the providers to use those payments to meet their bills. So one ends up confronting pushback for legitimate reasons in various ways. But I think that is an issue that's on the table.

FREELAND: And do you have a preferred fraud estimate? You said that one was higher than ones you've seen.

ELMENDORF: I don't actually have one offhand. I think one thing to be careful of here is that there is -- people often talk about waste, fraud and abuse. So there are -- fraud, as you know, is a narrower category of legally objectionable actions. There's a lot of payments that happen in Medicare that are discovered later to be inappropriate for reasons besides fraud. If you don't put in the right Social Security number for somebody, then you end up making an inappropriate payment. If that -- that may be done accidentally, in which case I think it's not fraud. And if you had them do it right, you might still make the payment.

So there are different levels of what one views as sort of the bad payments. The fraud is the -- I think, the most -- the thing people are most unhappy about, but it's the narrower category.

FREELAND: And how about if you took all the bad payments you have in there?

ELMENDORF: Well, once you get into the question of waste, then you get -- then that becomes a much fuzzier -- a much fuzzier term and, I think, a sort of permeable boundary. And I don't know how to put a number on that.

FREELAND: OK, good way of evading it.


ELMENDORF: Thank you.

FREELAND: Yeah. (Chuckles.)

Q: Frank Weil. I have a question. Without being pessimistic, assuming that the special committee fails, the goalposts on either side of them provide for certain things to happen. Have you scored the outcome of those events?

ELMENDORF: Yes, so if the special committee does not propose any changes in policy, or if they do and they're not enacted by Congress, then there'll be automatic cutbacks in spending in a way that was specified in law. We did an analysis of what that would lead. We released this in the middle of September, and those who are interested, we have a website, and everything we do publicly is on the website.

The effect would be substantial cutbacks in defense spending, substantial cutbacks in nondefense discretionary spending and substantial cutbacks in some mandatory spending and entitlement programs apart from those that are protected in certain ways. Social Security and Medicaid are insulated from those cutbacks, and Medicare's cutbacks are limited to 2 percent in most of the areas of Medicare. The remaining entitlement programs would take a significant whack, as would the defense and nondefense discretionary spending. And the precise amount I don't remember. There are hundreds of billions of dollars in each of those categories over the next decade, an amount meant to add up to $1.2 trillion, which was the minimum amount that the committee is supposed to achieve.


We'll take a question from the back.

Q: (Name and affiliation inaudible.) If you look back 40 years, as you did, would you say that the deficit we have, which sounds like structural, comes more from the revenue side or the tax side -- sorry, or the -- or the expenditure side? And does that suggest to you that the adjustment should be more on the -- on one side or the other? It seems 15 percent of GDP on the revenue side is much lower than the secular level going back 40 years.

ELMENDORF: Well, right at this point in time revenues are unusually low, as you say, relative to the past 40 years -- past 60 years, in fact, and spending is unusually high. But both of those are driven, importantly, by the current state of the economy. And that's not going away right away, of course. But as one thinks about structural changes in the budget, I think it's better to focus on where we're going down the road at the point at which many of these structural changes might be taking effect. That's why I tried to focus on 2021.

And under current law, as I said, revenues in 2021 will be a lot higher than in the history, and spending, apart from Social Security and the major health programs, will be a lot lower. That's actually enough to accommodate that growth in the Social Security and the health programs with pretty small deficits. The problem is that many people have legitimate concerns about revenue being so high and about spending on the other purposes being so low.

But which side the adjustment should come on is not really an analytic question. There's a gap. Alfred Marshall, the great 19th century economist, said -- talked about supply and demand and said you couldn't really ask which one was affecting price; they were like two blades of the scissors. And there's a little of that here. It's a -- it's a choice. I think the point I'm trying to emphasize is that it's a choice we need to make, and it's not something that one can just apply some past rule of thumb to the future.

We need to make a more or less explicit and deliberate choice about how we will deviate from past policies. But it's a -- that's a matter of values as well as of analysis, and that's why I don't get to make policy recommendations.

FREELAND: Just a quick follow-up there. But does it -- I mean, maybe it should be explicit and deliberate. It doesn't have to be, though. It's possible that there will be a muddle through. It might not work very well.

ELMENDORF: Well, it doesn't -- yes, there could be a muddle through. I mean, I think the point I'm trying to make is that the changes need to be pretty large by the standards of incremental fiscal policy. So I'm not suggesting that we will all sit down on one day and then there will be a solution. What I am saying is that the set of changes that will have to be made will be sufficiently large that they will not look like just another little piece of fiscal policy, that they will be something where we will have to -- we may not be explicit, but I think we will have to deliberately make certain changes that will be important relative to what we've done in the past.

FREELAND: OK, thank you.

Please, ma'am.

Q: Elizabeth Bramwell, Bramwell Capital. I wonder if you could talk about the interest expense and how big it is relative to the budget and how it's changed over time, how it compares to what we spend on defense. And we -- (inaudible) -- have so low interest rates -- you know, supposing they go up 2(00) or 300 basis points, you know, what does that do to the projections? And what are you assuming on interest rates?

ELMENDORF: Yeah, so I didn't talk much about -- at all about interest payments in my comments. I ducked that issue because the amount of interest that one would pay in 2021 depended on the path of policies between here and there, and I wanted to focus on sort of picturing a year in the future.

At the moment, interest payments are quite low because interest rates are extremely low. Our projections have interest payments by the federal government, I think, tripling in dollar terms over the coming decade, doubling roughly as a share of GDP. So if one -- you know, the debt that we have now, the debt held by the public from the federal government, is about 67 percent of GDP. That is nearly double, I think, the average over the past 40 years. So if interest rates go back to the average that we've seen, then we'll end up with debt service payments that are about double their average. And of course, that's what makes it impossible to have revenues and non- interest spending diverge by large amounts indefinitely, is that the compounding debt and the rising interest payments need to be serviced.

But -- so what our baseline, our projection envisions is a gradual rise in interest rates beginning, I think, in 2013 mostly -- certainly the short-term interest rates beginning to rise in 2013, the longer-term rates rising a little before that to levels that are about, I think, at the average we've seen over the past decades.

But there's -- as you -- many of you may understand better than I do, there's a tension here -- the federal government is doing a lot of borrowing, but a lot of private organizations are not doing as much borrowing. And for all the worry about our economy and budget that we're talking about here -- of course there are much worse problems in other parts of the world -- and that brings money into this country. So the rate at which interest rates are going to -- the path in which they're going to follow is one of the very uncertain parts in our forecast.

And I wouldn't put a lot of -- a lot of weight on our forecast for any particular year. I think the underlying point is just that if one has an outstanding level of debt that's twice as large relative to the economies we've had in the past, that's going to be much more interest-burden. And that is one of the costs of that higher level of debt. And having people talk about trying to stop that from rising -- which of course is a necessary step at some point -- but if one wants to bring those interest payments back down under more normal interest rate assumptions, then you'll need to actually bring the debt back down over time. And that's a yet bigger challenge.

Q: (Off mic.)

ELMENDORF: It's not -- I don't remember for sure. It's not very -- it's not very large right now. But I -- again, I think that's a misleading -- I think that offers a misleading picture of the role of interest payments in the federal budget challenge.

FREELAND: And looking forward to 2021, how much do these different assumptions of where interest rates might be move the dial?

ELMENDORF: Interest rate assumptions matter a fair bit. They matter more than they used to because of the larger amount of debt that we have. But it depends -- and of course, they -- we could be off in that projection by a great deal. We're trying to have a projection that's in the middle of the range of possible outcomes. But the range is quite wide.


Q: Steve Robert (sp). I just heard you use a phrase that we hear a lot, "debt held by the public." And I understand that's 67 percent. I believe if you use total debt, including the debt to the various trust funds, Social Security, et cetera, it would be closer to hundred percent.

Why do we use this "debt held by the public" number, which is so often used in debate, showing us at 67 percent? Is that debt -- is this a distinction without a difference? Is this kind of a cheat to make our debt-to-GDP look lower? Or is there some rationale for this?

ELMENDORF: Option C. There's a rationale to it. (Laughter.) The rationale is that the -- almost all the work that CBO does and that many budget analysts do, we look at the overall federal budget, what's called the unified budget. This was adopted in the late 1960s, before CBO even existed, as a way of avoiding cheating, as a way of avoiding people doing things that were -- the federal government was engaged in and somehow hiding them off to the side. And the view was we're going to put all the pieces together.

What that means is that when we look at the current liabilities of the federal government, what we mean by that are the liabilities of the government to people outside of the government. Debt held by the public, is the measure of that. We can talk about financial asset holdings and other things. That's why we chose debt held by the public.

If we look at the fiscal picture going forward, then what we think is most useful for Congress and the public is to see the payments and receipts that would occur under current law. So we do projections of overall spending and overall revenues of the sort that I've just provided to you, and we do that over a 10-year window much of the time, but also over a 25-year and 75-year horizon.

We don't think that total debt is a very useful measure either of the government's current posture vis-a-vis the rest of the private holders of securities or a very good measure of the future fiscal challenge. The money in the Social Security trust fund and in the Medicare Part A, the hospital insurance trust fund, is a small -- it does not reflect the total cost those programs would bear under current law, so you're not getting to the true size of the fiscal challenge by putting those pieces in.

But also, holdings of government debt by other funds in the government that actually don't correspond to future obligations in the same sense that Social Security and Medicare have future obligations, you don't want to put those pieces in really at all.

So it's true that number is larger and that in some ways reflects, you know, past policy decisions to put money in trust funds to help make future payments out of those funds, but it's a very partial picture and, we think, can be a misleading picture. And we think a much better way to get a sense of the overall budget situation is to look at the interactions of the government on a unified basis with the U.S. private sector or the rest of the world. And debt held by the public combined with projections of overall spending and revenues we think are the best way to do that. And I think many other analysts share that perspective as well.

FREELAND: OK. Good to know you're not engaged in the gaming yourself.


Q: Thanks. Ted Schell from Associated Partners. Two quick questions. One is --

FREELAND: Do you mind just doing one? Because there are lots of people.

Q: OK. You hear the argument that, you know, the Obama policies of the last few years have contributed to the deficit. The other side of it is that you have no idea how bad they would have been had -- it would have been had they not been put into effect. Is there an analytic answer to that question?

ELMENDORF: There could be. We have not done the analysis to produce that. So we do -- we've estimated -- and we actually are every quarter under law required to report on the economic effects of the American Recovery and Reinvestment Act of 2009. And we believe that that act raised output and employment over the past 2 1/2 years relative to what would have occurred in the absence of those policies. And that higher level of output has itself had a positive feedback effect on the budget.

But we have not tried to do an encompassing estimate of all the of the policies that have been enacted and their budgetary feedbacks in an effort to produce what I think you're asking for, which would be a net budgetary effect of those fiscal policies after accounting for the economic feedbacks.

We do that sort of analysis in a variety of forward-looking ways. I think that sort of work is very important. And when we look at large changes in fiscal policy, every year the analysis of the president's budget that we do, when we did analysis of the effects of extending the expiring tax provisions in certain ways, and many other products, we do try to look at the effect of budget policies on the economy and the feedback into the budget and try to present that.

But we've done that for proposals looking forward, not for the historical perspective.

FREELAND: If we asked you to do a historical analysis like that, would you?

ELMENDORF: Well, as much as I'd like to, I couldn't do it for you --

FREELAND: It would be quite politically --

ELMENDORF: -- but if you -- if you -- if you worked --

FREELAND: Council on Foreign Relations?

ELMENDORF: Well, even for the Council on Foreign Relations, if you found a chair or a ranking member of one of the principal congressional committees and they wanted to put that request ahead of the many others from them -- (chuckles) -- then we could do that, yes.

FREELAND: It would be pretty politically explosive, though, wouldn't it? Because it would effectively giving like a report card to the historic performance of --

ELMENDORF: We don't -- we don't think about the political aspects of things. If we're asked to do analysis by members of Congress, we do the analysis. And the political ramifications of that are not what I'm getting paid to worry about.

FREELAND: OK. Let's take another question from the back so it's not too -- and also, OK, I've been caught out from a -- (inaudible) -- reporter, which we should have.

Q: Hi -- (inaudible) -- Matsura (sp). I'm a columnist for Japan's Sankei News. I'd like to (assess ?) the confidence level or optimism of the -- (inaudible) -- forecast by asking the specifics. What is the standard error of those three numbers, the revenue -- the revenue and expending -- revenue and spending and the sum of this, which will be zero in 2021? I just -- and what's your confidence level?

ELMENDORF: So I don't have numerical answers for you. We publish regularly analyses of our macroeconomic forecasts and the errors in them, and the confidences regions are wide. We don't do that for the budget in that same way, and the reason -- what we do do instead is, every time we release a new set of budget projections we decompose the revisions from the previous set into those that are attributable to surprises in the performance of the overall economy, those that are attributable to legislative changes that were enacted, and those that are attributable to what we call technical errors, everything else -- but you could think about that as, conditional on the unemployment rate, did we predict the take-up for food stamps correctly?

We do decompositions each time. One can accumulate those over time. People sometimes do. I used to, before I came to CBO, occasionally for papers I was writing. We don't do that in a formal way ourselves. But I think what you -- but it's worth noting here that part of the reason we'll be wrong in our projections of spending and revenues is that Congress will respond to those projections or to other developments in the world and change policy. So, you know, projections we've just been talking about are not what I think will happen in 2021. I think other things will happen; don't know exactly what they are.

So I think the thing that we're trying to do is to minimize the errors for economic and technical reasons, not those for legislative reasons. So in terms of checking our own work -- well, we want to take out the legislative piece. And so we haven't done that. But I'm pretty confident though that confidence regions will be wide there as well, and one should recognize that. And if one looks back at our projections a decade or so ago, which were for large surpluses for the first decade of this century, there's a very big error. And we have actually reported a sort of crude decomposition of that, and there are some technical errors and some economic errors and important legislative changes as well.

But -- well, all of the specific numbers that I offer and that we offer in all of our work should be all viewed as, hopefully, the midpoint of a substantial range. For communication purposes, we sometimes report ranges, but more often, we stick with the -- with the point estimates. But there's no doubt our point estimates in ranges.

FREELAND: OK. I think we have time just for one last question, because it's 9:00.


Q: Thank you. Michael Fellman (sp), from Morgan Stanley. How much has the CBO looked at potential savings from preventative health care and economic incentives used to encourage people to lead healthier lives, therefore avoiding longer-term costs?

ELMENDORF: That's an important question. We've done some work on preventive health care, and have much more under way. A particular project we're doing now is to look at the effects of changes in tobacco policies on smoking, and then on health, and then on the federal budget. And in the past when we've looked at policies like that, we've incorporated some feedback effects, but we're now engaged in a very large effort to try to trace through effects in all of the components of the federal budget.

I think -- and that's under way, and we hope to finish that shortly. But we view that as important for its own sake, but also as establishing a template that we hope we can follow and other people can comment on in the analytic community to look at more issues of the sort.

The challenges are, first of all, finding federal policies that affect people's behavior. Tobacco is an area where there's been a lot of work for decades now about how to reduce smoking. Obesity is another health condition that has a lot of effects on health. The state of knowledge of how to affect people's weight is much less advanced than how to affect their smoking, the more complicated problem in many ways. Many things are being tried. There's a lot of research that we look at. But it's less clear what the policy -- federal policy changes would be that would affect obesity.

And then there's the issue about how changes in health affect the budget. And you got to recognize there that there are effects in different directions. People who are healthier need less health care, and that reduces spending on Medicare and Medicaid, perhaps. But people who are healthier live longer, and that can increase spending on Social Security. That doesn't mean it's bad, of course -- (laughter) -- but if you're asking me about the budget effects -- (chuckles) -- then I need to take account of that. So in the estimates that we're working on now for tobacco policy, but for anything of this sort, one needs to take account of the things that are good for the budget and bad for the budget.

And people -- for this point in this conversation, people always sort of laugh and occasionally accuse me of not caring. (Laughter.) And it's worth just saying -- (chuckles) -- you know, lots of things are worth spending money on. Preventive health care -- and people in this business are -- say this a lot when we talk with them. They don't understand why spending on health treatments -- somehow we expect to spend money on that, but preventive health care, the advocates often feel, are in a box of -- they have to show that it saves money to do. That's probably not the right way to think about the problem. We should spend -- if we're spending money on health care, we should be prepared to spend money on prevention.

It's -- for us, it's very important to try to give the most accurate sense we can about how much net money one is spending to do a certain sort of prevention, which is why we're doing this work. But I think nobody should believe that these things have to pay for themselves to be worth doing. It may well be that there are effective ways to -- if the goal is to improve health, it may well be that an effective allocation of federal resources is partly on health care and partly on prevention of disease or efforts to keep people healthier.

FREELAND: Thank you very much. One of the -- just as a final question, one of the strains running through our conversation has been, I think, our collective effort to evade the choice you're saying the country faces. (Laughter.) So you know, maybe fraud -- you know, if you cut out fraud, that's the way you avoid this choice. I asked you about making health care more efficient, a question about preventive health care. Is there an area that we haven't asked you about that would help to evade this big choice -- (laughter) -- between cutting government provision and raising taxes?

ELMENDORF: There are lots of ways to evade choices.

FREELAND: But a good one, like one that you score as, you know, working.

ELMENDORF: An effective evasion.


ELMENDORF: No, I'm sorry, I -- (chuckles) -- I don't know of -- have any at hand, and if I did, I'd -- certainly not going to encourage a distinguished audience like this to follow that path. (Laughter.) I think you should, as I know you are, think hard about the choices and what choices you want the country to make.

FREELAND: OK, well, thank you very much -- (inaudible). (Applause.)

ELMENDORF: Thank you. It's good to be here.






Top Stories on CFR



Forget regional integration. Basic diplomacy will remain an uphill battle until the region’s leaders return to a shared definition of democracy.

United States

Now that Congress voted to pass a bill based on the Biden-McCarthy compromise, an immediate debt ceiling crisis appears to have been averted. Still, a much larger debt problem awaits.